Showing preview only (895K chars total). Download the full file or copy to clipboard to get everything.
Repository: gpakosz/PackedArray
Branch: master
Commit: a4cffc21a76f
Files: 39
Total size: 862.6 KB
Directory structure:
gitextract_wmqt4bm2/
├── .gitattributes
├── .gitignore
├── .travis.yml
├── LICENSE
├── PackedArray.c
├── PackedArray.h
├── PackedArraySIMD.c
├── README.md
├── _gnu-make/
│ └── Makefile
├── _ios-xcode/
│ ├── .gitignore
│ ├── PackedArray-Info.plist
│ └── PackedArray.xcodeproj/
│ ├── project.pbxproj
│ └── project.xcworkspace/
│ └── contents.xcworkspacedata
├── _mac-xcode/
│ ├── .gitignore
│ └── PackedArray.xcodeproj/
│ ├── project.pbxproj
│ └── project.xcworkspace/
│ └── contents.xcworkspacedata
├── _win-vs11/
│ ├── .gitignore
│ ├── Common.props
│ ├── Debug.props
│ ├── PackedArray.sln
│ ├── PackedArraySIMDSelfBench.vcxproj
│ ├── PackedArraySIMDSelfTest.vcxproj
│ ├── PackedArraySelfBench.vcxproj
│ ├── PackedArraySelfTest.vcxproj
│ ├── Release.props
│ ├── x64.props
│ └── x86.props
└── benchmark/
├── PackedArraySIMDSelfBench-unrolled-galaxy-note-cortex-a9-1.4GHz.txt
├── PackedArraySIMDSelfBench-unrolled-ipad2-cortex-a9-1GHz.txt
├── PackedArraySIMDSelfBench-unrolled-iphone5-a6-1.3GHz.txt
├── PackedArraySIMDSelfBench-unrolled-mbp-corei7-M620-2.67GHz.txt
├── PackedArraySelfBench-reference-galaxy-note-cortex-a9-1.4GHz.txt
├── PackedArraySelfBench-reference-ipad2-cortex-a9-1GHz.txt
├── PackedArraySelfBench-reference-iphone5-a6-1.3GHz.txt
├── PackedArraySelfBench-reference-mbp-corei7-M620-2.67GHz.txt
├── PackedArraySelfBench-unrolled-galaxy-note-cortex-a9-1.4GHz.txt
├── PackedArraySelfBench-unrolled-ipad2-cortex-a9-1GHz.txt
├── PackedArraySelfBench-unrolled-iphone5-a6-1.3GHz.txt
└── PackedArraySelfBench-unrolled-mbp-corei7-M620-2.67GHz.txt
================================================
FILE CONTENTS
================================================
================================================
FILE: .gitattributes
================================================
* text=auto
# sources
*.h text diff=cpp
*.c text diff=cpp
*.cpp text diff=cpp
*.rb text diff=ruby
*.html text diff=html
*.m text diff=objc
# shell scripts
*.sh eol=lf
# GNU Makefile
Makefile text eol=lf
# Autotools
*.am text eol=lf
# Android
*.mk text eol=lf
# Xcode files
*.pbxproj text eol=lf merge=union
# Visual Studio files
*.sln text eol=crlf merge=union
*.vcxproj text eol=crlf merge=union
*.vcxproj.filters text eol=crlf merge=union
*.props text eol=crlf
================================================
FILE: .gitignore
================================================
.DS_Store
*~
*.swp
/bin
PackedArray.pp.c
PackedArray.cut.c
PackedArraySIMD.pp.c
PackedArraySIMD.cut.c
================================================
FILE: .travis.yml
================================================
language: cpp
compiler:
- clang
- gcc
env:
- TARGET=build
- TARGET=test
- TARGET=preprocess
- TARGET=cut
- TARGET=assembly
script: make -j -C ./_gnu-make $TARGET
================================================
FILE: LICENSE
================================================
DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE
Version 2, December 2004
Copyright (C) 2004 Sam Hocevar <sam@hocevar.net>
Everyone is permitted to copy and distribute verbatim or modified
copies of this license document, and changing it is allowed as long
as the name is changed.
DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE
TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
0. You just DO WHAT THE FUCK YOU WANT TO.
1. Bla bla bla
2. Montesqieu et camembert, vive la France, zut alors!
================================================
FILE: PackedArray.c
================================================
// see README.md for usage instructions.
// (‑●‑●)> released under the WTFPL v2 license, by Gregory Pakosz (@gpakosz)
#ifndef PACKEDARRAY_SELF
#define PACKEDARRAY_SELF "PackedArray.c"
#endif
#ifdef PACKEDARRAY_IMPL
#ifndef PACKEDARRAY_JOIN
#define PACKEDARRAY_JOIN(lhs, rhs) PACKEDARRAY_JOIN_(lhs, rhs)
#define PACKEDARRAY_JOIN_(lhs, rhs) PACKEDARRAY_JOIN__(lhs, rhs)
#define PACKEDARRAY_JOIN__(lhs, rhs) lhs##rhs
#endif // #ifndef PACKEDARRAY_JOIN
#ifndef PACKEDARRAY_IMPL_BITS_PER_ITEM
#error PACKEDARRAY_IMPL_BITS_PER_ITEM undefined
#endif // #ifndef PACKEDARRAY_IMPL_BITS_PER_ITEM
#if defined(PACKEDARRAY_IMPL_PACK_CASES) || defined(PACKEDARRAY_IMPL_UNPACK_CASES)
#ifndef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 0
#elif PACKEDARRAY_IMPL_CASE_I == 0
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 1
#elif PACKEDARRAY_IMPL_CASE_I == 1
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 2
#elif PACKEDARRAY_IMPL_CASE_I == 2
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 3
#elif PACKEDARRAY_IMPL_CASE_I == 3
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 4
#elif PACKEDARRAY_IMPL_CASE_I == 4
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 5
#elif PACKEDARRAY_IMPL_CASE_I == 5
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 6
#elif PACKEDARRAY_IMPL_CASE_I == 6
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 7
#elif PACKEDARRAY_IMPL_CASE_I == 7
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 8
#elif PACKEDARRAY_IMPL_CASE_I == 8
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 9
#elif PACKEDARRAY_IMPL_CASE_I == 9
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 10
#elif PACKEDARRAY_IMPL_CASE_I == 10
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 11
#elif PACKEDARRAY_IMPL_CASE_I == 11
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 12
#elif PACKEDARRAY_IMPL_CASE_I == 12
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 13
#elif PACKEDARRAY_IMPL_CASE_I == 13
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 14
#elif PACKEDARRAY_IMPL_CASE_I == 14
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 15
#elif PACKEDARRAY_IMPL_CASE_I == 15
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 16
#elif PACKEDARRAY_IMPL_CASE_I == 16
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 17
#elif PACKEDARRAY_IMPL_CASE_I == 17
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 18
#elif PACKEDARRAY_IMPL_CASE_I == 18
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 19
#elif PACKEDARRAY_IMPL_CASE_I == 19
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 20
#elif PACKEDARRAY_IMPL_CASE_I == 20
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 21
#elif PACKEDARRAY_IMPL_CASE_I == 21
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 22
#elif PACKEDARRAY_IMPL_CASE_I == 22
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 23
#elif PACKEDARRAY_IMPL_CASE_I == 23
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 24
#elif PACKEDARRAY_IMPL_CASE_I == 24
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 25
#elif PACKEDARRAY_IMPL_CASE_I == 25
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 26
#elif PACKEDARRAY_IMPL_CASE_I == 26
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 27
#elif PACKEDARRAY_IMPL_CASE_I == 27
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 28
#elif PACKEDARRAY_IMPL_CASE_I == 28
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 29
#elif PACKEDARRAY_IMPL_CASE_I == 29
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 30
#elif PACKEDARRAY_IMPL_CASE_I == 30
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 31
#elif PACKEDARRAY_IMPL_CASE_I == 31
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 32
#endif // #ifndef PACKEDARRAY_IMPL_CASE_I
#ifndef PACKEDARRAY_IMPL_BITS_AVAILABLE
#define PACKEDARRAY_IMPL_BITS_AVAILABLE (32 - ((PACKEDARRAY_IMPL_CASE_I * PACKEDARRAY_IMPL_BITS_PER_ITEM) % 32))
#endif
#ifndef PACKEDARRAY_IMPL_START_BIT
#define PACKEDARRAY_IMPL_START_BIT ((PACKEDARRAY_IMPL_CASE_I * PACKEDARRAY_IMPL_BITS_PER_ITEM) % 32)
#endif
#ifndef PACKEDARRAY_IMPL_MASK
#define PACKEDARRAY_IMPL_MASK (uint32_t)((1ULL << PACKEDARRAY_IMPL_BITS_PER_ITEM) - 1)
#endif
#if defined(PACKEDARRAY_IMPL_PACK_CASES)
#ifndef PACKEDARRAY_IMPL_PACK_CASE_BREAK
#define PACKEDARRAY_IMPL_PACK_CASE_BREAK
#endif
case PACKEDARRAY_IMPL_CASE_I:
#if (PACKEDARRAY_IMPL_BITS_PER_ITEM <= PACKEDARRAY_IMPL_BITS_AVAILABLE)
packed |= *in++ << PACKEDARRAY_IMPL_START_BIT;
#if (PACKEDARRAY_IMPL_BITS_PER_ITEM == PACKEDARRAY_IMPL_BITS_AVAILABLE)
*out++ = packed;
packed = 0;
#endif
#else
packed |= *in << PACKEDARRAY_IMPL_START_BIT;
*out++ = packed;
packed = *in++ >> PACKEDARRAY_IMPL_BITS_AVAILABLE;
#endif
PACKEDARRAY_IMPL_PACK_CASE_BREAK
#if PACKEDARRAY_IMPL_CASE_I < 31
#include PACKEDARRAY_SELF
#else
#undef PACKEDARRAY_IMPL_CASE_I
#undef PACKEDARRAY_IMPL_PACK_CASE_BREAK
#undef PACKEDARRAY_IMPL_PACK_CASES
#endif
#elif defined(PACKEDARRAY_IMPL_UNPACK_CASES) // #if defined(PACKEDARRAY_IMPL_PACK_CASES)
#ifndef PACKEDARRAY_IMPL_UNPACK_CASE_BREAK
#define PACKEDARRAY_IMPL_UNPACK_CASE_BREAK
#endif
case PACKEDARRAY_IMPL_CASE_I:
#if (PACKEDARRAY_IMPL_BITS_PER_ITEM <= PACKEDARRAY_IMPL_BITS_AVAILABLE)
*out++ = (packed >> PACKEDARRAY_IMPL_START_BIT) & PACKEDARRAY_IMPL_MASK;
PACKEDARRAY_IMPL_UNPACK_CASE_BREAK
#if (PACKEDARRAY_IMPL_CASE_I < 31) && (PACKEDARRAY_IMPL_BITS_PER_ITEM == PACKEDARRAY_IMPL_BITS_AVAILABLE)
packed = *++in;
#endif
#else
{
uint32_t low, high;
low = packed >> PACKEDARRAY_IMPL_START_BIT;
packed = *++in;
high = packed << PACKEDARRAY_IMPL_BITS_AVAILABLE;
*out++ = (low | high) & PACKEDARRAY_IMPL_MASK;
}
PACKEDARRAY_IMPL_UNPACK_CASE_BREAK
#endif
#if PACKEDARRAY_IMPL_CASE_I < 31
#include PACKEDARRAY_SELF
#else
#undef PACKEDARRAY_IMPL_CASE_I
#undef PACKEDARRAY_IMPL_UNPACK_CASE_BREAK
#undef PACKEDARRAY_IMPL_UNPACK_CASES
#endif
#endif // #elif defined(PACKEDARRAY_IMPL_UNPACK_CASES)
#else // #if defined(PACKEDARRAY_IMPL_PACK_CASES) || defined(PACKEDARRAY_IMPL_UNPACK_CASES)
void PACKEDARRAY_JOIN(__PackedArray_pack_, PACKEDARRAY_IMPL_BITS_PER_ITEM)(uint32_t* __restrict out, uint32_t offset, const uint32_t* __restrict in, uint32_t count)
{
uint32_t startBit;
uint32_t packed;
const uint32_t* __restrict end;
out += ((uint64_t)offset * (uint64_t)PACKEDARRAY_IMPL_BITS_PER_ITEM) / 32;
startBit = ((uint64_t)offset * (uint64_t)PACKEDARRAY_IMPL_BITS_PER_ITEM) % 32;
packed = *out & (uint32_t)((1ULL << startBit) - 1);
offset = offset % 32;
if (count >= 32 - offset)
{
int32_t n;
n = (count + offset) / 32;
count -= 32 * n - offset;
switch (offset)
{
do
{
#define PACKEDARRAY_IMPL_PACK_CASES
#include PACKEDARRAY_SELF
} while (--n > 0);
}
if (count == 0)
return;
offset = 0;
startBit = 0;
}
end = in + count;
switch (offset)
{
#define PACKEDARRAY_IMPL_PACK_CASES
#define PACKEDARRAY_IMPL_PACK_CASE_BREAK \
if (in == end)\
break;
#include PACKEDARRAY_SELF
}
PACKEDARRAY_ASSERT(in == end);
if ((count * PACKEDARRAY_IMPL_BITS_PER_ITEM + startBit) % 32)
{
packed |= *out & ~((uint32_t)(1ULL << ((((uint64_t)count * (uint64_t)PACKEDARRAY_IMPL_BITS_PER_ITEM + startBit - 1) % 32) + 1)) - 1);
*out = packed;
}
}
void PACKEDARRAY_JOIN(__PackedArray_unpack_, PACKEDARRAY_IMPL_BITS_PER_ITEM)(const uint32_t* __restrict in, uint32_t offset, uint32_t* __restrict out, uint32_t count)
{
uint32_t packed;
const uint32_t* __restrict end;
in += ((uint64_t)offset * (uint64_t)PACKEDARRAY_IMPL_BITS_PER_ITEM) / 32;
packed = *in;
offset = offset % 32;
if (count >= 32 - offset)
{
int32_t n;
n = (count + offset) / 32;
count -= 32 * n - offset;
switch (offset)
{
do
{
packed = *++in;
#define PACKEDARRAY_IMPL_UNPACK_CASES
#include PACKEDARRAY_SELF
} while (--n > 0);
}
if (count == 0)
return;
packed = *++in;
offset = 0;
}
end = out + count;
switch (offset)
{
#define PACKEDARRAY_IMPL_UNPACK_CASES
#define PACKEDARRAY_IMPL_UNPACK_CASE_BREAK \
if (out == end)\
break;
#include PACKEDARRAY_SELF
}
PACKEDARRAY_ASSERT(out == end);
}
#undef PACKEDARRAY_IMPL_BITS_PER_ITEM
#undef PACKEDARRAY_IMPL_BITS_AVAILABLE
#undef PACKEDARRAY_IMPL_START_BIT
#undef PACKEDARRAY_IMPL_START_MASK
#endif // #if defined(PACKEDARRAY_IMPL_PACK_CASES) || defined(PACKEDARRAY_IMPL_UNPACK_CASES)
#else
#include "PackedArray.h"
#if !defined(PACKEDARRAY_ASSERT)
#include <assert.h>
#define PACKEDARRAY_ASSERT(expression) assert(expression)
#endif
#define PACKEDARRAY_IMPL
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 1
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 2
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 3
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 4
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 5
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 6
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 7
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 8
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 9
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 10
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 11
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 12
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 13
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 14
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 15
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 16
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 17
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 18
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 19
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 20
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 21
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 22
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 23
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 24
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 25
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 26
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 27
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 28
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 29
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 30
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 31
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 32
#include PACKEDARRAY_SELF
#undef PACKEDARRAY_IMPL
#if !defined(PACKEDARRAY_MALLOC) || !defined(PACKEDARRAY_FREE)
#include <stdlib.h>
#endif
#if !defined(PACKEDARRAY_MALLOC)
#define PACKEDARRAY_MALLOC(size) malloc(size)
#endif
#if !defined(PACKEDARRAY_FREE)
#define PACKEDARRAY_FREE(p) free(p)
#endif
#include <stddef.h>
PackedArray* PackedArray_create(uint32_t bitsPerItem, uint32_t count)
{
PackedArray* a;
size_t bufferSize;
PACKEDARRAY_ASSERT(bitsPerItem > 0);
PACKEDARRAY_ASSERT(bitsPerItem <= 32);
bufferSize = sizeof(uint32_t) * (((uint64_t)bitsPerItem * (uint64_t)count + 31) / 32);
a = (PackedArray*)PACKEDARRAY_MALLOC(sizeof(PackedArray) + bufferSize);
if (a != NULL)
{
a->buffer[((uint64_t)bitsPerItem * (uint64_t)count + 31) / 32 - 1] = 0;
a->bitsPerItem = bitsPerItem;
a->count = count;
}
return a;
}
void PackedArray_destroy(PackedArray* a)
{
PACKEDARRAY_ASSERT(a);
PACKEDARRAY_FREE(a);
}
void PackedArray_pack(PackedArray* a, const uint32_t offset, const uint32_t* in, uint32_t count)
{
PACKEDARRAY_ASSERT(a != NULL);
PACKEDARRAY_ASSERT(in != NULL);
switch (a->bitsPerItem)
{
case 1: __PackedArray_pack_1(a->buffer, offset, in, count); break;
case 2: __PackedArray_pack_2(a->buffer, offset, in, count); break;
case 3: __PackedArray_pack_3(a->buffer, offset, in, count); break;
case 4: __PackedArray_pack_4(a->buffer, offset, in, count); break;
case 5: __PackedArray_pack_5(a->buffer, offset, in, count); break;
case 6: __PackedArray_pack_6(a->buffer, offset, in, count); break;
case 7: __PackedArray_pack_7(a->buffer, offset, in, count); break;
case 8: __PackedArray_pack_8(a->buffer, offset, in, count); break;
case 9: __PackedArray_pack_9(a->buffer, offset, in, count); break;
case 10: __PackedArray_pack_10(a->buffer, offset, in, count); break;
case 11: __PackedArray_pack_11(a->buffer, offset, in, count); break;
case 12: __PackedArray_pack_12(a->buffer, offset, in, count); break;
case 13: __PackedArray_pack_13(a->buffer, offset, in, count); break;
case 14: __PackedArray_pack_14(a->buffer, offset, in, count); break;
case 15: __PackedArray_pack_15(a->buffer, offset, in, count); break;
case 16: __PackedArray_pack_16(a->buffer, offset, in, count); break;
case 17: __PackedArray_pack_17(a->buffer, offset, in, count); break;
case 18: __PackedArray_pack_18(a->buffer, offset, in, count); break;
case 19: __PackedArray_pack_19(a->buffer, offset, in, count); break;
case 20: __PackedArray_pack_20(a->buffer, offset, in, count); break;
case 21: __PackedArray_pack_21(a->buffer, offset, in, count); break;
case 22: __PackedArray_pack_22(a->buffer, offset, in, count); break;
case 23: __PackedArray_pack_23(a->buffer, offset, in, count); break;
case 24: __PackedArray_pack_24(a->buffer, offset, in, count); break;
case 25: __PackedArray_pack_25(a->buffer, offset, in, count); break;
case 26: __PackedArray_pack_26(a->buffer, offset, in, count); break;
case 27: __PackedArray_pack_27(a->buffer, offset, in, count); break;
case 28: __PackedArray_pack_28(a->buffer, offset, in, count); break;
case 29: __PackedArray_pack_29(a->buffer, offset, in, count); break;
case 30: __PackedArray_pack_30(a->buffer, offset, in, count); break;
case 31: __PackedArray_pack_31(a->buffer, offset, in, count); break;
case 32: __PackedArray_pack_32(a->buffer, offset, in, count); break;
}
}
void PackedArray_unpack(const PackedArray* a, const uint32_t offset, uint32_t* out, uint32_t count)
{
PACKEDARRAY_ASSERT(a != NULL);
PACKEDARRAY_ASSERT(out != NULL);
switch (a->bitsPerItem)
{
case 1: __PackedArray_unpack_1(a->buffer, offset, out, count); break;
case 2: __PackedArray_unpack_2(a->buffer, offset, out, count); break;
case 3: __PackedArray_unpack_3(a->buffer, offset, out, count); break;
case 4: __PackedArray_unpack_4(a->buffer, offset, out, count); break;
case 5: __PackedArray_unpack_5(a->buffer, offset, out, count); break;
case 6: __PackedArray_unpack_6(a->buffer, offset, out, count); break;
case 7: __PackedArray_unpack_7(a->buffer, offset, out, count); break;
case 8: __PackedArray_unpack_8(a->buffer, offset, out, count); break;
case 9: __PackedArray_unpack_9(a->buffer, offset, out, count); break;
case 10: __PackedArray_unpack_10(a->buffer, offset, out, count); break;
case 11: __PackedArray_unpack_11(a->buffer, offset, out, count); break;
case 12: __PackedArray_unpack_12(a->buffer, offset, out, count); break;
case 13: __PackedArray_unpack_13(a->buffer, offset, out, count); break;
case 14: __PackedArray_unpack_14(a->buffer, offset, out, count); break;
case 15: __PackedArray_unpack_15(a->buffer, offset, out, count); break;
case 16: __PackedArray_unpack_16(a->buffer, offset, out, count); break;
case 17: __PackedArray_unpack_17(a->buffer, offset, out, count); break;
case 18: __PackedArray_unpack_18(a->buffer, offset, out, count); break;
case 19: __PackedArray_unpack_19(a->buffer, offset, out, count); break;
case 20: __PackedArray_unpack_20(a->buffer, offset, out, count); break;
case 21: __PackedArray_unpack_21(a->buffer, offset, out, count); break;
case 22: __PackedArray_unpack_22(a->buffer, offset, out, count); break;
case 23: __PackedArray_unpack_23(a->buffer, offset, out, count); break;
case 24: __PackedArray_unpack_24(a->buffer, offset, out, count); break;
case 25: __PackedArray_unpack_25(a->buffer, offset, out, count); break;
case 26: __PackedArray_unpack_26(a->buffer, offset, out, count); break;
case 27: __PackedArray_unpack_27(a->buffer, offset, out, count); break;
case 28: __PackedArray_unpack_28(a->buffer, offset, out, count); break;
case 29: __PackedArray_unpack_29(a->buffer, offset, out, count); break;
case 30: __PackedArray_unpack_30(a->buffer, offset, out, count); break;
case 31: __PackedArray_unpack_31(a->buffer, offset, out, count); break;
case 32: __PackedArray_unpack_32(a->buffer, offset, out, count); break;
}
}
void PackedArray_set(PackedArray* a, const uint32_t offset, const uint32_t in)
{
uint32_t* __restrict out;
uint32_t bitsPerItem;
uint32_t startBit;
uint32_t bitsAvailable;
uint32_t mask;
PACKEDARRAY_ASSERT(a != NULL);
bitsPerItem = a->bitsPerItem;
out = &a->buffer[((uint64_t)offset * (uint64_t)bitsPerItem) / 32];
startBit = ((uint64_t)offset * (uint64_t)bitsPerItem) % 32;
bitsAvailable = 32 - startBit;
mask = (uint32_t)(1ULL << bitsPerItem) - 1;
PACKEDARRAY_ASSERT(0 == (~mask & in));
if (bitsPerItem <= bitsAvailable)
{
out[0] = (out[0] & ~(mask << startBit)) | (in << startBit);
}
else
{
// value spans 2 buffer cells
uint32_t low, high;
low = in << startBit;
high = in >> bitsAvailable;
out[0] = (out[0] & ~(mask << startBit)) | low;
out[1] = (out[1] & ~(mask >> (32 - startBit))) | high;
}
}
uint32_t PackedArray_get(const PackedArray* a, const uint32_t offset)
{
const uint32_t* __restrict in;
uint32_t bitsPerItem;
uint32_t startBit;
uint32_t bitsAvailable;
uint32_t mask;
uint32_t out;
PACKEDARRAY_ASSERT(a != NULL);
bitsPerItem = a->bitsPerItem;
in = &a->buffer[((uint64_t)offset * (uint64_t)bitsPerItem) / 32];
startBit = ((uint64_t)offset * (uint64_t)bitsPerItem) % 32;
bitsAvailable = 32 - startBit;
mask = (uint32_t)(1ULL << bitsPerItem) - 1;
if (bitsPerItem <= bitsAvailable)
{
out = (in[0] >> startBit) & mask;
}
else
{
// out spans 2 buffer cells
uint32_t low, high;
low = in[0] >> startBit;
high = in[1] << (32 - startBit);
out = low ^ ((low ^ high) & (mask >> bitsAvailable << bitsAvailable));
}
return out;
}
uint32_t PackedArray_bufferSize(const PackedArray* a)
{
PACKEDARRAY_ASSERT(a != NULL);
return (uint32_t)(((uint64_t)a->bitsPerItem * (uint64_t)a->count + 31) / 32);
}
#if !(defined(_MSC_VER) && _MSC_VER >= 1400) && !defined(__GNUC__)
// log base 2 of an integer, aka the position of the highest bit set
static uint32_t __PackedArray_log2(uint32_t v)
{
// references
// http://aggregate.org/MAGIC
// http://graphics.stanford.edu/~seander/bithacks.html
static const uint32_t multiplyDeBruijnBitPosition[32] =
{
0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30,
8, 12, 20, 28, 15, 17, 24, 7, 19, 27, 23, 6, 26, 5, 4, 31
};
v |= v >> 1;
v |= v >> 2;
v |= v >> 4;
v |= v >> 8;
v |= v >> 16;
return multiplyDeBruijnBitPosition[(uint32_t)(v * 0x7C4ACDDU) >> 27];
}
#endif
// position of the highest bit set
static int __PackedArray_highestBitSet(uint32_t v)
{
#if defined(_MSC_VER) && _MSC_VER >= 1400
unsigned long index;
return _BitScanReverse(&index, v) ? index : -1;
#elif defined(__GNUC__)
return v == 0 ? -1 : 31 - __builtin_clz(v);
#else
return v != 0 ? __PackedArray_log2(v) : -1;
#endif
}
uint32_t PackedArray_computeBitsPerItem(const uint32_t* in, uint32_t count)
{
uint32_t i, in_max, bitsPerItem;
in_max = 0;
for (i = 0; i < count; ++i)
in_max = in[i] > in_max ? in[i] : in_max;
bitsPerItem = __PackedArray_highestBitSet(in_max) + 1;
return bitsPerItem == 0 ? 1 : bitsPerItem;
}
// - 8< ------------------------------------------------------------------------
#if defined(PACKEDARRAY_SELF_TEST) && defined(PACKEDARRAY_SELF_BENCH)
#error choose either PACKEDARRAY_SELF_TEST or PACKEDARRAY_SELF_BENCH
#endif
#if defined(PACKEDARRAY_SELF_TEST)
#undef NDEBUG // we want asserts
#include <assert.h>
#include <stdio.h>
#include <string.h> // memcmp
static void PackedArray_pack_reference(PackedArray* a, const uint32_t offset, const uint32_t* in, uint32_t count)
{
uint32_t* __restrict out;
uint32_t bitsPerItem;
uint32_t startBit;
uint32_t bitsAvailable;
uint32_t mask;
uint32_t packed;
assert(a != NULL);
assert(in != NULL);
assert(count != 0);
bitsPerItem = a->bitsPerItem;
out = &a->buffer[(uint64_t)offset * (uint64_t)bitsPerItem / 32];
startBit = ((uint64_t)offset * (uint64_t)bitsPerItem) % 32;
bitsAvailable = 32 - startBit;
mask = (uint32_t)(1ULL << bitsPerItem) - 1;
packed = *out;
while (count--)
{
uint32_t value = *in++;
assert(0 == (~mask & value));
if (bitsPerItem <= bitsAvailable)
{
packed = (packed & ~(mask << startBit)) | (value << startBit);
startBit += bitsPerItem;
bitsAvailable -= bitsPerItem;
}
else if (bitsAvailable == 0)
{
*out++ = packed;
packed = *out;
startBit = 0;
bitsAvailable = 32;
packed = (packed & ~mask) | value;
startBit += bitsPerItem;
bitsAvailable -= bitsPerItem;
}
else
{
// value spans 2 buffer cells
uint32_t low, high;
low = value << startBit;
high = value >> bitsAvailable;
packed = (packed & ~(mask << startBit)) | low;
*out++ = packed;
packed = *out;
packed = (packed & ~(mask >> (32 - startBit))) | high;
startBit = (startBit + bitsPerItem) % 32;
bitsAvailable = 32 - startBit;
}
}
*out = packed;
}
static void PackedArray_unpack_reference(const PackedArray* a, const uint32_t offset, uint32_t* out, uint32_t count)
{
const uint32_t* __restrict in;
uint32_t bitsPerItem;
uint32_t startBit;
uint32_t bitsAvailable;
uint32_t mask;
uint32_t packed;
assert(a != NULL);
assert(out != NULL);
assert(count != 0);
bitsPerItem = a->bitsPerItem;
in = &a->buffer[(uint64_t)offset * (uint64_t)bitsPerItem / 32];
startBit = ((uint64_t)offset * (uint64_t)bitsPerItem) % 32;
bitsAvailable = 32 - startBit;
mask = (uint32_t)(1ULL << bitsPerItem) - 1;
packed = *in;
while (count--)
{
uint32_t value;
if (bitsPerItem <= bitsAvailable)
{
value = (packed >> startBit) & mask;
*out++ = value;
startBit += bitsPerItem;
bitsAvailable -= bitsPerItem;
}
else if (bitsAvailable == 0)
{
packed = *++in;
value = packed & mask;
*out++ = value;
startBit = bitsPerItem;
bitsAvailable = 32 - bitsPerItem;
}
else
{
// value spans 2 buffer cells
uint32_t low, high;
low = packed >> startBit;
packed = *++in;
high = packed << (32 - startBit);
value = low ^ ((low ^ high) & (mask >> bitsAvailable << bitsAvailable));
*out++ = value;
startBit = (startBit + bitsPerItem) % 32;
bitsAvailable = 32 - startBit;
}
}
}
int main(void)
{
uint32_t bitsPerItem;
printf("-- PackedArray self test -------------------------------------------------------\n");
printf("\n");
printf("sizeof(PackedArray) = %d\n", (int)sizeof(PackedArray));
printf("\n");
printf("1 by 1 packing / unpacking:\n");
for (bitsPerItem = 1; bitsPerItem <= 32; ++bitsPerItem)
{
uint32_t mask = (uint32_t)(1ULL << bitsPerItem) - 1;
int count;
for (count = 1; count <= 1024; ++count)
{
PackedArray* a1 = PackedArray_create(bitsPerItem, count);
PackedArray* a2 = PackedArray_create(bitsPerItem, count);
int i;
assert(a1->count == (uint32_t)count);
assert(a2->count == (uint32_t)count);
assert(a1->bitsPerItem == bitsPerItem);
assert(a2->bitsPerItem == bitsPerItem);
assert(PackedArray_bufferSize(a1) == PackedArray_bufferSize(a2));
for (i = 0; i < (int)PackedArray_bufferSize(a1); ++i)
a1->buffer[i] = a2->buffer[i] = rand();
for (i = 0; i < count; ++i)
{
uint32_t v1, v2;
v1 = rand() & mask;
v2 = v1 + 1;
PackedArray_pack(a1, i, &v1, 1);
PackedArray_pack_reference(a2, i, &v1, 1);
assert(memcmp(a1->buffer, a2->buffer, sizeof(a1->buffer[0]) * PackedArray_bufferSize(a1)) == 0);
PackedArray_set(a1, i, v1);
assert(memcmp(a1->buffer, a2->buffer, sizeof(a1->buffer[0]) * PackedArray_bufferSize(a1)) == 0);
PackedArray_unpack(a1, i, &v2, 1);
assert(v1 == v2);
PackedArray_unpack_reference(a2, i, &v2, 1);
assert(v1 == v2);
v2 = PackedArray_get(a2, i);
assert(v1 == v2);
}
for (i = count - 1; i >= 0; --i)
{
uint32_t v1, v2;
v1 = rand() & mask;
v2 = v1 + 1;
PackedArray_pack(a1, i, &v1, 1);
PackedArray_pack_reference(a2, i, &v1, 1);
assert(memcmp(a1->buffer, a2->buffer, sizeof(a1->buffer[0]) * PackedArray_bufferSize(a1)) == 0);
PackedArray_set(a1, i, v1);
assert(memcmp(a1->buffer, a2->buffer, sizeof(a1->buffer[0]) * PackedArray_bufferSize(a1)) == 0);
PackedArray_unpack(a1, i, &v2, 1);
assert(v1 == v2);
PackedArray_unpack_reference(a2, i, &v2, 1);
assert(v1 == v2);
v2 = PackedArray_get(a2, i);
assert(v1 == v2);
}
PackedArray_destroy(a1);
PackedArray_destroy(a2);
}
printf(" %2d bits per item -- success.\n", bitsPerItem);
}
printf("\n");
printf("bulk packing / unpacking:\n");
for (bitsPerItem = 1; bitsPerItem <= 32; ++bitsPerItem)
{
uint32_t mask = (uint32_t)(1ULL << bitsPerItem) - 1;
int count;
for (count = 1; count <= 128; ++count)
{
uint32_t* v1;
uint32_t* v2;
PackedArray* a1;
PackedArray* a2;
int i, j;
v1 = (uint32_t*)malloc(sizeof(uint32_t) * count);
assert(v1 != NULL);
v2 = (uint32_t*)malloc(sizeof(uint32_t) * count);
assert(v2 != NULL);
a1 = PackedArray_create(bitsPerItem, count);
assert(a1 != NULL);
a2 = PackedArray_create(bitsPerItem, count);
assert(a2 != NULL);
for (i = 0; i < (int)PackedArray_bufferSize(a1); ++i)
a1->buffer[i] = a2->buffer[i] = rand();
for (i = 0; i < count; ++i)
v1[i] = rand() & mask;
assert(bitsPerItem >= PackedArray_computeBitsPerItem(v1, count));
for (i = 0; i < count; ++i)
{
for (j = 1; j <= count - i; ++j)
{
PackedArray_pack(a1, i, v1, j);
PackedArray_pack_reference(a2, i, v1, j);
assert(memcmp(a1->buffer, a2->buffer, sizeof(a1->buffer[0]) * PackedArray_bufferSize(a1)) == 0);
PackedArray_unpack(a1, i, v2, j);
assert(bitsPerItem >= PackedArray_computeBitsPerItem(v2, j));
assert(memcmp(v1, v2, j * sizeof(uint32_t)) == 0);
PackedArray_unpack_reference(a2, i, v2, j);
assert(memcmp(v1, v2, j * sizeof(uint32_t)) == 0);
}
}
PackedArray_destroy(a1);
PackedArray_destroy(a2);
free(v1);
free(v2);
}
printf(" %2d bits per item -- success.\n", bitsPerItem);
}
return 0;
}
#elif defined(PACKEDARRAY_SELF_BENCH) // #if defined(PACKEDARRAY_SELF_TEST)
#ifndef NDEBUG
#error please define NDEBUG to inhibit asserts when compiling the benchmark
#endif
#include <stdio.h>
#include <string.h>
#include <float.h>
#ifndef MIN
#define MIN(x, y) (((x) < (y)) ? (x) : (y))
#endif
#ifndef MAX
#define MAX(x, y) (((x) < (y)) ? (y) : (x))
#endif
#ifdef _MSC_VER
#pragma warning(push, 3)
#include <windows.h>
#pragma warning(pop)
static double getChronometerTime(void)
{
LARGE_INTEGER frequency;
LARGE_INTEGER t;
QueryPerformanceFrequency(&frequency);
QueryPerformanceCounter(&t);
return (double)t.QuadPart / (double)frequency.QuadPart * 1000;
}
#else
#include <sys/time.h>
static double getChronometerTime()
{
struct timeval now = { 0 };
gettimeofday(&now, NULL);
return (double)now.tv_sec + (double)now.tv_usec * 1e-6;
}
#endif
#define LOOP_COUNT 1000
static double bench_memcpy(uint32_t* in, uint32_t* out, uint32_t count)
{
double start, end;
uint32_t i;
start = getChronometerTime();
for (i = 0; i < LOOP_COUNT; ++i)
memcpy(out, in, count * sizeof(uint32_t));
end = getChronometerTime();
return 1e6 * (end - start) / LOOP_COUNT;
}
static double bench_loopcpy(uint32_t* in, uint32_t* out, uint32_t count)
{
double start, end;
uint32_t i;
start = getChronometerTime();
for (i = 0; i < LOOP_COUNT; ++i)
{
uint32_t j;
for (j = 0; j < count; ++j)
out[j] = in[j];
}
end = getChronometerTime();
return 1e6 * (end - start) / LOOP_COUNT;
}
static double bench_pack(uint32_t* in, PackedArray* out, uint32_t count)
{
double start, end;
int i;
start = getChronometerTime();
for (i = 0; i < LOOP_COUNT; ++i)
PackedArray_pack(out, 0, in, count);
end = getChronometerTime();
return 1e6 * (end - start) / LOOP_COUNT;
}
static double bench_unpack(PackedArray* in, uint32_t* out, uint32_t count)
{
double start, end;
int i;
start = getChronometerTime();
for (i = 0; i < LOOP_COUNT; ++i)
PackedArray_unpack(in, 0, out, count);
end = getChronometerTime();
return 1e6 * (end - start) / LOOP_COUNT;
}
#define MAX_ELEMENT_COUNT (1 << 18)
#define LOG2_MAX_ELEMENT_COUNT 18
int main(void)
{
double start, end;
uint32_t* b1;
uint32_t* b2;
uint32_t count, bitsPerItem;
PackedArray** packed;
uint32_t i;
double* speed_memcpy;
double avg_memcpy, min_memcpy, max_memcpy;
double* speed_loopcpy;
double avg_loopcpy, min_loopcpy, max_loopcpy;
double* speed_pack[32];
double avg_pack, min_pack, max_pack;
double* speed_unpack[32];
double avg_unpack, min_unpack, max_unpack;
printf("-- PackedArray self bench ------------------------------------------------------\n");
start = getChronometerTime();
b1 = (uint32_t*)malloc(sizeof(uint32_t) * MAX_ELEMENT_COUNT);
assert(b1 != NULL);
b2 = (uint32_t*)malloc(sizeof(uint32_t) * MAX_ELEMENT_COUNT);
assert(b2 != NULL);
packed = (PackedArray**)malloc(sizeof(PackedArray*) * 32);
assert(packed != NULL);
for (i = 0; i < 32; ++i)
packed[i] = PackedArray_create(i + 1, MAX_ELEMENT_COUNT);
for (i = 0; i < MAX_ELEMENT_COUNT; ++i)
b1[i] = rand();
speed_memcpy = (double*)malloc(sizeof(double) * (LOG2_MAX_ELEMENT_COUNT + 1));
assert(speed_memcpy != NULL);
avg_memcpy = 0;
min_memcpy = DBL_MAX;
max_memcpy = 0;
printf("memcpy:\n");
printf("bits\tsize (B)\ttime (µs)\tspeed (B/µs)\n");
for (count = 1, i = 0; count <= MAX_ELEMENT_COUNT; count *= 2, ++i)
{
double elapsed = bench_memcpy(b1, b2, count);
double speed = count * sizeof(uint32_t) / elapsed;
printf("%4d\t%8d\t%9.3f\t%12.3f\n", 32, (uint32_t)(count * sizeof(uint32_t)), elapsed, speed);
avg_memcpy += speed;
min_memcpy = MIN(min_memcpy, speed);
max_memcpy = MAX(max_memcpy, speed);
speed_memcpy[i] = speed;
}
avg_memcpy /= i;
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)\n");
printf("%10.3f\t%10.3f\t%10.3f\n", avg_memcpy, min_memcpy, max_memcpy);
printf("\n");
speed_loopcpy = (double*)malloc(sizeof(double) * (LOG2_MAX_ELEMENT_COUNT + 1));
assert(speed_loopcpy != NULL);
avg_loopcpy = 0;
min_loopcpy = DBL_MAX;
max_loopcpy = 0;
printf("loopcpy:\n");
printf("bits\tsize (B)\ttime (µs)\tspeed (B/µs)\n");
for (count = 1, i = 0; count <= MAX_ELEMENT_COUNT; count *= 2, ++i)
{
double elapsed = bench_loopcpy(b1, b2, count);
double speed = count * sizeof(uint32_t) / elapsed;
printf("%4d\t%8d\t%9.3f\t%12.3f\n", 32, (uint32_t)(count * sizeof(uint32_t)), elapsed, speed);
avg_loopcpy += speed;
min_loopcpy = MIN(min_loopcpy, speed);
max_loopcpy = MAX(max_loopcpy, speed);
speed_loopcpy[i] = speed;
}
avg_loopcpy /= i;
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)\n");
printf("%10.3f\t%10.3f\t%10.3f\n", avg_loopcpy, min_loopcpy, max_loopcpy);
printf("\n");
for (bitsPerItem = 1; bitsPerItem <= 32; ++bitsPerItem)
{
avg_pack = 0;
min_pack = DBL_MAX;
max_pack = 0;
avg_unpack = 0;
min_unpack = DBL_MAX;
max_unpack = 0;
printf("pack:\t \t \t \t");
printf("unpack:\t \t \t \t");
printf("\n");
printf("bits\tsize (B)\ttime (µs)\tspeed (B/µs)");
printf("\t");
printf("bits\tsize (B)\ttime (µs)\tspeed (B/µs)");
printf("\n");
speed_pack[bitsPerItem - 1] = (double*)malloc(sizeof(double) * (LOG2_MAX_ELEMENT_COUNT + 1));
assert(speed_pack[bitsPerItem - 1] != NULL);
speed_unpack[bitsPerItem - 1] = (double*)malloc(sizeof(double) * (LOG2_MAX_ELEMENT_COUNT + 1));
assert(speed_unpack[bitsPerItem - 1] != NULL);
for (count = 1, i = 0; count <= MAX_ELEMENT_COUNT; count *= 2, ++i)
{
uint32_t mask = (uint32_t)(1ULL << bitsPerItem) - 1;
uint32_t j;
double elapsed, speed;
for (j = 0; j < count; ++j)
b2[j] = b1[j] & mask;
elapsed = bench_pack(b2, packed[bitsPerItem - 1], count);
speed = count * sizeof(uint32_t) / elapsed;
printf("%4d\t%8d\t%9.3f\t%12.3f", bitsPerItem, (uint32_t)(count * sizeof(uint32_t)), elapsed, speed);
avg_pack += speed;
min_pack = MIN(min_pack, speed);
max_pack = MAX(max_pack, speed);
speed_pack[bitsPerItem - 1][i] = speed;
printf("\t");
elapsed = bench_unpack(packed[bitsPerItem - 1], b2, count);
speed = count * sizeof(uint32_t) / elapsed;
printf("%4d\t%8d\t%9.3f\t%12.3f", bitsPerItem, (uint32_t)(count * sizeof(uint32_t)), elapsed, speed);
avg_unpack += speed;
min_unpack = MIN(min_unpack, speed);
max_unpack = MAX(max_unpack, speed);
speed_unpack[bitsPerItem - 1][i] = speed;
printf("\n");
}
assert(i == LOG2_MAX_ELEMENT_COUNT + 1);
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\t\t");
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\n");
avg_pack /= i;
printf("%10.3f\t%10.3f\t%10.3f", avg_pack, min_pack, max_pack);
printf("\t\t");
avg_unpack /= i;
printf("%10.3f\t%10.3f\t%10.3f", avg_unpack, min_unpack, max_unpack);
printf("\n");
printf("\n");
}
printf("\n");
printf("stats by bits per item\n");
printf("pack:\t \t \t \t");
printf("unpack:\t \t \t \t");
printf("\n");
printf("bits\tavg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\t");
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\n");
for (bitsPerItem = 1; bitsPerItem <= 32; ++bitsPerItem)
{
avg_pack = 0;
min_pack = DBL_MAX;
max_pack = 0;
avg_unpack = 0;
min_unpack = DBL_MAX;
max_unpack = 0;
for (count = 1, i = 0; count <= MAX_ELEMENT_COUNT; count *= 2, ++i)
{
double speed;
speed = speed_pack[bitsPerItem - 1][i];
avg_pack += speed;
min_pack = MIN(min_pack, speed);
max_pack = MAX(max_pack, speed);
speed = speed_unpack[bitsPerItem - 1][i];
avg_unpack += speed;
min_unpack = MIN(min_unpack, speed);
max_unpack = MAX(max_unpack, speed);
}
assert(i == LOG2_MAX_ELEMENT_COUNT + 1);
printf("%4d\t", bitsPerItem);
avg_pack /= i;
printf("%10.3f\t%10.3f\t%10.3f", avg_pack, min_pack, max_pack);
printf("\t");
avg_unpack /= i;
printf("%10.3f\t%10.3f\t%10.3f", avg_unpack, min_unpack, max_unpack);
printf("\n");
}
printf("\n");
printf("stats by size\n");
printf("pack:\t \t \t \t");
printf("unpack:\t \t \t \t");
printf("\n");
printf("size(B)\tavg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\t");
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\n");
for (count = 1, i = 0; count <= MAX_ELEMENT_COUNT; count *= 2, ++i)
{
avg_pack = 0;
min_pack = DBL_MAX;
max_pack = 0;
avg_unpack = 0;
min_unpack = DBL_MAX;
max_unpack = 0;
for (bitsPerItem = 1; bitsPerItem <= 32; ++bitsPerItem)
{
double speed;
speed = speed_pack[bitsPerItem - 1][i];
avg_pack += speed;
min_pack = MIN(min_pack, speed);
max_pack = MAX(max_pack, speed);
speed = speed_unpack[bitsPerItem - 1][i];
avg_unpack += speed;
min_unpack = MIN(min_unpack, speed);
max_unpack = MAX(max_unpack, speed);
}
printf("%7d\t", (uint32_t)sizeof(uint32_t) * count);
avg_pack /= 32;
printf("%10.3f\t%10.3f\t%10.3f", avg_pack, min_pack, max_pack);
printf("\t");
avg_unpack /= 32;
printf("%10.3f\t%10.3f\t%10.3f", avg_unpack, min_unpack, max_unpack);
printf("\n");
}
printf("\n");
free(b1);
free(b2);
free(speed_memcpy);
free(speed_loopcpy);
for (i = 0; i < 32; ++i)
{
PackedArray_destroy(packed[i]);
free(speed_pack[i]);
free(speed_unpack[i]);
}
free(packed);
end = getChronometerTime();
printf("total time (s): %f\n", (end - start));
printf("\n");
return 0;
}
#endif // #elif defined(PACKEDARRAY_SELF_BENCH)
#endif // #ifdef PACKEDARRAY_IMPL
================================================
FILE: PackedArray.h
================================================
#ifndef PACKEDARRAY_H
#define PACKEDARRAY_H
#ifdef __cplusplus
extern "C" {
#endif
#include <stdint.h>
/*
PackedArray principle:
. compact storage of <= 32 bits items
. items are tightly packed into a buffer of uint32_t integers
PackedArray requirements:
. you must know in advance how many bits are needed to hold a single item
. you must know in advance how many items you want to store
. when packing, behavior is undefined if items have more than bitsPerItem bits
PackedArray general in memory representation:
|-------------------------------------------------- - - -
| b0 | b1 | b2 |
|-------------------------------------------------- - - -
| i0 | i1 | i2 | i3 | i4 | i5 | i6 | i7 | i8 | i9 |
|-------------------------------------------------- - - -
. items are tightly packed together
. several items end up inside the same buffer cell, e.g. i0, i1, i2
. some items span two buffer cells, e.g. i3, i6
*/
struct _PackedArray
{
uint32_t bitsPerItem;
uint32_t count;
uint32_t padding[2];
#ifdef _MSC_VER
#pragma warning(push)
#pragma warning(disable: 4200)
#endif // #ifdef _MSC_VER
uint32_t buffer[];
#ifdef _MSC_VER
#pragma warning(pop)
#endif // #ifdef _MSC_VER
};
typedef struct _PackedArray PackedArray;
// creation / destruction
PackedArray* PackedArray_create(uint32_t bitsPerItem, uint32_t count);
void PackedArray_destroy(PackedArray* a);
// packing / unpacking
// offset is expressed in number of elements
void PackedArray_pack(PackedArray* a, const uint32_t offset, const uint32_t* in, uint32_t count);
void PackedArray_unpack(const PackedArray* a, const uint32_t offset, uint32_t* out, uint32_t count);
// single item access
void PackedArray_set(PackedArray* a, const uint32_t offset, const uint32_t in);
uint32_t PackedArray_get(const PackedArray* a, const uint32_t offset);
// helpers
uint32_t PackedArray_bufferSize(const PackedArray* a);
uint32_t PackedArray_computeBitsPerItem(const uint32_t* in, uint32_t count);
#ifdef __cplusplus
}
#endif
#endif // #ifndef PACKEDARRAY_H
================================================
FILE: PackedArraySIMD.c
================================================
// see README.md for usage instructions.
// (‑●‑●)> released under the WTFPL v2 license, by Gregory Pakosz (@gpakosz)
#ifndef PACKEDARRAY_SELF
#define PACKEDARRAY_SELF "PackedArraySIMD.c"
#endif
#ifdef PACKEDARRAY_IMPL
#ifndef PACKEDARRAY_JOIN
#define PACKEDARRAY_JOIN(lhs, rhs) PACKEDARRAY_JOIN_(lhs, rhs)
#define PACKEDARRAY_JOIN_(lhs, rhs) PACKEDARRAY_JOIN__(lhs, rhs)
#define PACKEDARRAY_JOIN__(lhs, rhs) lhs##rhs
#endif // #ifndef PACKEDARRAY_JOIN
#ifndef PACKEDARRAY_IMPL_BITS_PER_ITEM
#error PACKEDARRAY_IMPL_BITS_PER_ITEM undefined
#endif // #ifndef PACKEDARRAY_IMPL_BITS_PER_ITEM
#if defined(PACKEDARRAY_IMPL_PACK_CASES) || defined(PACKEDARRAY_IMPL_UNPACK_CASES)
#ifndef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 0
#elif PACKEDARRAY_IMPL_CASE_I == 0
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 1
#elif PACKEDARRAY_IMPL_CASE_I == 1
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 2
#elif PACKEDARRAY_IMPL_CASE_I == 2
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 3
#elif PACKEDARRAY_IMPL_CASE_I == 3
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 4
#elif PACKEDARRAY_IMPL_CASE_I == 4
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 5
#elif PACKEDARRAY_IMPL_CASE_I == 5
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 6
#elif PACKEDARRAY_IMPL_CASE_I == 6
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 7
#elif PACKEDARRAY_IMPL_CASE_I == 7
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 8
#elif PACKEDARRAY_IMPL_CASE_I == 8
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 9
#elif PACKEDARRAY_IMPL_CASE_I == 9
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 10
#elif PACKEDARRAY_IMPL_CASE_I == 10
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 11
#elif PACKEDARRAY_IMPL_CASE_I == 11
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 12
#elif PACKEDARRAY_IMPL_CASE_I == 12
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 13
#elif PACKEDARRAY_IMPL_CASE_I == 13
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 14
#elif PACKEDARRAY_IMPL_CASE_I == 14
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 15
#elif PACKEDARRAY_IMPL_CASE_I == 15
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 16
#elif PACKEDARRAY_IMPL_CASE_I == 16
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 17
#elif PACKEDARRAY_IMPL_CASE_I == 17
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 18
#elif PACKEDARRAY_IMPL_CASE_I == 18
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 19
#elif PACKEDARRAY_IMPL_CASE_I == 19
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 20
#elif PACKEDARRAY_IMPL_CASE_I == 20
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 21
#elif PACKEDARRAY_IMPL_CASE_I == 21
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 22
#elif PACKEDARRAY_IMPL_CASE_I == 22
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 23
#elif PACKEDARRAY_IMPL_CASE_I == 23
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 24
#elif PACKEDARRAY_IMPL_CASE_I == 24
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 25
#elif PACKEDARRAY_IMPL_CASE_I == 25
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 26
#elif PACKEDARRAY_IMPL_CASE_I == 26
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 27
#elif PACKEDARRAY_IMPL_CASE_I == 27
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 28
#elif PACKEDARRAY_IMPL_CASE_I == 28
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 29
#elif PACKEDARRAY_IMPL_CASE_I == 29
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 30
#elif PACKEDARRAY_IMPL_CASE_I == 30
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 31
#elif PACKEDARRAY_IMPL_CASE_I == 31
#undef PACKEDARRAY_IMPL_CASE_I
#define PACKEDARRAY_IMPL_CASE_I 32
#endif // #ifndef PACKEDARRAY_IMPL_CASE_I
#ifndef PACKEDARRAY_IMPL_BITS_AVAILABLE
#define PACKEDARRAY_IMPL_BITS_AVAILABLE (32 - ((PACKEDARRAY_IMPL_CASE_I * PACKEDARRAY_IMPL_BITS_PER_ITEM) % 32))
#endif
#ifndef PACKEDARRAY_IMPL_START_BIT
#define PACKEDARRAY_IMPL_START_BIT ((PACKEDARRAY_IMPL_CASE_I * PACKEDARRAY_IMPL_BITS_PER_ITEM) % 32)
#endif
#if defined(PACKEDARRAY_IMPL_PACK_CASES)
#ifndef PACKEDARRAY_IMPL_PACK_CASE_BREAK
#define PACKEDARRAY_IMPL_PACK_CASE_BREAK
#endif
case PACKEDARRAY_IMPL_CASE_I:
#if (PACKEDARRAY_IMPL_BITS_PER_ITEM <= PACKEDARRAY_IMPL_BITS_AVAILABLE)
in_4 = PackedArray_loadu_uint32x4(in);
packed = PackedArray_vsli0_uint32x4(packed, in_4, PACKEDARRAY_IMPL_START_BIT);
in += 4;
#if (PACKEDARRAY_IMPL_BITS_PER_ITEM == PACKEDARRAY_IMPL_BITS_AVAILABLE)
PackedArray_store_uint32x4(out, packed);
out += 4;
packed = PackedArray_uint32x4_zero;
#endif
#else
in_4 = PackedArray_loadu_uint32x4(in);
packed = PackedArray_vsli0_uint32x4(packed, in_4, PACKEDARRAY_IMPL_START_BIT);
PackedArray_store_uint32x4(out, packed);
out += 4;
packed = PackedArray_shr_uint32x4(in_4, PACKEDARRAY_IMPL_BITS_AVAILABLE);
in += 4;
#endif
PACKEDARRAY_IMPL_PACK_CASE_BREAK
#if PACKEDARRAY_IMPL_CASE_I < 31
#include PACKEDARRAY_SELF
#else
#undef PACKEDARRAY_IMPL_CASE_I
#undef PACKEDARRAY_IMPL_PACK_CASE_BREAK
#undef PACKEDARRAY_IMPL_PACK_CASES
#endif
#elif defined(PACKEDARRAY_IMPL_UNPACK_CASES) // #if defined(PACKEDARRAY_IMPL_PACK_CASES)
#ifndef PACKEDARRAY_IMPL_UNPACK_CASE_BREAK
#define PACKEDARRAY_IMPL_UNPACK_CASE_BREAK
#endif
case PACKEDARRAY_IMPL_CASE_I:
#if (PACKEDARRAY_IMPL_BITS_PER_ITEM <= PACKEDARRAY_IMPL_BITS_AVAILABLE)
out_4 = PackedArray_and_uint32x4(PackedArray_shr_uint32x4(packed, PACKEDARRAY_IMPL_START_BIT), PackedArray_set_uint32x4(PACKEDARRAY_IMPL_MASK));
PackedArray_storeu_uint32x4(out, out_4);
out += 4;
PACKEDARRAY_IMPL_UNPACK_CASE_BREAK
#if (PACKEDARRAY_IMPL_CASE_I < 31) && (PACKEDARRAY_IMPL_BITS_PER_ITEM == PACKEDARRAY_IMPL_BITS_AVAILABLE)
in += 4;
packed = PackedArray_load_uint32x4(in);
#endif
#else
out_4 = PackedArray_shr_uint32x4(packed, PACKEDARRAY_IMPL_START_BIT);
in += 4;
packed = PackedArray_load_uint32x4(in);
out_4 = PackedArray_vsli0_uint32x4(out_4, packed, PACKEDARRAY_IMPL_BITS_AVAILABLE);
out_4 = PackedArray_and_uint32x4(out_4, PackedArray_set_uint32x4(PACKEDARRAY_IMPL_MASK));
PackedArray_storeu_uint32x4(out, out_4);
out += 4;
PACKEDARRAY_IMPL_UNPACK_CASE_BREAK
#endif
#if PACKEDARRAY_IMPL_CASE_I < 31
#include PACKEDARRAY_SELF
#else
#undef PACKEDARRAY_IMPL_CASE_I
#undef PACKEDARRAY_IMPL_UNPACK_CASE_BREAK
#undef PACKEDARRAY_IMPL_UNPACK_CASES
#endif
#endif // #elif defined(PACKEDARRAY_IMPL_UNPACK_CASES)
#else // #if defined(PACKEDARRAY_IMPL_PACK_CASES) || defined(PACKEDARRAY_IMPL_UNPACK_CASES)
#ifndef PACKEDARRAY_IMPL_MASK
#define PACKEDARRAY_IMPL_MASK (uint32_t)((1ULL << PACKEDARRAY_IMPL_BITS_PER_ITEM) - 1)
#endif
void PACKEDARRAY_JOIN(__PackedArray_pack_, PACKEDARRAY_IMPL_BITS_PER_ITEM)(uint32_t* __restrict buffer, uint32_t offset, const uint32_t* __restrict in, uint32_t count)
{
uint32_t pre, post;
uint32_t* __restrict out;
const uint32_t* __restrict end;
uint32_t startBit;
PackedArray_uint32x4_t packed, in_4, mask;
uint32_t offset_4;
pre = (offset + 3) / 4 * 4 - offset;
pre = pre > count ? count : pre;
if (pre > 0)
{
__PackedArray_pack_scalar(buffer, PACKEDARRAY_IMPL_BITS_PER_ITEM, PACKEDARRAY_IMPL_MASK, offset, in, pre);
offset += pre;
in += pre;
count -= pre;
}
post = count % 4;
count -= post;
if (count > 0)
{
out = &buffer[(offset / 4 * PACKEDARRAY_IMPL_BITS_PER_ITEM) / 32 * 4];
startBit = (offset / 4 * PACKEDARRAY_IMPL_BITS_PER_ITEM) % 32;
packed = PackedArray_load_uint32x4(out);
mask = PackedArray_sub_uint32x4(PackedArray_shl_uint32x4(PackedArray_set_uint32x4(1), startBit), PackedArray_set_uint32x4(1));
packed = PackedArray_and_uint32x4(packed, mask);
offset_4 = offset % 128;
offset += count;
if (count >= 128 - offset_4)
{
int32_t n;
n = (count + offset_4) / 128;
count -= 128 * n - offset_4;
switch (offset_4 / 4)
{
do
{
#define PACKEDARRAY_IMPL_PACK_CASES
#include PACKEDARRAY_SELF
} while (--n > 0);
}
if (count == 0)
goto PACKEDARRAY_JOIN(PACKEDARRAY_JOIN(__PackedArray_pack_, PACKEDARRAY_IMPL_BITS_PER_ITEM), _post);
offset_4 = 0;
startBit = 0;
}
end = in + count;
switch (offset_4 / 4)
{
#define PACKEDARRAY_IMPL_PACK_CASES
#define PACKEDARRAY_IMPL_PACK_CASE_BREAK \
if (in == end)\
break;
#include PACKEDARRAY_SELF
}
PACKEDARRAY_ASSERT(in == end);
if ((count / 4 * PACKEDARRAY_IMPL_BITS_PER_ITEM + startBit) % 32)
{
in_4 = PackedArray_loadu_uint32x4(out);
mask = PackedArray_sub_uint32x4(PackedArray_shl_uint32x4(PackedArray_set_uint32x4(1), ((count / 4 * PACKEDARRAY_IMPL_BITS_PER_ITEM + startBit - 1) % 32) + 1), PackedArray_set_uint32x4(1));
in_4 = PackedArray_andnot_uint32x4(in_4, mask);
packed = PackedArray_or_uint32x4(packed, in_4);
PackedArray_store_uint32x4(out, packed);
}
}
PACKEDARRAY_JOIN(PACKEDARRAY_JOIN(__PackedArray_pack_, PACKEDARRAY_IMPL_BITS_PER_ITEM), _post):
if (post > 0)
__PackedArray_pack_scalar(buffer, PACKEDARRAY_IMPL_BITS_PER_ITEM, PACKEDARRAY_IMPL_MASK, offset, in, post);
}
void PACKEDARRAY_JOIN(__PackedArray_unpack_, PACKEDARRAY_IMPL_BITS_PER_ITEM)(const uint32_t* __restrict buffer, uint32_t offset, uint32_t* __restrict out, uint32_t count)
{
uint32_t pre, post;
const uint32_t* __restrict in;
const uint32_t* __restrict end;
PackedArray_uint32x4_t packed, out_4;
uint32_t offset_4;
pre = (offset + 3) / 4 * 4 - offset;
pre = pre > count ? count : pre;
if (pre > 0)
{
__PackedArray_unpack_scalar(buffer, PACKEDARRAY_IMPL_BITS_PER_ITEM, PACKEDARRAY_IMPL_MASK, offset, out, pre);
offset += pre;
out += pre;
count -= pre;
}
post = count % 4;
count -= post;
if (count > 0)
{
in = &buffer[(offset / 4 * PACKEDARRAY_IMPL_BITS_PER_ITEM) / 32 * 4];
packed = PackedArray_load_uint32x4(in);
offset_4 = offset % 128;
offset += count;
if (count >= 128 - offset_4)
{
int32_t n;
n = (count + offset_4) / 128;
count -= 128 * n - offset_4;
switch (offset_4 / 4)
{
do
{
in += 4;
packed = PackedArray_load_uint32x4(in);
#define PACKEDARRAY_IMPL_UNPACK_CASES
#include PACKEDARRAY_SELF
} while (--n > 0);
}
if (count == 0)
goto PACKEDARRAY_JOIN(PACKEDARRAY_JOIN(__PackedArray_unpack_, PACKEDARRAY_IMPL_BITS_PER_ITEM), _post);
in += 4;
packed = PackedArray_loadu_uint32x4(in);
offset_4 = 0;
}
end = out + count;
switch (offset_4 / 4)
{
#define PACKEDARRAY_IMPL_UNPACK_CASES
#define PACKEDARRAY_IMPL_UNPACK_CASE_BREAK \
if (out == end)\
break;
#include PACKEDARRAY_SELF
}
PACKEDARRAY_ASSERT(out == end);
}
PACKEDARRAY_JOIN(PACKEDARRAY_JOIN(__PackedArray_unpack_, PACKEDARRAY_IMPL_BITS_PER_ITEM), _post):
if (post > 0)
__PackedArray_unpack_scalar(buffer, PACKEDARRAY_IMPL_BITS_PER_ITEM, PACKEDARRAY_IMPL_MASK, offset, out, post);
}
#undef PACKEDARRAY_IMPL_BITS_PER_ITEM
#undef PACKEDARRAY_IMPL_BITS_AVAILABLE
#undef PACKEDARRAY_IMPL_START_BIT
#undef PACKEDARRAY_IMPL_START_MASK
#endif // #if defined(PACKEDARRAY_IMPL_PACK_CASES) || defined(PACKEDARRAY_IMPL_UNPACK_CASES)
#else
#include "PackedArray.h"
#if !defined(PACKEDARRAY_ASSERT)
#include <assert.h>
#define PACKEDARRAY_ASSERT(expression) assert(expression)
#endif
#include <stddef.h>
static void __PackedArray_pack_scalar(uint32_t* buffer, const uint32_t bitsPerItem, const uint32_t mask, uint32_t offset, const uint32_t* in, uint32_t count)
{
uint32_t* __restrict out;
uint32_t startBit;
uint32_t bitsAvailable;
PACKEDARRAY_ASSERT(buffer != NULL);
PACKEDARRAY_ASSERT(in != NULL);
PACKEDARRAY_ASSERT(count != 0);
while (count--)
{
uint32_t value = *in++;
PACKEDARRAY_ASSERT(0 == (~mask & value));
out = &buffer[((uint64_t)offset / 4 * (uint64_t)bitsPerItem) / 32 * 4 + offset % 4];
startBit = ((uint64_t)offset / 4 * (uint64_t)bitsPerItem) % 32;
++offset;
bitsAvailable = 32 - startBit;
if (bitsPerItem <= bitsAvailable)
{
out[0] = (out[0] & ~(mask << startBit)) | (value << startBit);
}
else
{
// value spans 2 buffer cells
uint32_t low, high;
low = value << startBit;
high = value >> bitsAvailable;
out[0] = (out[0] & ~(mask << startBit)) | low;
out[4] = (out[4] & ~(mask >> bitsAvailable)) | high;
}
}
}
static void __PackedArray_unpack_scalar(const uint32_t* buffer, const uint32_t bitsPerItem, const uint32_t mask, uint32_t offset, uint32_t* out, uint32_t count)
{
const uint32_t* __restrict in;
uint32_t startBit;
uint32_t bitsAvailable;
PACKEDARRAY_ASSERT(buffer != NULL);
PACKEDARRAY_ASSERT(out != NULL);
PACKEDARRAY_ASSERT(count != 0);
while (count--)
{
uint32_t value;
in = &buffer[((uint64_t)offset / 4 * (uint64_t)bitsPerItem) / 32 * 4 + offset % 4];
startBit = ((uint64_t)offset / 4 * (uint64_t)bitsPerItem) % 32;
++offset;
bitsAvailable = 32 - startBit;
if (bitsPerItem <= bitsAvailable)
{
value = (in[0] >> startBit) & mask;
}
else
{
// value spans 2 buffer cells
uint32_t low, high;
low = in[0] >> startBit;
high = in[4] << bitsAvailable;
value = (low | high) & mask;
}
*out++ = value;
}
}
#if defined(__SSE2__) || defined(_M_IX86) || defined(_M_X64)
#include <emmintrin.h>
#define PackedArray_uint32x4_t __m128i
#define PackedArray_uint32x4_zero _mm_setzero_si128()
#define PackedArray_set_uint32x4(i) _mm_set1_epi32(i)
#define PackedArray_sub_uint32x4(lhs, rhs) _mm_sub_epi32(lhs, rhs)
#define PackedArray_loadu_uint32x4(ptr) _mm_loadu_si128((const __m128i*)ptr)
#define PackedArray_storeu_uint32x4(ptr, v) _mm_storeu_si128((__m128i*)ptr, v)
#define PackedArray_load_uint32x4(ptr) _mm_load_si128((const __m128i*)ptr)
#define PackedArray_store_uint32x4(ptr, v) _mm_store_si128((__m128i*)ptr, v)
#define PackedArray_shl_uint32x4(v, shift) _mm_slli_epi32(v, shift)
#define PackedArray_shr_uint32x4(v, shift) _mm_srli_epi32(v, shift)
#define PackedArray_or_uint32x4(lhs, rhs) _mm_or_si128(lhs, rhs)
#define PackedArray_and_uint32x4(lhs, rhs) _mm_and_si128(lhs, rhs)
#define PackedArray_andnot_uint32x4(lhs, rhs) _mm_andnot_si128(rhs, lhs)
// assumes dst bits are cleared at inserted bit positions
#define PackedArray_vsli0_uint32x4(dst, src, shift) PackedArray_or_uint32x4(dst, PackedArray_shl_uint32x4(src, shift))
#elif defined(__ARM_NEON__) || defined(_M_ARM)
#include <arm_neon.h>
#define PackedArray_uint32x4_t uint32x4_t
#define PackedArray_uint32x4_zero vdupq_n_u32(0)
#define PackedArray_set_uint32x4(i) vdupq_n_u32(i)
#define PackedArray_sub_uint32x4(lhs, rhs) vsubq_u32(lhs, rhs)
#define PackedArray_loadu_uint32x4(ptr) vld1q_u32((const uint32_t*)ptr)
#define PackedArray_storeu_uint32x4(ptr, v) vst1q_u32(ptr, v)
#if defined(__GNUC__)
// because __builtin_assume_aligned isn't always available...
typedef uint32_t __attribute__((aligned(16))) PackedArray_aligned_uint32_t;
#define PackedArray_load_uint32x4(ptr) vld1q_u32((const PackedArray_aligned_uint32_t*)ptr)
#define PackedArray_store_uint32x4(ptr, v) vst1q_u32((PackedArray_aligned_uint32_t*)ptr, v)
#else
#define PackedArray_load_uint32x4(ptr) vld1q_u32((const uint32_t*)ptr)
#define PackedArray_store_uint32x4(ptr, v) vst1q_u32(ptr, v)
#endif
#define PackedArray_shl_uint32x4(v, shift) vshlq_u32(v, vdupq_n_s32(shift))
#define PackedArray_shr_uint32x4(v, shift) vshlq_u32(v, vdupq_n_s32(-shift))
#define PackedArray_or_uint32x4(lhs, rhs) vorrq_u32(lhs, rhs)
#define PackedArray_and_uint32x4(lhs, rhs) vandq_u32(lhs, rhs)
#define PackedArray_andnot_uint32x4(lhs, rhs) vbicq_u32(lhs, rhs)
// assumes dst bits are cleared at inserted bit positions
#define PackedArray_vsli0_uint32x4(dst, src, shift) vsliq_n_u32(dst, src, shift)
#else
#error unsupported SIMD platform
#endif
#define PACKEDARRAY_IMPL
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 1
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 2
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 3
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 4
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 5
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 6
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 7
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 8
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 9
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 10
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 11
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 12
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 13
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 14
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 15
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 16
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 17
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 18
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 19
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 20
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 21
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 22
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 23
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 24
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 25
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 26
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 27
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 28
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 29
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 30
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 31
#include PACKEDARRAY_SELF
#define PACKEDARRAY_IMPL_BITS_PER_ITEM 32
#include PACKEDARRAY_SELF
#undef PACKEDARRAY_IMPL
#if !defined(PACKEDARRAY_ALIGNED_MALLOC) || !defined(PACKEDARRAY_FREE)
#include <stdlib.h>
#endif
#if !defined(PACKEDARRAY_ALIGNED_MALLOC)
#if defined (_MSC_VER)
#define PACKEDARRAY_ALIGNED_MALLOC(alignment, size) _aligned_malloc(size, alignment)
#elif defined (ANDROID) || defined (__ANDROID__)
#define PACKEDARRAY_ALIGNED_MALLOC(alignment, size) memalign(alignment, size)
#else
static void* __PackedArray_aligned_malloc(size_t alignment, size_t size)
{
void* p = NULL;
posix_memalign(&p, alignment, size);
return p;
}
#define PACKEDARRAY_ALIGNED_MALLOC(alignment, size) __PackedArray_aligned_malloc(alignment, size)
#endif
#endif
#if !defined(PACKEDARRAY_FREE)
#if defined (_MSC_VER)
#define PACKEDARRAY_FREE(p) _aligned_free(p)
#else
#define PACKEDARRAY_FREE(p) free(p)
#endif
#endif
PackedArray* PackedArray_create(uint32_t bitsPerItem, uint32_t count)
{
PackedArray* a;
size_t bufferSize;
PACKEDARRAY_ASSERT(bitsPerItem > 0);
PACKEDARRAY_ASSERT(bitsPerItem <= 32);
bufferSize = sizeof(uint32_t) * (((uint64_t)count / 4 * (uint64_t)bitsPerItem + 31) / 32 * 4);
bufferSize += count < 4 ? sizeof(uint32_t) * count : sizeof(uint32_t) * 4;
a = (PackedArray*)PACKEDARRAY_ALIGNED_MALLOC(16, sizeof(PackedArray) + bufferSize);
PACKEDARRAY_ASSERT((uint64_t)a->buffer % 16 == 0);
if (a != NULL)
{
a->buffer[bufferSize / sizeof(uint32_t) - 1] = 0;
a->bitsPerItem = bitsPerItem;
a->count = count;
}
return a;
}
void PackedArray_destroy(PackedArray* a)
{
PACKEDARRAY_ASSERT(a);
PACKEDARRAY_FREE(a);
}
void PackedArray_pack(PackedArray* a, const uint32_t offset, const uint32_t* in, uint32_t count)
{
PACKEDARRAY_ASSERT(a != NULL);
PACKEDARRAY_ASSERT(in != NULL);
switch (a->bitsPerItem)
{
case 1: __PackedArray_pack_1(a->buffer, offset, in, count); break;
case 2: __PackedArray_pack_2(a->buffer, offset, in, count); break;
case 3: __PackedArray_pack_3(a->buffer, offset, in, count); break;
case 4: __PackedArray_pack_4(a->buffer, offset, in, count); break;
case 5: __PackedArray_pack_5(a->buffer, offset, in, count); break;
case 6: __PackedArray_pack_6(a->buffer, offset, in, count); break;
case 7: __PackedArray_pack_7(a->buffer, offset, in, count); break;
case 8: __PackedArray_pack_8(a->buffer, offset, in, count); break;
case 9: __PackedArray_pack_9(a->buffer, offset, in, count); break;
case 10: __PackedArray_pack_10(a->buffer, offset, in, count); break;
case 11: __PackedArray_pack_11(a->buffer, offset, in, count); break;
case 12: __PackedArray_pack_12(a->buffer, offset, in, count); break;
case 13: __PackedArray_pack_13(a->buffer, offset, in, count); break;
case 14: __PackedArray_pack_14(a->buffer, offset, in, count); break;
case 15: __PackedArray_pack_15(a->buffer, offset, in, count); break;
case 16: __PackedArray_pack_16(a->buffer, offset, in, count); break;
case 17: __PackedArray_pack_17(a->buffer, offset, in, count); break;
case 18: __PackedArray_pack_18(a->buffer, offset, in, count); break;
case 19: __PackedArray_pack_19(a->buffer, offset, in, count); break;
case 20: __PackedArray_pack_20(a->buffer, offset, in, count); break;
case 21: __PackedArray_pack_21(a->buffer, offset, in, count); break;
case 22: __PackedArray_pack_22(a->buffer, offset, in, count); break;
case 23: __PackedArray_pack_23(a->buffer, offset, in, count); break;
case 24: __PackedArray_pack_24(a->buffer, offset, in, count); break;
case 25: __PackedArray_pack_25(a->buffer, offset, in, count); break;
case 26: __PackedArray_pack_26(a->buffer, offset, in, count); break;
case 27: __PackedArray_pack_27(a->buffer, offset, in, count); break;
case 28: __PackedArray_pack_28(a->buffer, offset, in, count); break;
case 29: __PackedArray_pack_29(a->buffer, offset, in, count); break;
case 30: __PackedArray_pack_30(a->buffer, offset, in, count); break;
case 31: __PackedArray_pack_31(a->buffer, offset, in, count); break;
case 32: __PackedArray_pack_32(a->buffer, offset, in, count); break;
}
}
void PackedArray_unpack(const PackedArray* a, const uint32_t offset, uint32_t* out, uint32_t count)
{
PACKEDARRAY_ASSERT(a != NULL);
PACKEDARRAY_ASSERT(out != NULL);
switch (a->bitsPerItem)
{
case 1: __PackedArray_unpack_1(a->buffer, offset, out, count); break;
case 2: __PackedArray_unpack_2(a->buffer, offset, out, count); break;
case 3: __PackedArray_unpack_3(a->buffer, offset, out, count); break;
case 4: __PackedArray_unpack_4(a->buffer, offset, out, count); break;
case 5: __PackedArray_unpack_5(a->buffer, offset, out, count); break;
case 6: __PackedArray_unpack_6(a->buffer, offset, out, count); break;
case 7: __PackedArray_unpack_7(a->buffer, offset, out, count); break;
case 8: __PackedArray_unpack_8(a->buffer, offset, out, count); break;
case 9: __PackedArray_unpack_9(a->buffer, offset, out, count); break;
case 10: __PackedArray_unpack_10(a->buffer, offset, out, count); break;
case 11: __PackedArray_unpack_11(a->buffer, offset, out, count); break;
case 12: __PackedArray_unpack_12(a->buffer, offset, out, count); break;
case 13: __PackedArray_unpack_13(a->buffer, offset, out, count); break;
case 14: __PackedArray_unpack_14(a->buffer, offset, out, count); break;
case 15: __PackedArray_unpack_15(a->buffer, offset, out, count); break;
case 16: __PackedArray_unpack_16(a->buffer, offset, out, count); break;
case 17: __PackedArray_unpack_17(a->buffer, offset, out, count); break;
case 18: __PackedArray_unpack_18(a->buffer, offset, out, count); break;
case 19: __PackedArray_unpack_19(a->buffer, offset, out, count); break;
case 20: __PackedArray_unpack_20(a->buffer, offset, out, count); break;
case 21: __PackedArray_unpack_21(a->buffer, offset, out, count); break;
case 22: __PackedArray_unpack_22(a->buffer, offset, out, count); break;
case 23: __PackedArray_unpack_23(a->buffer, offset, out, count); break;
case 24: __PackedArray_unpack_24(a->buffer, offset, out, count); break;
case 25: __PackedArray_unpack_25(a->buffer, offset, out, count); break;
case 26: __PackedArray_unpack_26(a->buffer, offset, out, count); break;
case 27: __PackedArray_unpack_27(a->buffer, offset, out, count); break;
case 28: __PackedArray_unpack_28(a->buffer, offset, out, count); break;
case 29: __PackedArray_unpack_29(a->buffer, offset, out, count); break;
case 30: __PackedArray_unpack_30(a->buffer, offset, out, count); break;
case 31: __PackedArray_unpack_31(a->buffer, offset, out, count); break;
case 32: __PackedArray_unpack_32(a->buffer, offset, out, count); break;
}
}
void PackedArray_set(PackedArray* a, const uint32_t offset, const uint32_t in)
{
uint32_t* __restrict out;
uint32_t bitsPerItem;
uint32_t startBit;
uint32_t bitsAvailable;
uint32_t mask;
PACKEDARRAY_ASSERT(a != NULL);
bitsPerItem = a->bitsPerItem;
out = &a->buffer[4 * (((uint64_t)offset / 4 * (uint64_t)bitsPerItem) / 32) + (offset % 4)];
startBit = ((uint64_t)offset / 4 * (uint64_t)bitsPerItem) % 32;
bitsAvailable = 32 - startBit;
mask = (uint32_t)(1ULL << bitsPerItem) - 1;
PACKEDARRAY_ASSERT(0 == (~mask & in));
if (bitsPerItem <= bitsAvailable)
{
out[0] = (out[0] & ~(mask << startBit)) | (in << startBit);
}
else
{
// value spans 2 buffer cells
uint32_t low, high;
low = in << startBit;
high = in >> bitsAvailable;
out[0] = (out[0] & ~(mask << startBit)) | low;
out[4] = (out[4] & ~(mask >> (32 - startBit))) | high;
}
}
uint32_t PackedArray_get(const PackedArray* a, const uint32_t offset)
{
const uint32_t* __restrict in;
uint32_t bitsPerItem;
uint32_t startBit;
uint32_t bitsAvailable;
uint32_t mask;
uint32_t out;
PACKEDARRAY_ASSERT(a != NULL);
bitsPerItem = a->bitsPerItem;
in = &a->buffer[4 * (((uint64_t)offset / 4 * (uint64_t)bitsPerItem) / 32) + (offset % 4)];
startBit = ((uint64_t)offset / 4 * (uint64_t)bitsPerItem) % 32;
bitsAvailable = 32 - startBit;
mask = (uint32_t)(1ULL << bitsPerItem) - 1;
if (bitsPerItem <= bitsAvailable)
{
out = (in[0] >> startBit) & mask;
}
else
{
// value spans 2 buffer cells
uint32_t low, high;
low = in[0] >> startBit;
high = in[4] << (32 - startBit);
out = low ^ ((low ^ high) & (mask >> bitsAvailable << bitsAvailable));
}
return out;
}
uint32_t PackedArray_bufferSize(const PackedArray* a)
{
size_t bufferSize;
uint32_t bitsPerItem, count;
PACKEDARRAY_ASSERT(a != NULL);
bitsPerItem = a->bitsPerItem;
count = a->count;
bufferSize = ((uint64_t)count / 4 * (uint64_t)bitsPerItem + 31) / 32 * 4;
bufferSize += count < 4 ? count : 4;
return (uint32_t)bufferSize;
}
#if !(defined(_MSC_VER) && _MSC_VER >= 1400) && !defined(__GNUC__)
// log base 2 of an integer, aka the position of the highest bit set
static uint32_t __PackedArray_log2(uint32_t v)
{
// references
// http://aggregate.org/MAGIC
// http://graphics.stanford.edu/~seander/bithacks.html
static const uint32_t multiplyDeBruijnBitPosition[32] =
{
0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30,
8, 12, 20, 28, 15, 17, 24, 7, 19, 27, 23, 6, 26, 5, 4, 31
};
v |= v >> 1;
v |= v >> 2;
v |= v >> 4;
v |= v >> 8;
v |= v >> 16;
return multiplyDeBruijnBitPosition[(uint32_t)(v * 0x7C4ACDDU) >> 27];
}
#endif
// position of the highest bit set
static int __PackedArray_highestBitSet(uint32_t v)
{
#if defined(_MSC_VER) && _MSC_VER >= 1400
unsigned long index;
return _BitScanReverse(&index, v) ? index : -1;
#elif defined(__GNUC__)
return v == 0 ? -1 : 31 - __builtin_clz(v);
#else
return v != 0 ? __PackedArray_log2(v) : -1;
#endif
}
uint32_t PackedArray_computeBitsPerItem(const uint32_t* in, uint32_t count)
{
uint32_t i, in_max, bitsPerItem;
in_max = 0;
for (i = 0; i < count; ++i)
in_max = in[i] > in_max ? in[i] : in_max;
bitsPerItem = __PackedArray_highestBitSet(in_max) + 1;
return bitsPerItem == 0 ? 1 : bitsPerItem;
}
// - 8< ------------------------------------------------------------------------
#if defined(PACKEDARRAY_SELF_TEST) && defined(PACKEDARRAY_SELF_BENCH)
#error choose either PACKEDARRAY_SELF_TEST or PACKEDARRAY_SELF_BENCH
#endif
#if defined(PACKEDARRAY_SELF_TEST)
#undef NDEBUG // we want asserts
#include <assert.h>
#include <stdio.h>
#include <string.h> // memcmp
static void PackedArray_pack_reference(PackedArray* a, uint32_t offset, const uint32_t* in, uint32_t count)
{
uint32_t* __restrict out;
uint32_t bitsPerItem;
uint32_t startBit;
uint32_t bitsAvailable;
uint32_t mask;
assert(a != NULL);
assert(in != NULL);
bitsPerItem = a->bitsPerItem;
mask = (uint32_t)((1ULL << bitsPerItem) - 1);
while (count--)
{
uint32_t value = *in++;
assert(0 == (~mask & value));
out = &a->buffer[((uint64_t)offset / 4 * (uint64_t)bitsPerItem) / 32 * 4 + offset % 4];
startBit = ((uint64_t)offset / 4 * (uint64_t)bitsPerItem) % 32;
++offset;
bitsAvailable = 32 - startBit;
if (bitsPerItem <= bitsAvailable)
{
out[0] = (out[0] & ~(mask << startBit)) | (value << startBit);
}
else
{
// value spans 2 buffer cells
uint32_t low, high;
low = value << startBit;
high = value >> bitsAvailable;
out[0] = (out[0] & ~(mask << startBit)) | low;
out[4] = (out[4] & ~(mask >> bitsAvailable)) | high;
}
}
}
static void PackedArray_unpack_reference(const PackedArray* a, uint32_t offset, uint32_t* out, uint32_t count)
{
const uint32_t* __restrict in;
uint32_t bitsPerItem;
uint32_t startBit;
uint32_t bitsAvailable;
uint32_t mask;
assert(a != NULL);
assert(out != NULL);
bitsPerItem = a->bitsPerItem;
mask = (uint32_t)((1ULL << bitsPerItem) - 1);
while (count--)
{
uint32_t value;
in = &a->buffer[((uint64_t)offset / 4 * (uint64_t)bitsPerItem) / 32 * 4 + offset % 4];
startBit = ((uint64_t)offset / 4 * (uint64_t)bitsPerItem) % 32;
++offset;
bitsAvailable = 32 - startBit;
if (bitsPerItem <= bitsAvailable)
{
value = (in[0] >> startBit) & mask;
}
else
{
// value spans 2 buffer cells
uint32_t low, high;
low = in[0] >> startBit;
high = in[4] << bitsAvailable;
value = low ^ ((low ^ high) & (mask >> bitsAvailable << bitsAvailable));
}
*out++ = value;
}
}
int main(void)
{
uint32_t bitsPerItem;
printf("-- PackedArray self test -------------------------------------------------------\n");
printf("\n");
printf("sizeof(PackedArray) = %d\n", (int)sizeof(PackedArray));
printf("\n");
printf("1 by 1 packing / unpacking:\n");
for (bitsPerItem = 1; bitsPerItem <= 32; ++bitsPerItem)
{
uint32_t mask = (uint32_t)(1ULL << bitsPerItem) - 1;
int count;
for (count = 1; count <= 1024; ++count)
{
PackedArray* a1 = PackedArray_create(bitsPerItem, count);
PackedArray* a2 = PackedArray_create(bitsPerItem, count);
int i;
assert(a1->count == (uint32_t)count);
assert(a2->count == (uint32_t)count);
assert(a1->bitsPerItem == bitsPerItem);
assert(a2->bitsPerItem == bitsPerItem);
assert(PackedArray_bufferSize(a1) == PackedArray_bufferSize(a2));
for (i = 0; i < (int)PackedArray_bufferSize(a1); ++i)
a1->buffer[i] = a2->buffer[i] = rand();
for (i = 0; i < count; ++i)
{
uint32_t v1, v2;
v1 = rand() & mask;
v2 = v1 + 1;
PackedArray_pack(a1, i, &v1, 1);
PackedArray_pack_reference(a2, i, &v1, 1);
assert(memcmp(a1->buffer, a2->buffer, sizeof(a1->buffer[0]) * PackedArray_bufferSize(a1)) == 0);
PackedArray_set(a1, i, v1);
assert(memcmp(a1->buffer, a2->buffer, sizeof(a1->buffer[0]) * PackedArray_bufferSize(a1)) == 0);
PackedArray_unpack(a1, i, &v2, 1);
assert(v1 == v2);
PackedArray_unpack_reference(a2, i, &v2, 1);
assert(v1 == v2);
v2 = PackedArray_get(a2, i);
assert(v1 == v2);
}
for (i = count - 1; i >= 0; --i)
{
uint32_t v1, v2;
v1 = rand() & mask;
v2 = v1 + 1;
PackedArray_pack(a1, i, &v1, 1);
PackedArray_pack_reference(a2, i, &v1, 1);
assert(memcmp(a1->buffer, a2->buffer, sizeof(a1->buffer[0]) * PackedArray_bufferSize(a1)) == 0);
PackedArray_set(a1, i, v1);
assert(memcmp(a1->buffer, a2->buffer, sizeof(a1->buffer[0]) * PackedArray_bufferSize(a1)) == 0);
PackedArray_unpack(a1, i, &v2, 1);
assert(v1 == v2);
PackedArray_unpack_reference(a2, i, &v2, 1);
assert(v1 == v2);
v2 = PackedArray_get(a2, i);
assert(v1 == v2);
}
PackedArray_destroy(a1);
PackedArray_destroy(a2);
}
printf(" %2d bits per item -- success.\n", bitsPerItem);
}
printf("\n");
printf("bulk packing / unpacking:\n");
for (bitsPerItem = 1; bitsPerItem <= 32; ++bitsPerItem)
{
uint32_t mask = (uint32_t)(1ULL << bitsPerItem) - 1;
int count;
for (count = 1; count <= 128; ++count)
{
uint32_t* v1;
uint32_t* v2;
PackedArray* a1;
PackedArray* a2;
int i, j;
v1 = (uint32_t*)malloc(sizeof(uint32_t) * count);
assert(v1 != NULL);
v2 = (uint32_t*)malloc(sizeof(uint32_t) * count);
assert(v2 != NULL);
a1 = PackedArray_create(bitsPerItem, count);
assert(a1 != NULL);
a2 = PackedArray_create(bitsPerItem, count);
assert(a2 != NULL);
for (i = 0; i < (int)PackedArray_bufferSize(a1); ++i)
a1->buffer[i] = a2->buffer[i] = rand();
for (i = 0; i < count; ++i)
v1[i] = rand() & mask;
assert(bitsPerItem >= PackedArray_computeBitsPerItem(v1, count));
for (i = 0; i < count; ++i)
{
for (j = 1; j <= count - i; ++j)
{
PackedArray_pack(a1, i, v1, j);
PackedArray_pack_reference(a2, i, v1, j);
assert(memcmp(a1->buffer, a2->buffer, sizeof(a1->buffer[0]) * PackedArray_bufferSize(a1)) == 0);
PackedArray_unpack(a1, i, v2, j);
assert(bitsPerItem >= PackedArray_computeBitsPerItem(v2, j));
assert(memcmp(v1, v2, j * sizeof(uint32_t)) == 0);
PackedArray_unpack_reference(a2, i, v2, j);
assert(memcmp(v1, v2, j * sizeof(uint32_t)) == 0);
}
}
PackedArray_destroy(a1);
PackedArray_destroy(a2);
free(v1);
free(v2);
}
printf(" %2d bits per item -- success.\n", bitsPerItem);
}
return 0;
}
#elif defined(PACKEDARRAY_SELF_BENCH) // #if defined(PACKEDARRAY_SELF_TEST)
#ifndef NDEBUG
#error please define NDEBUG to inhibit asserts when compiling the benchmark
#endif
#include <stdio.h>
#include <string.h>
#include <float.h>
#ifndef MIN
#define MIN(x, y) (((x) < (y)) ? (x) : (y))
#endif
#ifndef MAX
#define MAX(x, y) (((x) < (y)) ? (y) : (x))
#endif
#ifdef _MSC_VER
#pragma warning(push, 3)
#include <windows.h>
#pragma warning(pop)
static double getChronometerTime(void)
{
LARGE_INTEGER frequency;
LARGE_INTEGER t;
QueryPerformanceFrequency(&frequency);
QueryPerformanceCounter(&t);
return (double)t.QuadPart / (double)frequency.QuadPart * 1000;
}
#else
#include <sys/time.h>
static double getChronometerTime()
{
struct timeval now = { 0 };
gettimeofday(&now, NULL);
return (double)now.tv_sec + (double)now.tv_usec * 1e-6;
}
#endif
#define LOOP_COUNT 1000
static double bench_memcpy(uint32_t* in, uint32_t* out, uint32_t count)
{
double start, end;
uint32_t i;
start = getChronometerTime();
for (i = 0; i < LOOP_COUNT; ++i)
memcpy(out, in, count * sizeof(uint32_t));
end = getChronometerTime();
return 1e6 * (end - start) / LOOP_COUNT;
}
static double bench_loopcpy(uint32_t* in, uint32_t* out, uint32_t count)
{
double start, end;
uint32_t i;
start = getChronometerTime();
for (i = 0; i < LOOP_COUNT; ++i)
{
uint32_t j;
for (j = 0; j < count; ++j)
out[j] = in[j];
}
end = getChronometerTime();
return 1e6 * (end - start) / LOOP_COUNT;
}
static double bench_pack(uint32_t* in, PackedArray* out, uint32_t count)
{
double start, end;
int i;
start = getChronometerTime();
for (i = 0; i < LOOP_COUNT; ++i)
PackedArray_pack(out, 0, in, count);
end = getChronometerTime();
return 1e6 * (end - start) / LOOP_COUNT;
}
static double bench_unpack(PackedArray* in, uint32_t* out, uint32_t count)
{
double start, end;
int i;
start = getChronometerTime();
for (i = 0; i < LOOP_COUNT; ++i)
PackedArray_unpack(in, 0, out, count);
end = getChronometerTime();
return 1e6 * (end - start) / LOOP_COUNT;
}
#define MAX_ELEMENT_COUNT (1 << 18)
#define LOG2_MAX_ELEMENT_COUNT 18
int main(void)
{
double start, end;
uint32_t* b1;
uint32_t* b2;
uint32_t count, bitsPerItem;
PackedArray** packed;
uint32_t i;
double* speed_memcpy;
double avg_memcpy, min_memcpy, max_memcpy;
double* speed_loopcpy;
double avg_loopcpy, min_loopcpy, max_loopcpy;
double* speed_pack[32];
double avg_pack, min_pack, max_pack;
double* speed_unpack[32];
double avg_unpack, min_unpack, max_unpack;
printf("-- PackedArray self bench ------------------------------------------------------\n");
start = getChronometerTime();
b1 = (uint32_t*)malloc(sizeof(uint32_t) * MAX_ELEMENT_COUNT);
assert(b1 != NULL);
b2 = (uint32_t*)malloc(sizeof(uint32_t) * MAX_ELEMENT_COUNT);
assert(b2 != NULL);
packed = (PackedArray**)malloc(sizeof(PackedArray*) * 32);
assert(packed != NULL);
for (i = 0; i < 32; ++i)
packed[i] = PackedArray_create(i + 1, MAX_ELEMENT_COUNT);
for (i = 0; i < MAX_ELEMENT_COUNT; ++i)
b1[i] = rand();
speed_memcpy = (double*)malloc(sizeof(double) * (LOG2_MAX_ELEMENT_COUNT + 1));
assert(speed_memcpy != NULL);
avg_memcpy = 0;
min_memcpy = DBL_MAX;
max_memcpy = 0;
printf("memcpy:\n");
printf("bits\tsize (B)\ttime (µs)\tspeed (B/µs)\n");
for (count = 1, i = 0; count <= MAX_ELEMENT_COUNT; count *= 2, ++i)
{
double elapsed = bench_memcpy(b1, b2, count);
double speed = count * sizeof(uint32_t) / elapsed;
printf("%4d\t%8d\t%9.3f\t%12.3f\n", 32, (uint32_t)(count * sizeof(uint32_t)), elapsed, speed);
avg_memcpy += speed;
min_memcpy = MIN(min_memcpy, speed);
max_memcpy = MAX(max_memcpy, speed);
speed_memcpy[i] = speed;
}
avg_memcpy /= i;
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)\n");
printf("%10.3f\t%10.3f\t%10.3f\n", avg_memcpy, min_memcpy, max_memcpy);
printf("\n");
speed_loopcpy = (double*)malloc(sizeof(double) * (LOG2_MAX_ELEMENT_COUNT + 1));
assert(speed_loopcpy != NULL);
avg_loopcpy = 0;
min_loopcpy = DBL_MAX;
max_loopcpy = 0;
printf("loopcpy:\n");
printf("bits\tsize (B)\ttime (µs)\tspeed (B/µs)\n");
for (count = 1, i = 0; count <= MAX_ELEMENT_COUNT; count *= 2, ++i)
{
double elapsed = bench_loopcpy(b1, b2, count);
double speed = count * sizeof(uint32_t) / elapsed;
printf("%4d\t%8d\t%9.3f\t%12.3f\n", 32, (uint32_t)(count * sizeof(uint32_t)), elapsed, speed);
avg_loopcpy += speed;
min_loopcpy = MIN(min_loopcpy, speed);
max_loopcpy = MAX(max_loopcpy, speed);
speed_loopcpy[i] = speed;
}
avg_loopcpy /= i;
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)\n");
printf("%10.3f\t%10.3f\t%10.3f\n", avg_loopcpy, min_loopcpy, max_loopcpy);
printf("\n");
for (bitsPerItem = 1; bitsPerItem <= 32; ++bitsPerItem)
{
avg_pack = 0;
min_pack = DBL_MAX;
max_pack = 0;
avg_unpack = 0;
min_unpack = DBL_MAX;
max_unpack = 0;
printf("pack:\t \t \t \t");
printf("unpack:\t \t \t \t");
printf("\n");
printf("bits\tsize (B)\ttime (µs)\tspeed (B/µs)");
printf("\t");
printf("bits\tsize (B)\ttime (µs)\tspeed (B/µs)");
printf("\n");
speed_pack[bitsPerItem - 1] = (double*)malloc(sizeof(double) * (LOG2_MAX_ELEMENT_COUNT + 1));
assert(speed_pack[bitsPerItem - 1] != NULL);
speed_unpack[bitsPerItem - 1] = (double*)malloc(sizeof(double) * (LOG2_MAX_ELEMENT_COUNT + 1));
assert(speed_unpack[bitsPerItem - 1] != NULL);
for (count = 1, i = 0; count <= MAX_ELEMENT_COUNT; count *= 2, ++i)
{
uint32_t mask = (uint32_t)(1ULL << bitsPerItem) - 1;
uint32_t j;
double elapsed, speed;
for (j = 0; j < count; ++j)
b2[j] = b1[j] & mask;
elapsed = bench_pack(b2, packed[bitsPerItem - 1], count);
speed = count * sizeof(uint32_t) / elapsed;
printf("%4d\t%8d\t%9.3f\t%12.3f", bitsPerItem, (uint32_t)(count * sizeof(uint32_t)), elapsed, speed);
avg_pack += speed;
min_pack = MIN(min_pack, speed);
max_pack = MAX(max_pack, speed);
speed_pack[bitsPerItem - 1][i] = speed;
printf("\t");
elapsed = bench_unpack(packed[bitsPerItem - 1], b2, count);
speed = count * sizeof(uint32_t) / elapsed;
printf("%4d\t%8d\t%9.3f\t%12.3f", bitsPerItem, (uint32_t)(count * sizeof(uint32_t)), elapsed, speed);
avg_unpack += speed;
min_unpack = MIN(min_unpack, speed);
max_unpack = MAX(max_unpack, speed);
speed_unpack[bitsPerItem - 1][i] = speed;
printf("\n");
}
printf("%d\n",i);
assert(i == LOG2_MAX_ELEMENT_COUNT + 1);
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\t\t");
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\n");
avg_pack /= i;
printf("%10.3f\t%10.3f\t%10.3f", avg_pack, min_pack, max_pack);
printf("\t\t");
avg_unpack /= i;
printf("%10.3f\t%10.3f\t%10.3f", avg_unpack, min_unpack, max_unpack);
printf("\n");
printf("\n");
}
printf("\n");
printf("stats by bits per item\n");
printf("pack:\t \t \t \t");
printf("unpack:\t \t \t \t");
printf("\n");
printf("bits\tavg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\t");
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\n");
for (bitsPerItem = 1; bitsPerItem <= 32; ++bitsPerItem)
{
avg_pack = 0;
min_pack = DBL_MAX;
max_pack = 0;
avg_unpack = 0;
min_unpack = DBL_MAX;
max_unpack = 0;
for (count = 1, i = 0; count <= MAX_ELEMENT_COUNT; count *= 2, ++i)
{
double speed;
speed = speed_pack[bitsPerItem - 1][i];
avg_pack += speed;
min_pack = MIN(min_pack, speed);
max_pack = MAX(max_pack, speed);
speed = speed_unpack[bitsPerItem - 1][i];
avg_unpack += speed;
min_unpack = MIN(min_unpack, speed);
max_unpack = MAX(max_unpack, speed);
}
assert(i == LOG2_MAX_ELEMENT_COUNT + 1);
printf("%4d\t", bitsPerItem);
avg_pack /= i;
printf("%10.3f\t%10.3f\t%10.3f", avg_pack, min_pack, max_pack);
printf("\t");
avg_unpack /= i;
printf("%10.3f\t%10.3f\t%10.3f", avg_unpack, min_unpack, max_unpack);
printf("\n");
}
printf("\n");
printf("stats by size\n");
printf("pack:\t \t \t \t");
printf("unpack:\t \t \t \t");
printf("\n");
printf("size(B)\tavg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\t");
printf("avg (B/µs)\tmin (B/µs)\tmax (B/µs)");
printf("\n");
for (count = 1, i = 0; count <= MAX_ELEMENT_COUNT; count *= 2, ++i)
{
avg_pack = 0;
min_pack = DBL_MAX;
max_pack = 0;
avg_unpack = 0;
min_unpack = DBL_MAX;
max_unpack = 0;
for (bitsPerItem = 1; bitsPerItem <= 32; ++bitsPerItem)
{
double speed;
speed = speed_pack[bitsPerItem - 1][i];
avg_pack += speed;
min_pack = MIN(min_pack, speed);
max_pack = MAX(max_pack, speed);
speed = speed_unpack[bitsPerItem - 1][i];
avg_unpack += speed;
min_unpack = MIN(min_unpack, speed);
max_unpack = MAX(max_unpack, speed);
}
printf("%7d\t", (uint32_t)sizeof(uint32_t) * count);
avg_pack /= 32;
printf("%10.3f\t%10.3f\t%10.3f", avg_pack, min_pack, max_pack);
printf("\t");
avg_unpack /= 32;
printf("%10.3f\t%10.3f\t%10.3f", avg_unpack, min_unpack, max_unpack);
printf("\n");
}
printf("\n");
free(b1);
free(b2);
free(speed_memcpy);
free(speed_loopcpy);
for (i = 0; i < 32; ++i)
{
PackedArray_destroy(packed[i]);
free(speed_pack[i]);
free(speed_unpack[i]);
}
free(packed);
end = getChronometerTime();
printf("total time (s): %f\n", (end - start));
printf("\n");
return 0;
}
#endif // #elif defined(PACKEDARRAY_SELF_BENCH)
#endif // #ifdef PACKEDARRAY_IMPL
================================================
FILE: README.md
================================================
# PackedArray: random access array of tightly packed unsigned integers
[](https://travis-ci.org/gpakosz/PackedArray)
## TLDR
*PackedArray comes to the rescue when you're in a desperate need for an uint9_t
or uint17_t array.*
## What?
When you want to hold an unordered sequence of unsigned integers into memory,
the C programming language lets you choose among 4 data types:
- `uint8_t`
- `uint16_t`
- `uint32_t`
- `uint64_t`
If your numbers are within the [0, 100000] range, only 17 bits per integer are
needed since 2<sup>17</sup> = 131072. However, you can't use an array of
`uint16_t` because 16 bits are not enough to store numbers between 65536 and
100000. When you use the next available type, `uint32_t`, you're wasting 15 bits
per integer which represents a 47% overhead in terms of storage requirements.
`PackedArray` saves memory by packing integers/items together at the bit-level:
<table class="monospace">
<tr>
<td colspan="32">b0</td>
<td colspan="32">b1</td>
<td colspan="32">b2</td>
<td style="border-style: dashed; border-right: none;">...</td>
</tr>
<tr>
<td colspan="9">i0</td>
<td colspan="9">i1</td>
<td colspan="9">i2</td>
<td colspan="9">i3</td>
<td colspan="9">i4</td>
<td colspan="9">i5</td>
<td colspan="9">i6</td>
<td colspan="9">i7</td>
<td colspan="9">i8</td>
<td colspan="9">i9</td>
<td style="border-style: dashed; border-right: none;">...</td>
</tr>
</table>
A `PackedArray` is backed by an `uint32_t` buffer. Several items end up being
stored inside the same buffer cell, e.g. i0, i1, and i2. Some items span two
buffer cells, e.g. i3, and i7. `PackedArray` is responsible for
encoding/decoding items into/from the storage buffer.
`PackedArraySIMD` is a `PackedArray` variant that makes use of SSE2 or NEON
instructions.
Going SIMD processes integers 4 by 4 but imposes an interleaved layout in the
storage buffer.
`PackedArraySIMD` interleaved layout, 13 bits per item:
<table class="monospace">
<tr>
<td colspan="32">b0</td>
<td colspan="32">b1</td>
<td colspan="32">b2</td>
<td colspan="32">b3</td>
<td style="border-style: dashed; border-right: none;">...</td>
</tr>
<tr>
<td colspan="13">i0</td>
<td colspan="13">i4</td>
<td colspan="6">i8a</td>
<td colspan="13">i1</td>
<td colspan="13">i5</td>
<td colspan="6">i9a</td>
<td colspan="13">i2</td>
<td colspan="13">i6</td>
<td colspan="6">i10a</td>
<td colspan="13">i3</td>
<td colspan="13">i7</td>
<td colspan="6">i11a</td>
<td style="border-style: dashed;">i8b</td>
<td style="border-style: dashed; border-right: none;">...</td>
</tr>
</table>
As a consequence, the data layout of `PackedArraySIMD` isn't compatible with its
non SIMD counterpart. In other words, you cannot use `PackedArray` to unpack
data packed with `PackedArraySIMD` or the other way around.
It is also worth noting the implementations of `PackedArraySIMD_pack` and
`PackedArraySIMD_unpack` require more plumbing than their non-SIMD counterparts.
Additional computations are needed to find out and adjust a data window that can
be processed 4 by 4 with SIMD instructions.
`PackedArray` and `PackedArraySIMD` are released under the WTFPL v2 license.
For more information, see the [PackedArray announcement on my personal website].
[PackedArray announcement on my personal website]: http://pempek.net/articles/2013/08/03/packedarray-random-access-array-tightly-packed-unsigned-integers/
## Why?
`PackedArray` is designed as a drop-in replacement for an unsigned integer
array. I couldn't find such a data structure in the wild, so I implemented one.
Instead of writing:
uint32_t* a = (uint32_t*)malloc(sizeof(uint32_t) * count);
...
value = a[i];
...
a[j] = value;
You write:
PackedArray* a = PackedArray_create(bitsPerItem, count);
...
value = PackedArray_get(a, i);
...
PackedArray_set(a, j, value);
The `PackedArray_computeBitsPerItem` helper scans a `uint32_t` array and returns
the number of bits needed to create a `PackedArray` capable of holding its
content.
There are also `PackedArray_pack` and `PackedArray_unpack` that operate on
several items in a row. Those two could really have been named
`PackedArray_write` and `PackedArray_read` but I decided "pack" / "unpack"
conveys better something is happening under the hood.
// bulk packing / unpacking
PackedArray_pack(a, j, in, count);
PackedArray_unpack(a, j, out, count);
// the following are semantically equivalent
PackedArray_set(a, j, value);
PackedArray_pack(a, j, &value, 1);
value = PackedArray_get(a, i);
PackedArray_unpack(a, i, &value, 1);
--------------------------------------------------------------------------------
## Compiling
In order to use `PackedArray` or `PackedArraySIMD` in your own project, you just
have to bring in the two `PackedArray.h` and `PackedArray.c` (or
`PackedArraySIMD.c`) files. It's that simple.
You can customize `PackedArray.c`'s behavior by defining the following macros:
- `PACKEDARRAY_ASSERT`
- `PACKEDARRAY_MALLOC`
- `PACKEDARARY_FREE`
You can customize `PackedArraySIMD.c`'s behavior by defining the following
macros:
- `PACKEDARRAY_ASSERT`
- `PACKEDARRAY_ALIGNED_MALLOC`
- `PACKEDARARY_FREE`
`PackedArray.c` and `PackedArraySIMD.c` can compile themselves into either a
test program or a micro-benchmark. For that, you have to use one of the
following preprocessor directives:
- `PACKEDARRAY_SELF_TEST`
- `PACKEDARRAY_SELF_BENCH`
For example, from command line:
$ cc -o PackedArraySelfTest -DPACKEDARRAY_SELF_TEST -O2 -g PackedArray.c
$ cc -o PackedArraySelfBench -DPACKEDARRAY_SELF_BENCH -DNDEBUG -O2 -g PackedArray.c
$ cc -o PackedArraySIMDSelfTest -DPACKEDARRAY_SELF_TEST -O2 -g PackedArraySIMD.c
$ cc -o PackedArraySIMDSelfBench -DPACKEDARRAY_SELF_BENCH -DNDEBUG -O2 -g PackedArraySIMD.c
### Compiling for Windows
There is a Visual Studio 2012 solution in the `_win-vs11/` folder.
### Compiling for Linux or Mac
There is a GNU Make 3.81 `MakeFile` in the `_gnu-make/` folder:
$ make -C _gnu-make/
### Compiling for Mac
See above if you want to compile from command line. Otherwise there is an Xcode
project located in the `_mac-xcode/` folder.
### Compiling for iOS
There is an Xcode project located in the `_ios-xcode/` folder.
If you prefer compiling from command line and deploying to a jailbroken device
through SSH, use:
$ make -C _gnu-make/ binsubdir=ios CC="$(xcrun --sdk iphoneos --find clang) -isysroot $(xcrun --sdk iphoneos --show-sdk-path) -arch armv7 -arch armv7s -arch arm64" postbuild="codesign -s 'iPhone Developer'"
### Compiling for Android
You will have to install the Android NDK, and point the `$NDK_ROOT` environment
variable to the NDK path: e.g. `export NDK_ROOT=/opt/android-ndk` (without a
trailing `/` character).
Next, the easy way is to make a standalone Android toolchain with the following
command:
$ $NDK_ROOT/build/tools/make-standalone-toolchain.sh --system=$(uname -s | tr [A-Z] [a-z])-$(uname -m) --platform=android-3 --toolchain=arm-linux-androideabi-clang3.3 --install-dir=/tmp/android-clang
Now you can compile the self test and self benchmark programs by running:
$ make -C _gnu-make/ binsubdir=android CC=/tmp/android-clang/bin/clang CFLAGS='-march=armv7-a -mfloat-abi=softfp -mfpu=neon -O2'
--------------------------------------------------------------------------------
## Implementation details, what the hell is going on?
First, in `PackedArray.c` or `PackedArraySIMD.c`, everything that comes below
the `- 8< ----` marker is the code for the self test and self micro-benchmark
programs and can be discarded if you really want to:
If you want to cut down your anxiety, you can use the provided GNU Makefile and
invoke:
$ make -C _gnu-make/ cut
This produces the `PackedArray.cut.c` and `PackedArraySIMD.cut.c` files.
You may also be troubled by `PackedArray.c` and `PackedArraySIMD.c` including
themselves with `#include PACKEDARRAY_SELF`. By combining preprocessing tricks
and including themselves, `PackedArray.c` and `PackedArraySIMD.c`
"generate the code" for the unrolled pack and unpack implementations.
By default `PACKEDARRAY_SELF` is defined to `"PackedArray.c"` which assumes the
compiler is going to look for the file in the same directory as the file from
which the `#include` statement is being evaluated. This helps compiling when the
build system refers to the source files with relative paths. Depending on your
compiler/build system combination you may want to override `PACKEDARRAY_SELF` to
`__FILE__`.
If you want to see the generated code, you can use the provided GNU Makefile and
invoke:
$ make -C _gnu-make/ preprocess
This produces the `PackedArray.pp.c` and `PackedArraySIMD.pp.c` files.
--------------------------------------------------------------------------------
If you find `PackedArray` or `PackedArraySIMD` useful and decide to use it in
your own projects please drop me a line [@gpakosz].
If you use it in a commercial project, consider using [Gittip].
[@gpakosz]: https://twitter.com/gpakosz
[Gittip]: https://www.gittip.com/gpakosz/
================================================
FILE: _gnu-make/Makefile
================================================
.PHONY: build test preprocess cut assembly clean
# directories
ifeq ($(realpath .),)
$(error your version of Make doesn't support $$(realpath names...) - please use GNU Make 3.81 or later)
endif
ifeq ($(platform),)
__uname_s := $(shell sh -c 'uname -s 2>/dev/null | tr [A-Z] [a-z] || echo unknown-platform')
__uname_m := $(shell sh -c 'uname -m 2>/dev/null | tr [A-Z] [a-z] || echo unknown-architecture')
ifeq ($(__uname_s),linux)
override platform := linux
override architecture := $(__uname_m)
endif
ifeq ($(__uname_s),darwin)
override platform := mac
override architecture := $(__uname_m)
endif
endif
ifeq ($(architecture),)
override architecture := unknown-architecture
endif
prefix := $(realpath ..)
srcdir := $(realpath ..)
buildir := $(realpath .)/build
binsubdir := $(platform)-$(architecture)
bindir := $(prefix)/bin/$(binsubdir)
CFLAGS := -O2 -g
define _generate_rules
.PHONY: build-$(1)
build: build-$(1)
build-$(1): $(bindir)/$(1)SelfTest $(bindir)/$(1)SelfBench
$(bindir)/$(1)SelfTest: $(srcdir)/$(1).c $(srcdir)/PackedArray.h
mkdir -p $$(@D)
$(CC) -o $$@ -DPACKEDARRAY_SELF_TEST -std=c99 -pedantic $(CFLAGS) $$<
$$(if $(postbuild),$(postbuild) $$@)
$(bindir)/$(1)SelfBench: $(srcdir)/$(1).c $(srcdir)/PackedArray.h
mkdir -p $$(@D)
$(CC) -o $$@ -DPACKEDARRAY_SELF_BENCH -DNDEBUG -std=c99 -pedantic $(CFLAGS) $$<
$$(if $(postbuild),$(postbuild) $$@)
.PHONY: assembly-$(1)
assembly: assembly-$(1)
assembly-$(1): $(bindir)/$(1)SelfTest.s $(bindir)/$(1)SelfBench.s
$(bindir)/$(1)SelfTest.s: $(srcdir)/$(1).c $(srcdir)/PackedArray.h
@echo Generating $$(@F) assembly listing
@mkdir -p $$(@D)
@$(CC) -o $$@ -S -fverbose-asm -g -DPACKEDARRAY_SELF_TEST -std=c99 -pedantic $(CFLAGS) $$<
$(bindir)/$(1)SelfBench.s: $(srcdir)/$(1).c $(srcdir)/PackedArray.h
@echo Generating $$(@F) assembly listing
@mkdir -p $$(@D)
@$(CC) -o $$@ -S -fverbose-asm -g -DPACKEDARRAY_SELF_BENCH -DNDEBUG -std=c99 -pedantic $(CFLAGS) $$<
.PHONY: test-$(1)
test : test-$(1)
test-$(1): build-$(1)
$(bindir)/$(1)SelfTest
.PHONY: preprocess-$(1)
preprocess: preprocess-$(1)
preprocess-$(1): $(srcdir)/$(1).c
@echo Preprocessing $(1).c to $(1).pp.c
$$(eval tmp := $$(shell mktemp -t $(1)XXX))
@sed -E\
-e 's/#include (<|").+(>|")/pp(&)/'\
$(srcdir)/$(1).c > $$(tmp)
@$(CPP) $(CPPFLAGS) -DPACKEDARRAY_SELF=__FILE__ $(CFLAGS) -xc $$(tmp) | sed -E\
-e '/^# /d'\
-e 's/[ \t]*$$$$//'\
-e 's/pp\((.*)\)/\1/'\
| cat -s > $(srcdir)/$(1).pp.c
@rm -f $$(tmp)
.PHONY: cut-$(1)
cut: cut-$(1)
cut-$(1): $(srcdir)/$(1).c
@echo Cutting down your anxiety to $(1).cut.c
@sed -n '/- 8<.*/q;p' $(srcdir)/$(1).c > $(srcdir)/$(1).cut.c
endef
$(foreach p,PackedArray PackedArraySIMD, $(eval $(call _generate_rules,$(p))))
clean:
rm -rf $(buildir)
rm -rf $(bindir)
rm -rf $(srcdir)/*.pp.c
rm -rf $(srcdir)/*.cut.c
================================================
FILE: _ios-xcode/.gitignore
================================================
xcuserdata/
xcshareddata/
================================================
FILE: _ios-xcode/PackedArray-Info.plist
================================================
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
<key>CFBundleDevelopmentRegion</key>
<string>en</string>
<key>CFBundleDisplayName</key>
<string>${PRODUCT_NAME}</string>
<key>CFBundleExecutable</key>
<string>${EXECUTABLE_NAME}</string>
<key>CFBundleIdentifier</key>
<string>net.pempek.${PRODUCT_NAME:rfc1034identifier}</string>
<key>CFBundleInfoDictionaryVersion</key>
<string>6.0</string>
<key>CFBundleName</key>
<string>${PRODUCT_NAME}</string>
<key>CFBundlePackageType</key>
<string>APPL</string>
<key>CFBundleShortVersionString</key>
<string>1.0</string>
<key>CFBundleSignature</key>
<string>????</string>
<key>CFBundleVersion</key>
<string>1.0</string>
<key>LSRequiresIPhoneOS</key>
<true/>
<key>UIRequiredDeviceCapabilities</key>
<array>
<string>armv7</string>
</array>
<key>UISupportedInterfaceOrientations</key>
<array>
<string>UIInterfaceOrientationPortrait</string>
<string>UIInterfaceOrientationLandscapeLeft</string>
<string>UIInterfaceOrientationLandscapeRight</string>
</array>
<key>UISupportedInterfaceOrientations~ipad</key>
<array>
<string>UIInterfaceOrientationPortrait</string>
<string>UIInterfaceOrientationPortraitUpsideDown</string>
<string>UIInterfaceOrientationLandscapeLeft</string>
<string>UIInterfaceOrientationLandscapeRight</string>
</array>
</dict>
</plist>
================================================
FILE: _ios-xcode/PackedArray.xcodeproj/project.pbxproj
================================================
// !$*UTF8*$!
{
archiveVersion = 1;
classes = {
};
objectVersion = 46;
objects = {
/* Begin PBXBuildFile section */
1A98F9EC17A406A700BF09FF /* PackedArray.c in Sources */ = {isa = PBXBuildFile; fileRef = 1A98F9EB17A406A700BF09FF /* PackedArray.c */; };
1A98F9F017A408F000BF09FF /* PackedArray.c in Sources */ = {isa = PBXBuildFile; fileRef = 1A98F9EB17A406A700BF09FF /* PackedArray.c */; };
1AC5A25517AD052200249A68 /* PackedArraySIMD.c in Sources */ = {isa = PBXBuildFile; fileRef = 1AC5A24017AD049E00249A68 /* PackedArraySIMD.c */; };
1AC5A25617AD052400249A68 /* PackedArraySIMD.c in Sources */ = {isa = PBXBuildFile; fileRef = 1AC5A24017AD049E00249A68 /* PackedArraySIMD.c */; };
/* End PBXBuildFile section */
/* Begin PBXFileReference section */
1A98F9C917A4018400BF09FF /* PackedArraySelfTest.app */ = {isa = PBXFileReference; explicitFileType = wrapper.application; includeInIndex = 0; path = PackedArraySelfTest.app; sourceTree = BUILT_PRODUCTS_DIR; };
1A98F9EB17A406A700BF09FF /* PackedArray.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = PackedArray.c; sourceTree = "<group>"; };
1A98F9F617A408F000BF09FF /* PackedArraySelfBench.app */ = {isa = PBXFileReference; explicitFileType = wrapper.application; includeInIndex = 0; path = PackedArraySelfBench.app; sourceTree = BUILT_PRODUCTS_DIR; };
1A98FA1A17A4267A00BF09FF /* PackedArray.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = PackedArray.h; sourceTree = "<group>"; };
1AC5A24017AD049E00249A68 /* PackedArraySIMD.c */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.c.c; path = PackedArraySIMD.c; sourceTree = "<group>"; };
1AC5A24917AD04A600249A68 /* PackedArraySIMDSelfTest.app */ = {isa = PBXFileReference; explicitFileType = wrapper.application; includeInIndex = 0; path = PackedArraySIMDSelfTest.app; sourceTree = BUILT_PRODUCTS_DIR; };
1AC5A25317AD04A800249A68 /* PackedArraySIMDSelfBench.app */ = {isa = PBXFileReference; explicitFileType = wrapper.application; includeInIndex = 0; path = PackedArraySIMDSelfBench.app; sourceTree = BUILT_PRODUCTS_DIR; };
/* End PBXFileReference section */
/* Begin PBXFrameworksBuildPhase section */
1A98F9C617A4018400BF09FF /* Frameworks */ = {
isa = PBXFrameworksBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
1A98F9F117A408F000BF09FF /* Frameworks */ = {
isa = PBXFrameworksBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
1AC5A24417AD04A600249A68 /* Frameworks */ = {
isa = PBXFrameworksBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
1AC5A24E17AD04A800249A68 /* Frameworks */ = {
isa = PBXFrameworksBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
/* End PBXFrameworksBuildPhase section */
/* Begin PBXGroup section */
1A98F9C017A4018400BF09FF = {
isa = PBXGroup;
children = (
1A98F9ED17A406AE00BF09FF /* PackedArray */,
1A98F9CA17A4018400BF09FF /* Products */,
);
sourceTree = "<group>";
};
1A98F9CA17A4018400BF09FF /* Products */ = {
isa = PBXGroup;
children = (
1A98F9C917A4018400BF09FF /* PackedArraySelfTest.app */,
1A98F9F617A408F000BF09FF /* PackedArraySelfBench.app */,
1AC5A24917AD04A600249A68 /* PackedArraySIMDSelfTest.app */,
1AC5A25317AD04A800249A68 /* PackedArraySIMDSelfBench.app */,
);
name = Products;
sourceTree = "<group>";
};
1A98F9ED17A406AE00BF09FF /* PackedArray */ = {
isa = PBXGroup;
children = (
1A98F9EB17A406A700BF09FF /* PackedArray.c */,
1A98FA1A17A4267A00BF09FF /* PackedArray.h */,
1AC5A24017AD049E00249A68 /* PackedArraySIMD.c */,
);
name = PackedArray;
path = ..;
sourceTree = "<group>";
};
/* End PBXGroup section */
/* Begin PBXNativeTarget section */
1A98F9C817A4018400BF09FF /* PackedArraySelfTest */ = {
isa = PBXNativeTarget;
buildConfigurationList = 1A98F9E617A4018400BF09FF /* Build configuration list for PBXNativeTarget "PackedArraySelfTest" */;
buildPhases = (
1A98F9C517A4018400BF09FF /* Sources */,
1A98F9C617A4018400BF09FF /* Frameworks */,
1A98F9C717A4018400BF09FF /* Resources */,
);
buildRules = (
);
dependencies = (
);
name = PackedArraySelfTest;
productName = PackedArray;
productReference = 1A98F9C917A4018400BF09FF /* PackedArraySelfTest.app */;
productType = "com.apple.product-type.application";
};
1A98F9EE17A408F000BF09FF /* PackedArraySelfBench */ = {
isa = PBXNativeTarget;
buildConfigurationList = 1A98F9F317A408F000BF09FF /* Build configuration list for PBXNativeTarget "PackedArraySelfBench" */;
buildPhases = (
1A98F9EF17A408F000BF09FF /* Sources */,
1A98F9F117A408F000BF09FF /* Frameworks */,
1A98F9F217A408F000BF09FF /* Resources */,
);
buildRules = (
);
dependencies = (
);
name = PackedArraySelfBench;
productName = PackedArray;
productReference = 1A98F9F617A408F000BF09FF /* PackedArraySelfBench.app */;
productType = "com.apple.product-type.application";
};
1AC5A24117AD04A600249A68 /* PackedArraySIMDSelfTest */ = {
isa = PBXNativeTarget;
buildConfigurationList = 1AC5A24617AD04A600249A68 /* Build configuration list for PBXNativeTarget "PackedArraySIMDSelfTest" */;
buildPhases = (
1AC5A24217AD04A600249A68 /* Sources */,
1AC5A24417AD04A600249A68 /* Frameworks */,
1AC5A24517AD04A600249A68 /* Resources */,
);
buildRules = (
);
dependencies = (
);
name = PackedArraySIMDSelfTest;
productName = PackedArray;
productReference = 1AC5A24917AD04A600249A68 /* PackedArraySIMDSelfTest.app */;
productType = "com.apple.product-type.application";
};
1AC5A24B17AD04A800249A68 /* PackedArraySIMDSelfBench */ = {
isa = PBXNativeTarget;
buildConfigurationList = 1AC5A25017AD04A800249A68 /* Build configuration list for PBXNativeTarget "PackedArraySIMDSelfBench" */;
buildPhases = (
1AC5A24C17AD04A800249A68 /* Sources */,
1AC5A24E17AD04A800249A68 /* Frameworks */,
1AC5A24F17AD04A800249A68 /* Resources */,
);
buildRules = (
);
dependencies = (
);
name = PackedArraySIMDSelfBench;
productName = PackedArray;
productReference = 1AC5A25317AD04A800249A68 /* PackedArraySIMDSelfBench.app */;
productType = "com.apple.product-type.application";
};
/* End PBXNativeTarget section */
/* Begin PBXProject section */
1A98F9C117A4018400BF09FF /* Project object */ = {
isa = PBXProject;
attributes = {
LastUpgradeCheck = 0460;
ORGANIZATIONNAME = "Gregory Pakosz";
};
buildConfigurationList = 1A98F9C417A4018400BF09FF /* Build configuration list for PBXProject "PackedArray" */;
compatibilityVersion = "Xcode 3.2";
developmentRegion = English;
hasScannedForEncodings = 0;
knownRegions = (
en,
);
mainGroup = 1A98F9C017A4018400BF09FF;
productRefGroup = 1A98F9CA17A4018400BF09FF /* Products */;
projectDirPath = "";
projectRoot = "";
targets = (
1A98F9C817A4018400BF09FF /* PackedArraySelfTest */,
1A98F9EE17A408F000BF09FF /* PackedArraySelfBench */,
1AC5A24117AD04A600249A68 /* PackedArraySIMDSelfTest */,
1AC5A24B17AD04A800249A68 /* PackedArraySIMDSelfBench */,
);
};
/* End PBXProject section */
/* Begin PBXResourcesBuildPhase section */
1A98F9C717A4018400BF09FF /* Resources */ = {
isa = PBXResourcesBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
1A98F9F217A408F000BF09FF /* Resources */ = {
isa = PBXResourcesBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
1AC5A24517AD04A600249A68 /* Resources */ = {
isa = PBXResourcesBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
1AC5A24F17AD04A800249A68 /* Resources */ = {
isa = PBXResourcesBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
/* End PBXResourcesBuildPhase section */
/* Begin PBXSourcesBuildPhase section */
1A98F9C517A4018400BF09FF /* Sources */ = {
isa = PBXSourcesBuildPhase;
buildActionMask = 2147483647;
files = (
1A98F9EC17A406A700BF09FF /* PackedArray.c in Sources */,
);
runOnlyForDeploymentPostprocessing = 0;
};
1A98F9EF17A408F000BF09FF /* Sources */ = {
isa = PBXSourcesBuildPhase;
buildActionMask = 2147483647;
files = (
1A98F9F017A408F000BF09FF /* PackedArray.c in Sources */,
);
runOnlyForDeploymentPostprocessing = 0;
};
1AC5A24217AD04A600249A68 /* Sources */ = {
isa = PBXSourcesBuildPhase;
buildActionMask = 2147483647;
files = (
1AC5A25517AD052200249A68 /* PackedArraySIMD.c in Sources */,
);
runOnlyForDeploymentPostprocessing = 0;
};
1AC5A24C17AD04A800249A68 /* Sources */ = {
isa = PBXSourcesBuildPhase;
buildActionMask = 2147483647;
files = (
1AC5A25617AD052400249A68 /* PackedArraySIMD.c in Sources */,
);
runOnlyForDeploymentPostprocessing = 0;
};
/* End PBXSourcesBuildPhase section */
/* Begin XCBuildConfiguration section */
1A98F9E417A4018400BF09FF /* Debug */ = {
isa = XCBuildConfiguration;
buildSettings = {
ALWAYS_SEARCH_USER_PATHS = NO;
ARCHS = "$(ARCHS_STANDARD_INCLUDING_64_BIT)";
CLANG_CXX_LANGUAGE_STANDARD = "gnu++0x";
CLANG_CXX_LIBRARY = "libc++";
CLANG_WARN_CONSTANT_CONVERSION = YES;
CLANG_WARN_EMPTY_BODY = YES;
CLANG_WARN_ENUM_CONVERSION = YES;
CLANG_WARN_INT_CONVERSION = YES;
CLANG_WARN__DUPLICATE_METHOD_MATCH = YES;
"CODE_SIGN_IDENTITY[sdk=iphoneos*]" = "iPhone Developer";
COPY_PHASE_STRIP = NO;
GCC_C_LANGUAGE_STANDARD = gnu99;
GCC_DYNAMIC_NO_PIC = NO;
GCC_OPTIMIZATION_LEVEL = 0;
GCC_SYMBOLS_PRIVATE_EXTERN = NO;
GCC_WARN_ABOUT_RETURN_TYPE = YES;
GCC_WARN_UNINITIALIZED_AUTOS = YES;
GCC_WARN_UNUSED_VARIABLE = YES;
INFOPLIST_FILE = "PackedArray-Info.plist";
IPHONEOS_DEPLOYMENT_TARGET = 7.0;
ONLY_ACTIVE_ARCH = YES;
PRODUCT_NAME = "$(TARGET_NAME)";
SDKROOT = iphoneos;
TARGETED_DEVICE_FAMILY = "1,2";
};
name = Debug;
};
1A98F9E517A4018400BF09FF /* Release */ = {
isa = XCBuildConfiguration;
buildSettings = {
ALWAYS_SEARCH_USER_PATHS = NO;
ARCHS = "$(ARCHS_STANDARD_INCLUDING_64_BIT)";
CLANG_CXX_LANGUAGE_STANDARD = "gnu++0x";
CLANG_CXX_LIBRARY = "libc++";
CLANG_WARN_CONSTANT_CONVERSION = YES;
CLANG_WARN_EMPTY_BODY = YES;
CLANG_WARN_ENUM_CONVERSION = YES;
CLANG_WARN_INT_CONVERSION = YES;
CLANG_WARN__DUPLICATE_METHOD_MATCH = YES;
"CODE_SIGN_IDENTITY[sdk=iphoneos*]" = "iPhone Developer";
COPY_PHASE_STRIP = YES;
GCC_C_LANGUAGE_STANDARD = gnu99;
GCC_OPTIMIZATION_LEVEL = 2;
GCC_PREPROCESSOR_DEFINITIONS = NDEBUG;
GCC_WARN_ABOUT_RETURN_TYPE = YES;
GCC_WARN_UNINITIALIZED_AUTOS = YES;
GCC_WARN_UNUSED_VARIABLE = YES;
INFOPLIST_FILE = "PackedArray-Info.plist";
IPHONEOS_DEPLOYMENT_TARGET = 7.0;
OTHER_CFLAGS = "-DNS_BLOCK_ASSERTIONS=1";
PRODUCT_NAME = "$(TARGET_NAME)";
SDKROOT = iphoneos;
TARGETED_DEVICE_FAMILY = "1,2";
VALIDATE_PRODUCT = YES;
};
name = Release;
};
1A98F9E717A4018400BF09FF /* Debug */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = PACKEDARRAY_SELF_TEST;
WRAPPER_EXTENSION = app;
};
name = Debug;
};
1A98F9E817A4018400BF09FF /* Release */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = PACKEDARRAY_SELF_TEST;
WRAPPER_EXTENSION = app;
};
name = Release;
};
1A98F9F417A408F000BF09FF /* Debug */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_BENCH,
NDEBUG,
"$(inherited)",
);
WRAPPER_EXTENSION = app;
};
name = Debug;
};
1A98F9F517A408F000BF09FF /* Release */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_BENCH,
"$(inherited)",
);
WRAPPER_EXTENSION = app;
};
name = Release;
};
1AC5A24717AD04A600249A68 /* Debug */ = {
isa = XCBuildConfiguration;
buildSettings = {
ARCHS = "$(ARCHS_STANDARD_INCLUDING_64_BIT)";
GCC_PREPROCESSOR_DEFINITIONS = PACKEDARRAY_SELF_TEST;
WRAPPER_EXTENSION = app;
};
name = Debug;
};
1AC5A24817AD04A600249A68 /* Release */ = {
isa = XCBuildConfiguration;
buildSettings = {
ARCHS = "$(ARCHS_STANDARD_INCLUDING_64_BIT)";
GCC_PREPROCESSOR_DEFINITIONS = PACKEDARRAY_SELF_TEST;
WRAPPER_EXTENSION = app;
};
name = Release;
};
1AC5A25117AD04A800249A68 /* Debug */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_BENCH,
NDEBUG,
"$(inherited)",
);
WRAPPER_EXTENSION = app;
};
name = Debug;
};
1AC5A25217AD04A800249A68 /* Release */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_BENCH,
"$(inherited)",
);
WRAPPER_EXTENSION = app;
};
name = Release;
};
/* End XCBuildConfiguration section */
/* Begin XCConfigurationList section */
1A98F9C417A4018400BF09FF /* Build configuration list for PBXProject "PackedArray" */ = {
isa = XCConfigurationList;
buildConfigurations = (
1A98F9E417A4018400BF09FF /* Debug */,
1A98F9E517A4018400BF09FF /* Release */,
);
defaultConfigurationIsVisible = 0;
defaultConfigurationName = Release;
};
1A98F9E617A4018400BF09FF /* Build configuration list for PBXNativeTarget "PackedArraySelfTest" */ = {
isa = XCConfigurationList;
buildConfigurations = (
1A98F9E717A4018400BF09FF /* Debug */,
1A98F9E817A4018400BF09FF /* Release */,
);
defaultConfigurationIsVisible = 0;
defaultConfigurationName = Release;
};
1A98F9F317A408F000BF09FF /* Build configuration list for PBXNativeTarget "PackedArraySelfBench" */ = {
isa = XCConfigurationList;
buildConfigurations = (
1A98F9F417A408F000BF09FF /* Debug */,
1A98F9F517A408F000BF09FF /* Release */,
);
defaultConfigurationIsVisible = 0;
defaultConfigurationName = Release;
};
1AC5A24617AD04A600249A68 /* Build configuration list for PBXNativeTarget "PackedArraySIMDSelfTest" */ = {
isa = XCConfigurationList;
buildConfigurations = (
1AC5A24717AD04A600249A68 /* Debug */,
1AC5A24817AD04A600249A68 /* Release */,
);
defaultConfigurationIsVisible = 0;
defaultConfigurationName = Release;
};
1AC5A25017AD04A800249A68 /* Build configuration list for PBXNativeTarget "PackedArraySIMDSelfBench" */ = {
isa = XCConfigurationList;
buildConfigurations = (
1AC5A25117AD04A800249A68 /* Debug */,
1AC5A25217AD04A800249A68 /* Release */,
);
defaultConfigurationIsVisible = 0;
defaultConfigurationName = Release;
};
/* End XCConfigurationList section */
};
rootObject = 1A98F9C117A4018400BF09FF /* Project object */;
}
================================================
FILE: _ios-xcode/PackedArray.xcodeproj/project.xcworkspace/contents.xcworkspacedata
================================================
<?xml version="1.0" encoding="UTF-8"?>
<Workspace
version = "1.0">
<FileRef
location = "self:PackedArray.xcodeproj">
</FileRef>
</Workspace>
================================================
FILE: _mac-xcode/.gitignore
================================================
xcuserdata/
xcshareddata/
================================================
FILE: _mac-xcode/PackedArray.xcodeproj/project.pbxproj
================================================
// !$*UTF8*$!
{
archiveVersion = 1;
classes = {
};
objectVersion = 46;
objects = {
/* Begin PBXBuildFile section */
1A98FA1017A424D700BF09FF /* PackedArray.c in Sources */ = {isa = PBXBuildFile; fileRef = 1A98FA0E17A424D700BF09FF /* PackedArray.c */; };
1A98FA1317A4262700BF09FF /* PackedArray.c in Sources */ = {isa = PBXBuildFile; fileRef = 1A98FA0E17A424D700BF09FF /* PackedArray.c */; };
1AEF60E217AD035900CA6B64 /* PackedArraySIMD.c in Sources */ = {isa = PBXBuildFile; fileRef = 1AEF60CF17AD02B000CA6B64 /* PackedArraySIMD.c */; };
1AEF60E317AD035A00CA6B64 /* PackedArraySIMD.c in Sources */ = {isa = PBXBuildFile; fileRef = 1AEF60CF17AD02B000CA6B64 /* PackedArraySIMD.c */; };
/* End PBXBuildFile section */
/* Begin PBXCopyFilesBuildPhase section */
1A98F9FF17A4249200BF09FF /* CopyFiles */ = {
isa = PBXCopyFilesBuildPhase;
buildActionMask = 2147483647;
dstPath = /usr/share/man/man1/;
dstSubfolderSpec = 0;
files = (
);
runOnlyForDeploymentPostprocessing = 1;
};
1A98FA1517A4262700BF09FF /* CopyFiles */ = {
isa = PBXCopyFilesBuildPhase;
buildActionMask = 2147483647;
dstPath = /usr/share/man/man1/;
dstSubfolderSpec = 0;
files = (
);
runOnlyForDeploymentPostprocessing = 1;
};
1AEF60D417AD02BA00CA6B64 /* CopyFiles */ = {
isa = PBXCopyFilesBuildPhase;
buildActionMask = 2147483647;
dstPath = /usr/share/man/man1/;
dstSubfolderSpec = 0;
files = (
);
runOnlyForDeploymentPostprocessing = 1;
};
1AEF60DD17AD02BD00CA6B64 /* CopyFiles */ = {
isa = PBXCopyFilesBuildPhase;
buildActionMask = 2147483647;
dstPath = /usr/share/man/man1/;
dstSubfolderSpec = 0;
files = (
);
runOnlyForDeploymentPostprocessing = 1;
};
/* End PBXCopyFilesBuildPhase section */
/* Begin PBXFileReference section */
1A98FA0117A4249200BF09FF /* PackedArraySelfTest */ = {isa = PBXFileReference; explicitFileType = "compiled.mach-o.executable"; includeInIndex = 0; path = PackedArraySelfTest; sourceTree = BUILT_PRODUCTS_DIR; };
1A98FA0E17A424D700BF09FF /* PackedArray.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = PackedArray.c; sourceTree = "<group>"; };
1A98FA0F17A424D700BF09FF /* PackedArray.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = PackedArray.h; sourceTree = "<group>"; };
1A98FA1917A4262700BF09FF /* PackedArraySelfBench */ = {isa = PBXFileReference; explicitFileType = "compiled.mach-o.executable"; includeInIndex = 0; path = PackedArraySelfBench; sourceTree = BUILT_PRODUCTS_DIR; };
1AEF60CF17AD02B000CA6B64 /* PackedArraySIMD.c */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.c.c; path = PackedArraySIMD.c; sourceTree = "<group>"; };
1AEF60D817AD02BA00CA6B64 /* PackedArraySIMDSelfTest */ = {isa = PBXFileReference; explicitFileType = "compiled.mach-o.executable"; includeInIndex = 0; path = PackedArraySIMDSelfTest; sourceTree = BUILT_PRODUCTS_DIR; };
1AEF60E117AD02BD00CA6B64 /* PackedArraySIMDSelfBench */ = {isa = PBXFileReference; explicitFileType = "compiled.mach-o.executable"; includeInIndex = 0; path = PackedArraySIMDSelfBench; sourceTree = BUILT_PRODUCTS_DIR; };
/* End PBXFileReference section */
/* Begin PBXFrameworksBuildPhase section */
1A98F9FE17A4249200BF09FF /* Frameworks */ = {
isa = PBXFrameworksBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
1A98FA1417A4262700BF09FF /* Frameworks */ = {
isa = PBXFrameworksBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
1AEF60D317AD02BA00CA6B64 /* Frameworks */ = {
isa = PBXFrameworksBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
1AEF60DC17AD02BD00CA6B64 /* Frameworks */ = {
isa = PBXFrameworksBuildPhase;
buildActionMask = 2147483647;
files = (
);
runOnlyForDeploymentPostprocessing = 0;
};
/* End PBXFrameworksBuildPhase section */
/* Begin PBXGroup section */
1A98F9F817A4249200BF09FF = {
isa = PBXGroup;
children = (
1A98FA0D17A424BE00BF09FF /* PackedArray */,
1A98FA0217A4249200BF09FF /* Products */,
);
sourceTree = "<group>";
};
1A98FA0217A4249200BF09FF /* Products */ = {
isa = PBXGroup;
children = (
1A98FA0117A4249200BF09FF /* PackedArraySelfTest */,
1A98FA1917A4262700BF09FF /* PackedArraySelfBench */,
1AEF60D817AD02BA00CA6B64 /* PackedArraySIMDSelfTest */,
1AEF60E117AD02BD00CA6B64 /* PackedArraySIMDSelfBench */,
);
name = Products;
sourceTree = "<group>";
};
1A98FA0D17A424BE00BF09FF /* PackedArray */ = {
isa = PBXGroup;
children = (
1A98FA0E17A424D700BF09FF /* PackedArray.c */,
1A98FA0F17A424D700BF09FF /* PackedArray.h */,
1AEF60CF17AD02B000CA6B64 /* PackedArraySIMD.c */,
);
name = PackedArray;
path = ..;
sourceTree = "<group>";
};
/* End PBXGroup section */
/* Begin PBXNativeTarget section */
1A98FA0017A4249200BF09FF /* PackedArraySelfTest */ = {
isa = PBXNativeTarget;
buildConfigurationList = 1A98FA0A17A4249200BF09FF /* Build configuration list for PBXNativeTarget "PackedArraySelfTest" */;
buildPhases = (
1A98F9FD17A4249200BF09FF /* Sources */,
1A98F9FE17A4249200BF09FF /* Frameworks */,
1A98F9FF17A4249200BF09FF /* CopyFiles */,
);
buildRules = (
);
dependencies = (
);
name = PackedArraySelfTest;
productName = PackedArray;
productReference = 1A98FA0117A4249200BF09FF /* PackedArraySelfTest */;
productType = "com.apple.product-type.tool";
};
1A98FA1117A4262700BF09FF /* PackedArraySelfBench */ = {
isa = PBXNativeTarget;
buildConfigurationList = 1A98FA1617A4262700BF09FF /* Build configuration list for PBXNativeTarget "PackedArraySelfBench" */;
buildPhases = (
1A98FA1217A4262700BF09FF /* Sources */,
1A98FA1417A4262700BF09FF /* Frameworks */,
1A98FA1517A4262700BF09FF /* CopyFiles */,
);
buildRules = (
);
dependencies = (
);
name = PackedArraySelfBench;
productName = PackedArray;
productReference = 1A98FA1917A4262700BF09FF /* PackedArraySelfBench */;
productType = "com.apple.product-type.tool";
};
1AEF60D017AD02BA00CA6B64 /* PackedArraySIMDSelfTest */ = {
isa = PBXNativeTarget;
buildConfigurationList = 1AEF60D517AD02BA00CA6B64 /* Build configuration list for PBXNativeTarget "PackedArraySIMDSelfTest" */;
buildPhases = (
1AEF60D117AD02BA00CA6B64 /* Sources */,
1AEF60D317AD02BA00CA6B64 /* Frameworks */,
1AEF60D417AD02BA00CA6B64 /* CopyFiles */,
);
buildRules = (
);
dependencies = (
);
name = PackedArraySIMDSelfTest;
productName = PackedArray;
productReference = 1AEF60D817AD02BA00CA6B64 /* PackedArraySIMDSelfTest */;
productType = "com.apple.product-type.tool";
};
1AEF60D917AD02BD00CA6B64 /* PackedArraySIMDSelfBench */ = {
isa = PBXNativeTarget;
buildConfigurationList = 1AEF60DE17AD02BD00CA6B64 /* Build configuration list for PBXNativeTarget "PackedArraySIMDSelfBench" */;
buildPhases = (
1AEF60DA17AD02BD00CA6B64 /* Sources */,
1AEF60DC17AD02BD00CA6B64 /* Frameworks */,
1AEF60DD17AD02BD00CA6B64 /* CopyFiles */,
);
buildRules = (
);
dependencies = (
);
name = PackedArraySIMDSelfBench;
productName = PackedArray;
productReference = 1AEF60E117AD02BD00CA6B64 /* PackedArraySIMDSelfBench */;
productType = "com.apple.product-type.tool";
};
/* End PBXNativeTarget section */
/* Begin PBXProject section */
1A98F9F917A4249200BF09FF /* Project object */ = {
isa = PBXProject;
attributes = {
LastUpgradeCheck = 0460;
ORGANIZATIONNAME = "Gregory Pakosz";
};
buildConfigurationList = 1A98F9FC17A4249200BF09FF /* Build configuration list for PBXProject "PackedArray" */;
compatibilityVersion = "Xcode 3.2";
developmentRegion = English;
hasScannedForEncodings = 0;
knownRegions = (
en,
);
mainGroup = 1A98F9F817A4249200BF09FF;
productRefGroup = 1A98FA0217A4249200BF09FF /* Products */;
projectDirPath = "";
projectRoot = "";
targets = (
1A98FA0017A4249200BF09FF /* PackedArraySelfTest */,
1A98FA1117A4262700BF09FF /* PackedArraySelfBench */,
1AEF60D017AD02BA00CA6B64 /* PackedArraySIMDSelfTest */,
1AEF60D917AD02BD00CA6B64 /* PackedArraySIMDSelfBench */,
);
};
/* End PBXProject section */
/* Begin PBXSourcesBuildPhase section */
1A98F9FD17A4249200BF09FF /* Sources */ = {
isa = PBXSourcesBuildPhase;
buildActionMask = 2147483647;
files = (
1A98FA1017A424D700BF09FF /* PackedArray.c in Sources */,
);
runOnlyForDeploymentPostprocessing = 0;
};
1A98FA1217A4262700BF09FF /* Sources */ = {
isa = PBXSourcesBuildPhase;
buildActionMask = 2147483647;
files = (
1A98FA1317A4262700BF09FF /* PackedArray.c in Sources */,
);
runOnlyForDeploymentPostprocessing = 0;
};
1AEF60D117AD02BA00CA6B64 /* Sources */ = {
isa = PBXSourcesBuildPhase;
buildActionMask = 2147483647;
files = (
1AEF60E217AD035900CA6B64 /* PackedArraySIMD.c in Sources */,
);
runOnlyForDeploymentPostprocessing = 0;
};
1AEF60DA17AD02BD00CA6B64 /* Sources */ = {
isa = PBXSourcesBuildPhase;
buildActionMask = 2147483647;
files = (
1AEF60E317AD035A00CA6B64 /* PackedArraySIMD.c in Sources */,
);
runOnlyForDeploymentPostprocessing = 0;
};
/* End PBXSourcesBuildPhase section */
/* Begin XCBuildConfiguration section */
1A98FA0817A4249200BF09FF /* Debug */ = {
isa = XCBuildConfiguration;
buildSettings = {
ALWAYS_SEARCH_USER_PATHS = NO;
ARCHS = "$(ARCHS_STANDARD_64_BIT)";
CLANG_CXX_LANGUAGE_STANDARD = "gnu++0x";
CLANG_CXX_LIBRARY = "libc++";
CLANG_WARN_CONSTANT_CONVERSION = YES;
CLANG_WARN_EMPTY_BODY = YES;
CLANG_WARN_ENUM_CONVERSION = YES;
CLANG_WARN_INT_CONVERSION = YES;
CLANG_WARN__DUPLICATE_METHOD_MATCH = YES;
COPY_PHASE_STRIP = NO;
GCC_C_LANGUAGE_STANDARD = gnu99;
GCC_DYNAMIC_NO_PIC = NO;
GCC_ENABLE_OBJC_EXCEPTIONS = YES;
GCC_OPTIMIZATION_LEVEL = 0;
GCC_SYMBOLS_PRIVATE_EXTERN = NO;
GCC_WARN_64_TO_32_BIT_CONVERSION = YES;
GCC_WARN_ABOUT_RETURN_TYPE = YES;
GCC_WARN_UNINITIALIZED_AUTOS = YES;
GCC_WARN_UNUSED_VARIABLE = YES;
MACOSX_DEPLOYMENT_TARGET = 10.8;
ONLY_ACTIVE_ARCH = YES;
PRODUCT_NAME = "$(TARGET_NAME)";
SDKROOT = macosx;
};
name = Debug;
};
1A98FA0917A4249200BF09FF /* Release */ = {
isa = XCBuildConfiguration;
buildSettings = {
ALWAYS_SEARCH_USER_PATHS = NO;
ARCHS = "$(ARCHS_STANDARD_64_BIT)";
CLANG_CXX_LANGUAGE_STANDARD = "gnu++0x";
CLANG_CXX_LIBRARY = "libc++";
CLANG_WARN_CONSTANT_CONVERSION = YES;
CLANG_WARN_EMPTY_BODY = YES;
CLANG_WARN_ENUM_CONVERSION = YES;
CLANG_WARN_INT_CONVERSION = YES;
CLANG_WARN__DUPLICATE_METHOD_MATCH = YES;
COPY_PHASE_STRIP = YES;
DEBUG_INFORMATION_FORMAT = "dwarf-with-dsym";
GCC_C_LANGUAGE_STANDARD = gnu99;
GCC_ENABLE_OBJC_EXCEPTIONS = YES;
GCC_OPTIMIZATION_LEVEL = 2;
GCC_PREPROCESSOR_DEFINITIONS = NDEBUG;
GCC_WARN_64_TO_32_BIT_CONVERSION = YES;
GCC_WARN_ABOUT_RETURN_TYPE = YES;
GCC_WARN_UNINITIALIZED_AUTOS = YES;
GCC_WARN_UNUSED_VARIABLE = YES;
MACOSX_DEPLOYMENT_TARGET = 10.8;
PRODUCT_NAME = "$(TARGET_NAME)";
SDKROOT = macosx;
};
name = Release;
};
1A98FA0B17A4249200BF09FF /* Debug */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_TEST,
"$(inherited)",
);
};
name = Debug;
};
1A98FA0C17A4249200BF09FF /* Release */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_TEST,
"$(inherited)",
);
};
name = Release;
};
1A98FA1717A4262700BF09FF /* Debug */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_BENCH,
NDEBUG,
"$(inherited)",
);
};
name = Debug;
};
1A98FA1817A4262700BF09FF /* Release */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_BENCH,
NDEBUG,
"$(inherited)",
);
};
name = Release;
};
1AEF60D617AD02BA00CA6B64 /* Debug */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_TEST,
"$(inherited)",
);
};
name = Debug;
};
1AEF60D717AD02BA00CA6B64 /* Release */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_TEST,
"$(inherited)",
);
};
name = Release;
};
1AEF60DF17AD02BD00CA6B64 /* Debug */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_BENCH,
NDEBUG,
"$(inherited)",
);
};
name = Debug;
};
1AEF60E017AD02BD00CA6B64 /* Release */ = {
isa = XCBuildConfiguration;
buildSettings = {
GCC_PREPROCESSOR_DEFINITIONS = (
PACKEDARRAY_SELF_BENCH,
NDEBUG,
"$(inherited)",
);
};
name = Release;
};
/* End XCBuildConfiguration section */
/* Begin XCConfigurationList section */
1A98F9FC17A4249200BF09FF /* Build configuration list for PBXProject "PackedArray" */ = {
isa = XCConfigurationList;
buildConfigurations = (
1A98FA0817A4249200BF09FF /* Debug */,
1A98FA0917A4249200BF09FF /* Release */,
);
defaultConfigurationIsVisible = 0;
defaultConfigurationName = Release;
};
1A98FA0A17A4249200BF09FF /* Build configuration list for PBXNativeTarget "PackedArraySelfTest" */ = {
isa = XCConfigurationList;
buildConfigurations = (
1A98FA0B17A4249200BF09FF /* Debug */,
1A98FA0C17A4249200BF09FF /* Release */,
);
defaultConfigurationIsVisible = 0;
defaultConfigurationName = Release;
};
1A98FA1617A4262700BF09FF /* Build configuration list for PBXNativeTarget "PackedArraySelfBench" */ = {
isa = XCConfigurationList;
buildConfigurations = (
1A98FA1717A4262700BF09FF /* Debug */,
1A98FA1817A4262700BF09FF /* Release */,
);
defaultConfigurationIsVisible = 0;
defaultConfigurationName = Release;
};
1AEF60D517AD02BA00CA6B64 /* Build configuration list for PBXNativeTarget "PackedArraySIMDSelfTest" */ = {
isa = XCConfigurationList;
buildConfigurations = (
1AEF60D617AD02BA00CA6B64 /* Debug */,
1AEF60D717AD02BA00CA6B64 /* Release */,
);
defaultConfigurationIsVisible = 0;
defaultConfigurationName = Release;
};
1AEF60DE17AD02BD00CA6B64 /* Build configuration list for PBXNativeTarget "PackedArraySIMDSelfBench" */ = {
isa = XCConfigurationList;
buildConfigurations = (
1AEF60DF17AD02BD00CA6B64 /* Debug */,
1AEF60E017AD02BD00CA6B64 /* Release */,
);
defaultConfigurationIsVisible = 0;
defaultConfigurationName = Release;
};
/* End XCConfigurationList section */
};
rootObject = 1A98F9F917A4249200BF09FF /* Project object */;
}
================================================
FILE: _mac-xcode/PackedArray.xcodeproj/project.xcworkspace/contents.xcworkspacedata
================================================
<?xml version="1.0" encoding="UTF-8"?>
<Workspace
version = "1.0">
<FileRef
location = "self:PackedArray.xcodeproj">
</FileRef>
</Workspace>
================================================
FILE: _win-vs11/.gitignore
================================================
/build
/ipch
*.suo
*.sdf
*.opensdf
*.user
*.sln.docstates
================================================
FILE: _win-vs11/Common.props
================================================
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup Label="UserMacros">
<BIN_DIR>..\bin\Windows$(PLATFORM_SUFFIX)-$(ARCH)$(CONF_SUFFIX)</BIN_DIR>
<LIB_DIR>..\lib\Windows$(PLATFORM_SUFFIX)-$(ARCH)-vs11$(CONF_SUFFIX)</LIB_DIR>
<BUILD_DIR>build\$(ProjectName)-win$(PLATFORM_SUFFIX)-$(ARCH)-vs11$(CONF_SUFFIX)</BUILD_DIR>
</PropertyGroup>
<PropertyGroup>
<IntDir>$(BUILD_DIR)\</IntDir>
<OutDir>$(BIN_DIR)\</OutDir>
<LinkIncremental>false</LinkIncremental>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<PreprocessorDefinitions>_CRT_SECURE_NO_DEPRECATE;_CRT_NONSTDC_NO_DEPRECATE;_SCL_SECURE_NO_WARNINGS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<MultiProcessorCompilation>true</MultiProcessorCompilation>
<AdditionalOptions>/we4013 /wd4820 /we4289 /wd4342 /wd4347 /wd4514 /we4545 /we4546 /we4547 /we4548 /we4549 /we4619 /we4623 /we4625 /we4626 /wd4710 /we4836 /we4905 /we4906 /we4928 /we4946 /wd4986 /wd4711 /wd4350</AdditionalOptions>
<PrecompiledHeader>NotUsing</PrecompiledHeader>
<FunctionLevelLinking>true</FunctionLevelLinking>
<WarningLevel>EnableAllWarnings</WarningLevel>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
<SubSystem>Console</SubSystem>
<AdditionalOptions>/time %(AdditionalOptions)</AdditionalOptions>
</Link>
</ItemDefinitionGroup>
<ItemGroup>
<BuildMacro Include="BIN_DIR">
<Value>$(BIN_DIR)</Value>
</BuildMacro>
<BuildMacro Include="LIB_DIR">
<Value>$(LIB_DIR)</Value>
</BuildMacro>
<BuildMacro Include="BUILD_DIR">
<Value>$(BUILD_DIR)</Value>
</BuildMacro>
</ItemGroup>
</Project>
================================================
FILE: _win-vs11/Debug.props
================================================
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup Label="UserMacros">
<CONF_SUFFIX>-Debug</CONF_SUFFIX>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<Optimization>Disabled</Optimization>
<WarningLevel>Level3</WarningLevel>
<DebugInformationFormat>ProgramDatabase</DebugInformationFormat>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
<IntrinsicFunctions>true</IntrinsicFunctions>
<InlineFunctionExpansion>OnlyExplicitInline</InlineFunctionExpansion>
</ClCompile>
</ItemDefinitionGroup>
<ItemGroup>
<BuildMacro Include="CONF_SUFFIX">
<Value>$(CONF_SUFFIX)</Value>
</BuildMacro>
</ItemGroup>
</Project>
================================================
FILE: _win-vs11/PackedArray.sln
================================================
Microsoft Visual Studio Solution File, Format Version 12.00
# Visual Studio Express 2012 for Windows Desktop
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "PackedArraySelfTest", "PackedArraySelfTest.vcxproj", "{7BC8C873-A2C7-43A1-BD8A-2F1731F3CB81}"
EndProject
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "PackedArraySelfBench", "PackedArraySelfBench.vcxproj", "{686B991E-01AD-4433-897E-DFD5E751DAF5}"
EndProject
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "PackedArraySIMDSelfBench", "PackedArraySIMDSelfBench.vcxproj", "{6082A77F-4ADC-4E49-89BA-BFE1C9E69C9D}"
EndProject
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "PackedArraySIMDSelfTest", "PackedArraySIMDSelfTest.vcxproj", "{2F15B46B-D5F3-46B4-A5A2-ECD62B9FD848}"
EndProject
Global
GlobalSection(SolutionConfigurationPlatforms) = preSolution
Debug|Win32 = Debug|Win32
Debug|x64 = Debug|x64
Release|Win32 = Release|Win32
Release|x64 = Release|x64
EndGlobalSection
GlobalSection(ProjectConfigurationPlatforms) = postSolution
{7BC8C873-A2C7-43A1-BD8A-2F1731F3CB81}.Debug|Win32.ActiveCfg = Debug|Win32
{7BC8C873-A2C7-43A1-BD8A-2F1731F3CB81}.Debug|Win32.Build.0 = Debug|Win32
{7BC8C873-A2C7-43A1-BD8A-2F1731F3CB81}.Debug|x64.ActiveCfg = Debug|x64
{7BC8C873-A2C7-43A1-BD8A-2F1731F3CB81}.Debug|x64.Build.0 = Debug|x64
{7BC8C873-A2C7-43A1-BD8A-2F1731F3CB81}.Release|Win32.ActiveCfg = Release|Win32
{7BC8C873-A2C7-43A1-BD8A-2F1731F3CB81}.Release|Win32.Build.0 = Release|Win32
{7BC8C873-A2C7-43A1-BD8A-2F1731F3CB81}.Release|x64.ActiveCfg = Release|x64
{7BC8C873-A2C7-43A1-BD8A-2F1731F3CB81}.Release|x64.Build.0 = Release|x64
{686B991E-01AD-4433-897E-DFD5E751DAF5}.Debug|Win32.ActiveCfg = Debug|Win32
{686B991E-01AD-4433-897E-DFD5E751DAF5}.Debug|Win32.Build.0 = Debug|Win32
{686B991E-01AD-4433-897E-DFD5E751DAF5}.Debug|x64.ActiveCfg = Debug|x64
{686B991E-01AD-4433-897E-DFD5E751DAF5}.Debug|x64.Build.0 = Debug|x64
{686B991E-01AD-4433-897E-DFD5E751DAF5}.Release|Win32.ActiveCfg = Release|Win32
{686B991E-01AD-4433-897E-DFD5E751DAF5}.Release|Win32.Build.0 = Release|Win32
{686B991E-01AD-4433-897E-DFD5E751DAF5}.Release|x64.ActiveCfg = Release|x64
{686B991E-01AD-4433-897E-DFD5E751DAF5}.Release|x64.Build.0 = Release|x64
{6082A77F-4ADC-4E49-89BA-BFE1C9E69C9D}.Debug|Win32.ActiveCfg = Debug|Win32
{6082A77F-4ADC-4E49-89BA-BFE1C9E69C9D}.Debug|Win32.Build.0 = Debug|Win32
{6082A77F-4ADC-4E49-89BA-BFE1C9E69C9D}.Debug|x64.ActiveCfg = Debug|x64
{6082A77F-4ADC-4E49-89BA-BFE1C9E69C9D}.Debug|x64.Build.0 = Debug|x64
{6082A77F-4ADC-4E49-89BA-BFE1C9E69C9D}.Release|Win32.ActiveCfg = Release|Win32
{6082A77F-4ADC-4E49-89BA-BFE1C9E69C9D}.Release|Win32.Build.0 = Release|Win32
{6082A77F-4ADC-4E49-89BA-BFE1C9E69C9D}.Release|x64.ActiveCfg = Release|x64
{6082A77F-4ADC-4E49-89BA-BFE1C9E69C9D}.Release|x64.Build.0 = Release|x64
{2F15B46B-D5F3-46B4-A5A2-ECD62B9FD848}.Debug|Win32.ActiveCfg = Debug|Win32
{2F15B46B-D5F3-46B4-A5A2-ECD62B9FD848}.Debug|Win32.Build.0 = Debug|Win32
{2F15B46B-D5F3-46B4-A5A2-ECD62B9FD848}.Debug|x64.ActiveCfg = Debug|x64
{2F15B46B-D5F3-46B4-A5A2-ECD62B9FD848}.Debug|x64.Build.0 = Debug|x64
{2F15B46B-D5F3-46B4-A5A2-ECD62B9FD848}.Release|Win32.ActiveCfg = Release|Win32
{2F15B46B-D5F3-46B4-A5A2-ECD62B9FD848}.Release|Win32.Build.0 = Release|Win32
{2F15B46B-D5F3-46B4-A5A2-ECD62B9FD848}.Release|x64.ActiveCfg = Release|x64
{2F15B46B-D5F3-46B4-A5A2-ECD62B9FD848}.Release|x64.Build.0 = Release|x64
EndGlobalSection
GlobalSection(SolutionProperties) = preSolution
HideSolutionNode = FALSE
EndGlobalSection
EndGlobal
================================================
FILE: _win-vs11/PackedArraySIMDSelfBench.vcxproj
================================================
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|Win32">
<Configuration>Debug</Configuration>
<Platform>Win32</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|Win32">
<Configuration>Release</Configuration>
<Platform>Win32</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{6082A77F-4ADC-4E49-89BA-BFE1C9E69C9D}</ProjectGuid>
<Keyword>Win32Proj</Keyword>
<RootNamespace>PackedArraySIMD</RootNamespace>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>true</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>true</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>false</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>false</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
</ImportGroup>
<ImportGroup Label="PropertySheets" Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x86.props" />
<Import Project="Debug.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" Label="PropertySheets">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x64.props" />
<Import Project="Debug.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets" Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x86.props" />
<Import Project="Release.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" Label="PropertySheets">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x64.props" />
<Import Project="Release.props" />
<Import Project="Common.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" />
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_BENCH;NDEBUG;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_BENCH;NDEBUG;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_BENCH;NDEBUG%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_BENCH;NDEBUG%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemGroup>
<ClCompile Include="..\PackedArraySIMD.c" />
</ItemGroup>
<ItemGroup>
<ClInclude Include="..\PackedArray.h" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
</ImportGroup>
</Project>
================================================
FILE: _win-vs11/PackedArraySIMDSelfTest.vcxproj
================================================
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|Win32">
<Configuration>Debug</Configuration>
<Platform>Win32</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|Win32">
<Configuration>Release</Configuration>
<Platform>Win32</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{2F15B46B-D5F3-46B4-A5A2-ECD62B9FD848}</ProjectGuid>
<Keyword>Win32Proj</Keyword>
<RootNamespace>PackedArraySIMD</RootNamespace>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>true</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>true</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>false</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>false</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
</ImportGroup>
<ImportGroup Label="PropertySheets" Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x86.props" />
<Import Project="Debug.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" Label="PropertySheets">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x64.props" />
<Import Project="Debug.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets" Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x86.props" />
<Import Project="Release.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" Label="PropertySheets">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x64.props" />
<Import Project="Release.props" />
<Import Project="Common.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" />
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_TEST;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_TEST;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_TEST;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_TEST;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemGroup>
<ClCompile Include="..\PackedArraySIMD.c" />
</ItemGroup>
<ItemGroup>
<ClInclude Include="..\PackedArray.h" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
</ImportGroup>
</Project>
================================================
FILE: _win-vs11/PackedArraySelfBench.vcxproj
================================================
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|Win32">
<Configuration>Debug</Configuration>
<Platform>Win32</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|Win32">
<Configuration>Release</Configuration>
<Platform>Win32</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{686B991E-01AD-4433-897E-DFD5E751DAF5}</ProjectGuid>
<Keyword>Win32Proj</Keyword>
<RootNamespace>PackedArray</RootNamespace>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>true</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>true</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>false</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>false</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
</ImportGroup>
<ImportGroup Label="PropertySheets" Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x86.props" />
<Import Project="Debug.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" Label="PropertySheets">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x64.props" />
<Import Project="Debug.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets" Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x86.props" />
<Import Project="Release.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" Label="PropertySheets">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x64.props" />
<Import Project="Release.props" />
<Import Project="Common.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" />
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_BENCH;NDEBUG;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_BENCH;NDEBUG;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_BENCH;NDEBUG%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_BENCH;NDEBUG%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemGroup>
<ClCompile Include="..\PackedArray.c" />
</ItemGroup>
<ItemGroup>
<ClInclude Include="..\PackedArray.h" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
</ImportGroup>
</Project>
================================================
FILE: _win-vs11/PackedArraySelfTest.vcxproj
================================================
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<ItemGroup Label="ProjectConfigurations">
<ProjectConfiguration Include="Debug|Win32">
<Configuration>Debug</Configuration>
<Platform>Win32</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Debug|x64">
<Configuration>Debug</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|Win32">
<Configuration>Release</Configuration>
<Platform>Win32</Platform>
</ProjectConfiguration>
<ProjectConfiguration Include="Release|x64">
<Configuration>Release</Configuration>
<Platform>x64</Platform>
</ProjectConfiguration>
</ItemGroup>
<PropertyGroup Label="Globals">
<ProjectGuid>{7BC8C873-A2C7-43A1-BD8A-2F1731F3CB81}</ProjectGuid>
<Keyword>Win32Proj</Keyword>
<RootNamespace>PackedArray</RootNamespace>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>true</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>true</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>false</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>false</UseDebugLibraries>
<PlatformToolset>v110</PlatformToolset>
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
</ImportGroup>
<ImportGroup Label="PropertySheets" Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x86.props" />
<Import Project="Debug.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" Label="PropertySheets">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x64.props" />
<Import Project="Debug.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets" Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x86.props" />
<Import Project="Release.props" />
<Import Project="Common.props" />
</ImportGroup>
<ImportGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" Label="PropertySheets">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
<Import Project="x64.props" />
<Import Project="Release.props" />
<Import Project="Common.props" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" />
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_TEST;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_TEST;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_TEST;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'">
<ClCompile>
<PreprocessorDefinitions>PACKEDARRAY_SELF_TEST;%(PreprocessorDefinitions)</PreprocessorDefinitions>
</ClCompile>
</ItemDefinitionGroup>
<ItemGroup>
<ClCompile Include="..\PackedArray.c" />
</ItemGroup>
<ItemGroup>
<ClInclude Include="..\PackedArray.h" />
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
</ImportGroup>
</Project>
================================================
FILE: _win-vs11/Release.props
================================================
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup Label="UserMacros">
<CONF_SUFFIX />
</PropertyGroup>
<PropertyGroup>
<LinkIncremental>false</LinkIncremental>
</PropertyGroup>
<ItemDefinitionGroup>
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<InlineFunctionExpansion>AnySuitable</InlineFunctionExpansion>
<IntrinsicFunctions>true</IntrinsicFunctions>
<FavorSizeOrSpeed>Speed</FavorSizeOrSpeed>
<OmitFramePointers>true</OmitFramePointers>
<EnableFiberSafeOptimizations>true</EnableFiberSafeOptimizations>
<WholeProgramOptimization>true</WholeProgramOptimization>
<PreprocessorDefinitions>NDEBUG;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<StringPooling>true</StringPooling>
<WarningLevel>Level3</WarningLevel>
<DebugInformationFormat>ProgramDatabase</DebugInformationFormat>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
<BufferSecurityCheck>false</BufferSecurityCheck>
<AdditionalOptions>/d2Zi+ %(AdditionalOptions)</AdditionalOptions>
</ClCompile>
<Link>
<LinkTimeCodeGeneration>UseLinkTimeCodeGeneration</LinkTimeCodeGeneration>
<EnableCOMDATFolding>true</EnableCOMDATFolding>
</Link>
<Lib>
<LinkTimeCodeGeneration>true</LinkTimeCodeGeneration>
</Lib>
</ItemDefinitionGroup>
<ItemGroup>
<BuildMacro Include="CONF_SUFFIX">
<Value>$(CONF_SUFFIX)</Value>
</BuildMacro>
</ItemGroup>
</Project>
================================================
FILE: _win-vs11/x64.props
================================================
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup Label="UserMacros">
<ARCH>x64</ARCH>
</PropertyGroup>
<ItemDefinitionGroup>
<Link>
<TargetMachine>MachineX64</TargetMachine>
</Link>
</ItemDefinitionGroup>
<ItemGroup>
<BuildMacro Include="ARCH">
<Value>$(ARCH)</Value>
</BuildMacro>
</ItemGroup>
</Project>
================================================
FILE: _win-vs11/x86.props
================================================
<?xml version="1.0" encoding="utf-8"?>
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<PropertyGroup Label="UserMacros">
<ARCH>x86</ARCH>
</PropertyGroup>
<ItemDefinitionGroup>
<Link>
<TargetMachine>MachineX86</TargetMachine>
</Link>
<ClCompile>
<EnableEnhancedInstructionSet>StreamingSIMDExtensions2</EnableEnhancedInstructionSet>
</ClCompile>
</ItemDefinitionGroup>
<ItemGroup>
<BuildMacro Include="ARCH">
<Value>$(ARCH)</Value>
</BuildMacro>
</ItemGroup>
</Project>
================================================
FILE: benchmark/PackedArraySIMDSelfBench-unrolled-galaxy-note-cortex-a9-1.4GHz.txt
================================================
-- PackedArray self bench ------------------------------------------------------
memcpy:
bits size (B) time (µs) speed (B/µs)
32 4 0.260 15.392
32 8 0.225 35.583
32 16 0.223 71.774
32 32 0.363 88.127
32 64 0.424 150.976
32 128 0.509 251.462
32 256 0.767 333.771
32 512 1.049 488.175
32 1024 1.824 561.434
32 2048 3.328 615.413
32 4096 6.297 650.482
32 8192 12.142 674.686
32 16384 23.797 688.489
32 32768 10.177 3219.842
32 65536 27.135 2415.171
32 131072 55.138 2377.168
32 262144 132.314 1981.227
32 524288 351.372 1492.117
32 1048576 1400.464 748.735
avg (B/µs) min (B/µs) max (B/µs)
887.370 15.392 3219.842
loopcpy:
bits size (B) time (µs) speed (B/µs)
32 4 0.037 108.240
32 8 0.038 209.715
32 16 0.038 422.068
32 32 0.063 506.482
32 64 0.071 900.790
32 128 0.086 1491.308
32 256 0.116 2209.345
32 512 0.176 2909.869
32 1024 0.306 3344.990
32 2048 0.546 3751.063
32 4096 1.026 3991.605
32 8192 2.497 3280.792
32 16384 4.508 3634.413
32 32768 8.462 3872.392
32 65536 27.297 2400.848
32 131072 55.625 2356.342
32 262144 185.746 1411.304
32 524288 630.057 832.128
32 1048576 1377.527 761.202
avg (B/µs) min (B/µs) max (B/µs)
2020.784 108.240 3991.605
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
1 4 0.038 105.517 1 4 0.062 64.528
1 8 0.044 182.361 1 8 0.039 204.600
1 16 0.052 307.839 1 16 0.039 409.200
1 32 0.054 591.268 1 32 0.051 627.186
1 64 0.060 1065.220 1 64 0.065 986.895
1 128 0.074 1731.842 1 128 0.095 1345.541
1 256 0.106 2412.903 1 256 0.155 1651.910
1 512 0.166 3081.038 1 512 0.277 1848.093
1 1024 0.303 3379.203 1 1024 0.517 1980.160
1 2048 0.579 3537.864 1 2048 1.001 2046.197
1 4096 1.187 3450.466 1 4096 1.969 2080.139
1 8192 2.261 3623.298 1 8192 3.902 2099.459
1 16384 4.463 3670.912 1 16384 7.797 2101.320
1 32768 12.762 2567.609 1 32768 15.681 2089.659
1 65536 31.981 2049.217 1 65536 31.053 2110.468
1 131072 71.963 1821.379 1 131072 62.428 2099.571
1 262144 151.636 1728.770 1 262144 124.880 2099.166
1 524288 309.744 1692.649 1 524288 252.277 2078.224
1 1048576 773.824 1355.058 1 1048576 533.997 1963.636
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
2018.653 105.517 3670.912 1572.945 64.528 2110.468
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
2 4 0.039 102.928 2 4 0.067 59.705
2 8 0.049 162.886 2 8 0.043 185.384
2 16 0.054 296.942 2 16 0.043 372.827
2 32 0.057 561.580 2 32 0.050 642.190
2 64 0.063 1016.801 2 64 0.065 983.280
2 128 0.076 1682.981 2 128 0.095 1345.541
2 256 0.103 2485.513 2 256 0.155 1651.910
2 512 0.176 2909.869 2 512 0.279 1833.889
2 1024 0.316 3239.040 2 1024 0.521 1965.660
2 2048 0.637 3215.999 2 2048 1.007 2033.602
2 4096 1.179 3474.190 2 4096 1.987 2061.419
2 8192 2.334 3510.036 2 8192 3.953 2072.361
2 16384 4.602 3560.041 2 16384 7.857 2085.315
2 32768 13.053 2510.392 2 32768 15.936 2056.238
2 65536 34.460 1901.795 2 65536 31.283 2094.946
2 131072 73.934 1772.823 2 131072 62.702 2090.398
2 262144 149.351 1755.222 2 262144 133.610 1962.009
2 524288 432.895 1211.121 2 524288 275.559 1902.635
2 1048576 888.796 1179.771 2 1048576 556.340 1884.776
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1923.681 102.928 3560.041 1541.268 59.705 2094.946
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
3 4 0.042 95.325 3 4 0.070 57.260
3 8 0.054 147.817 3 8 0.047 170.327
3 16 0.056 285.570 3 16 0.039 409.200
3 32 0.058 550.073 3 32 0.051 627.186
3 64 0.064 1001.625 3 64 0.066 972.592
3 128 0.078 1636.802 3 128 0.096 1335.500
3 256 0.113 2265.278 3 256 0.165 1551.650
3 512 0.177 2890.288 3 512 0.296 1730.446
3 1024 0.322 3181.457 3 1024 0.556 1841.753
3 2048 0.614 3335.897 3 2048 1.076 1903.376
3 4096 1.248 3281.732 3 4096 2.116 1935.760
3 8192 2.378 3444.931 3 8192 4.213 1944.414
3 16384 4.734 3460.892 3 16384 8.385 1953.979
3 32768 13.400 2445.402 3 32768 16.724 1959.327
3 65536 34.637 1892.082 3 65536 33.432 1960.277
3 131072 77.480 1691.686 3 131072 67.093 1953.590
3 262144 173.980 1506.748 3 262144 136.542 1919.877
3 524288 337.184 1554.902 3 524288 287.618 1822.861
3 1048576 817.251 1283.052 3 1048576 647.711 1618.895
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1892.187 95.325 3460.892 1456.225 57.260 1960.277
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
4 4 0.083 48.210 4 4 0.034 117.323
4 8 0.051 156.796 4 8 0.044 182.361
4 16 0.057 279.620 4 16 0.040 399.458
4 32 0.058 552.336 4 32 0.049 651.542
4 64 0.065 983.280 4 64 0.063 1016.801
4 128 0.067 1910.573 4 128 0.094 1362.617
4 256 0.105 2440.322 4 256 0.155 1651.910
4 512 0.180 2844.349 4 512 0.281 1821.445
4 1024 0.330 3103.300 4 1024 0.526 1946.066
4 2048 0.630 3250.070 4 2048 1.071 1912.274
4 4096 1.245 3289.902 4 4096 2.081 1968.137
4 8192 2.497 3280.792 4 8192 4.149 1974.471
4 16384 4.888 3352.006 4 16384 8.034 2039.335
4 32768 14.355 2282.698 4 32768 16.054 2041.122
4 65536 37.628 1741.685 4 65536 45.911 1427.463
4 131072 91.421 1433.720 4 131072 70.325 1863.800
4 262144 176.782 1482.865 4 262144 151.455 1730.838
4 524288 377.989 1387.046 4 524288 287.518 1823.496
4 1048576 864.127 1213.451 4 1048576 601.741 1742.570
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1843.843 48.210 3352.006 1456.475 117.323 2041.122
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
5 4 0.082 48.771 5 4 0.040 99.864
5 8 0.057 140.395 5 8 0.053 150.468
5 16 0.055 291.778 5 16 0.038 422.068
5 32 0.058 552.336 5 32 0.051 630.130
5 64 0.085 754.032 5 64 0.066 969.081
5 128 0.083 1542.733 5 128 0.103 1242.757
5 256 0.116 2204.809 5 256 0.174 1470.879
5 512 0.185 2767.376 5 512 0.315 1625.650
5 1024 0.339 3020.371 5 1024 0.593 1726.967
5 2048 0.648 3160.388 5 2048 1.191 1719.707
5 4096 1.278 3204.602 5 4096 2.262 1810.694
5 8192 2.593 3159.516 5 8192 4.507 1817.591
5 16384 5.010 3270.176 5 16384 8.975 1825.558
5 32768 14.040 2333.904 5 32768 17.914 1829.202
5 65536 33.340 1965.688 5 65536 36.246 1808.086
5 131072 73.165 1791.459 5 131072 71.988 1820.751
5 262144 165.369 1585.208 5 262144 154.613 1695.485
5 524288 323.975 1618.297 5 524288 333.541 1571.885
5 1048576 975.722 1074.667 5 1048576 914.320 1146.837
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1815.079 48.771 3270.176 1335.982 99.864 1829.202
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
6 4 0.108 37.036 6 4 0.040 100.462
6 8 0.136 58.867 6 8 0.055 145.257
6 16 0.057 280.790 6 16 0.040 399.458
6 32 0.059 543.392 6 32 0.051 627.186
6 64 0.066 969.081 6 64 0.066 969.081
6 128 0.080 1597.830 6 128 0.106 1206.451
6 256 0.112 2284.557 6 256 0.174 1470.879
6 512 0.191 2681.003 6 512 0.316 1619.520
6 1024 0.349 2933.721 6 1024 0.596 1717.987
6 2048 0.665 3079.934 6 2048 1.155 1772.948
6 4096 3.067 1335.500 6 4096 2.279 1797.245
6 8192 2.619 3127.878 6 8192 4.562 1795.743
6 16384 5.137 3189.431 6 16384 10.831 1512.712
6 32768 16.548 1980.189 6 32768 21.795 1503.462
6 65536 42.263 1550.670 6 65536 38.935 1683.218
6 131072 87.775 1493.269 6 131072 73.146 1791.926
6 262144 165.623 1582.774 6 262144 163.291 1605.380
6 524288 328.345 1596.759 6 524288 372.375 1407.957
6 1048576 966.395 1085.039 6 1048576 723.185 1449.942
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1653.038 37.036 3189.431 1293.517 100.462 1797.245
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
7 4 0.079 50.534 7 4 0.050 80.274
7 8 0.057 140.395 7 8 0.055 145.257
7 16 0.056 285.570 7 16 0.038 419.430
7 32 0.059 541.201 7 32 0.051 627.186
7 64 0.065 983.280 7 64 0.066 969.081
7 128 0.080 1597.830 7 128 0.104 1231.355
7 256 0.121 2113.665 7 256 0.183 1398.101
7 512 0.194 2638.186 7 512 0.332 1541.625
7 1024 0.356 2876.736 7 1024 0.676 1514.980
7 2048 0.681 3006.627 7 2048 1.222 1676.085
7 4096 1.343 3049.861 7 4096 2.425 1689.103
7 8192 2.644 3098.263 7 8192 4.798 1707.316
7 16384 5.262 3113.564 7 16384 9.565 1712.934
7 32768 14.529 2255.353 7 32768 19.097 1715.885
7 65536 35.618 1839.965 7 65536 38.983 1681.138
7 131072 82.084 1596.800 7 131072 80.429 1629.659
7 262144 163.769 1600.694 7 262144 173.056 1514.792
7 524288 369.750 1417.953 7 524288 422.401 1241.209
7 1048576 1355.063 773.821 7 1048576 827.127 1267.733
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1735.805 50.534 3113.564 1250.692 80.274 1715.885
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
8 4 0.042 95.325 8 4 0.075 53.261
8 8 0.052 153.919 8 8 0.045 177.537
8 16 0.054 296.942 8 16 0.039 411.711
8 32 0.056 571.139 8 32 0.049 651.542
8 64 0.057 1123.161 8 64 0.063 1012.964
8 128 0.073 1754.480 8 128 0.095 1345.541
8 256 0.114 2246.322 8 256 0.159 1609.808
8 512 0.198 2587.330 8 512 0.286 1789.570
8 1024 0.364 2812.683 8 1024 0.539 1899.587
8 2048 0.697 2938.739 8 2048 1.106 1851.678
8 4096 1.376 2976.931 8 4096 2.062 1986.342
8 8192 2.726 3005.050 8 8192 4.112 1992.215
8 16384 5.475 2992.487 8 16384 8.238 1988.814
8 32768 15.495 2114.738 8 32768 16.305 2009.694
8 65536 38.939 1683.043 8 65536 33.180 1975.166
8 131072 85.077 1540.627 8 131072 73.070 1793.786
8 262144 178.622 1467.591 8 262144 203.704 1286.888
8 524288 380.350 1378.435 8 524288 325.536 1610.538
8 1048576 1532.529 684.213 8 1048576 1173.707 893.388
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1706.482 95.325 3005.050 1386.317 53.261 2009.694
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
9 4 0.065 61.455 9 4 0.110 36.393
9 8 0.086 92.949 9 8 0.084 95.325
9 16 0.082 195.653 9 16 0.060 266.305
9 32 0.086 372.827 9 32 0.074 432.960
9 64 0.106 603.226 9 64 0.101 633.102
9 128 0.137 935.315 9 128 0.161 795.364
9 256 0.194 1319.093 9 256 0.286 895.531
9 512 0.361 1418.417 9 512 0.526 973.474
9 1024 0.564 1815.286 9 1024 0.997 1027.258
9 2048 1.077 1901.690 9 2048 1.962 1043.861
9 4096 2.116 1935.760 9 4096 3.860 1061.141
9 8192 4.684 1748.943 9 8192 7.636 1072.803
9 16384 8.370 1957.429 9 16384 15.304 1070.564
9 32768 23.168 1414.359 9 32768 30.751 1065.592
9 65536 55.319 1184.691 9 65536 63.291 1035.470
9 131072 121.022 1083.043 9 131072 146.014 897.667
9 262144 263.057 996.529 9 262144 291.697 898.686
9 524288 544.491 962.896 9 524288 587.299 892.711
9 1048576 1402.228 747.793 9 1048576 1429.417 733.569
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1091.966 61.455 1957.429 785.672 36.393 1072.803
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
10 4 0.105 38.130 10 4 0.059 67.924
10 8 0.086 92.949 10 8 0.083 96.145
10 16 0.083 192.842 10 16 0.059 270.600
10 32 0.087 367.720 10 32 0.076 420.745
10 64 0.108 592.573 10 64 0.102 628.654
10 128 0.127 1007.262 10 128 0.170 752.975
10 256 0.181 1414.680 10 256 0.290 883.011
10 512 0.311 1645.581 10 512 0.529 967.771
10 1024 0.572 1790.316 10 1024 1.066 960.628
10 2048 1.650 1241.141 10 2048 1.955 1047.553
10 4096 2.206 1856.882 10 4096 3.874 1057.288
10 8192 4.287 1910.891 10 8192 7.675 1067.371
10 16384 8.569 1912.008 10 16384 15.414 1062.930
10 32768 23.449 1397.405 10 32768 30.574 1071.757
10 65536 58.764 1115.236 10 65536 62.770 1044.067
10 131072 125.489 1044.490 10 131072 136.313 961.552
10 262144 181.845 1441.580 10 262144 200.107 1310.020
10 524288 452.774 1157.947 10 524288 420.063 1248.117
10 1048576 1170.650 895.721 10 1048576 919.921 1139.854
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1111.334 38.130 1912.008 845.209 67.924 1310.020
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
11 4 0.079 50.686 11 4 0.049 81.443
11 8 0.057 140.395 11 8 0.054 148.471
11 16 0.057 279.620 11 16 0.039 411.711
11 32 0.059 543.392 11 32 0.052 615.678
11 64 0.068 941.879 11 64 0.075 852.176
11 128 0.084 1520.881 11 128 0.117 1093.423
11 256 0.133 1924.269 11 256 0.202 1267.700
11 512 0.216 2370.291 11 512 0.371 1380.131
11 1024 0.397 2579.560 11 1024 0.703 1456.415
11 2048 0.755 2712.325 11 2048 1.424 1438.368
11 4096 1.558 2628.901 11 4096 2.733 1498.724
11 8192 2.911 2814.065 11 8192 5.391 1519.536
11 16384 5.844 2803.618 11 16384 11.170 1466.798
11 32768 18.177 1802.714 11 32768 24.650 1329.325
11 65536 41.731 1570.444 11 65536 54.228 1208.525
11 131072 92.313 1419.864 11 131072 112.020 1170.077
11 262144 205.082 1278.240 11 262144 231.958 1130.135
11 524288 372.994 1405.621 11 524288 490.544 1068.789
11 1048576 1569.861 667.942 11 1048576 1417.162 739.913
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1550.248 50.686 2814.065 1046.176 81.443 1519.536
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
12 4 0.065 61.681 12 4 0.079 50.686
12 8 0.086 92.949 12 8 0.084 95.325
12 16 0.083 192.842 12 16 0.058 276.168
12 32 0.087 367.720 12 32 0.077 415.535
12 64 0.100 639.132 12 64 0.111 576.042
12 128 0.127 1007.262 12 128 0.174 735.440
12 256 0.195 1312.643 12 256 0.295 868.021
12 512 0.335 1528.458 12 512 0.786 651.345
12 1024 0.662 1547.178 12 1024 1.016 1007.972
12 2048 1.160 1765.297 12 2048 2.951 694.024
12 4096 2.268 1806.126 12 4096 4.365 938.380
12 8192 4.829 1696.442 12 8192 7.971 1027.719
12 16384 9.323 1757.397 12 16384 16.783 976.226
12 32768 26.191 1251.117 12 32768 33.979 964.362
12 65536 65.279 1003.937 12 65536 80.837 810.718
12 131072 140.986 929.681 12 131072 162.005 809.061
12 262144 278.053 942.784 12 262144 317.797 824.879
12 524288 599.059 875.186 12 524288 643.090 815.264
12 1048576 1606.330 652.777 12 1048576 995.421 1053.400
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1022.664 61.681 1806.126 715.293 50.686 1053.400
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
13 4 0.043 93.207 13 4 0.077 51.942
13 8 0.057 139.810 13 8 0.056 142.785
13 16 0.056 285.570 13 16 0.040 399.458
13 32 0.060 532.610 13 32 0.052 615.678
13 64 0.068 941.879 13 64 0.075 852.176
13 128 0.097 1319.093 13 128 0.119 1075.894
13 256 0.136 1883.758 13 256 0.211 1213.268
13 512 0.225 2274.877 13 512 0.390 1312.643
13 1024 0.413 2479.773 13 1024 0.800 1280.169
13 2048 0.789 2595.146 13 2048 1.444 1418.183
13 4096 1.546 2649.579 13 4096 2.866 1429.155
13 8192 3.095 2646.721 13 8192 5.691 1439.453
13 16384 6.112 2680.690 13 16384 12.151 1348.366
13 32768 17.636 1858.011 13 32768 28.618 1145.019
13 65536 39.021 1679.505 13 65536 61.931 1058.212
13 131072 122.128 1073.235 13 131072 116.900 1121.230
13 262144 204.158 1284.025 13 262144 252.843 1036.785
13 524288 411.603 1273.770 13 524288 494.814 1059.566
13 1048576 1636.328 640.810 13 1048576 1534.200 683.468
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1491.162 93.207 2680.690 983.339 51.942 1439.453
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
14 4 0.099 40.427 14 4 0.059 67.650
14 8 0.085 93.990 14 8 0.083 96.421
14 16 0.087 183.860 14 16 0.062 258.111
14 32 0.091 351.355 14 32 0.077 415.535
14 64 0.105 610.081 14 64 0.111 576.042
14 128 0.136 940.229 14 128 0.183 699.962
14 256 0.200 1279.788 14 256 0.321 797.136
14 512 0.347 1474.920 14 512 0.586 874.027
14 1024 0.634 1615.257 14 1024 1.166 878.316
14 2048 1.210 1692.598 14 2048 2.175 941.672
14 4096 2.395 1710.291 14 4096 5.866 698.255
14 8192 4.798 1707.401 14 8192 8.873 923.252
14 16384 9.408 1741.497 14 16384 17.200 952.559
14 32768 25.395 1290.325 14 32768 36.607 895.129
14 65536 62.286 1052.180 14 65536 81.975 799.463
14 131072 131.350 997.883 14 131072 196.949 665.512
14 262144 338.609 774.179 14 262144 375.820 697.525
14 524288 567.484 923.882 14 524288 705.064 743.603
14 1048576 1573.973 666.197 14 1048576 1619.872 647.320
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1007.702 40.427 1741.497 664.605 67.650 952.559
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
15 4 0.080 49.932 15 4 0.048 83.469
15 8 0.057 140.395 15 8 0.055 145.257
15 16 0.121 132.365 15 16 0.040 399.458
15 32 0.059 541.201 15 32 0.051 627.186
15 64 0.068 941.879 15 64 0.091 702.711
15 128 0.093 1376.592 15 128 0.123 1040.448
15 256 0.143 1789.570 15 256 0.220 1163.317
15 512 0.233 2198.039 15 512 0.408 1254.371
15 1024 0.433 2365.070 15 1024 0.777 1318.283
15 2048 0.825 2481.923 15 2048 1.545 1325.607
15 4096 1.616 2534.652 15 4096 2.997 1366.736
15 8192 3.202 2558.241 15 8192 5.980 1369.896
15 16384 6.434 2546.486 15 16384 11.934 1372.907
15 32768 17.127 1913.259 15 32768 25.486 1285.726
15 65536 40.597 1614.299 15 65536 57.525 1139.263
15 131072 98.897 1325.339 15 131072 145.304 902.054
15 262144 244.431 1072.466 15 262144 255.128 1027.500
15 524288 461.736 1135.471 15 524288 489.405 1071.276
15 1048576 1404.463 746.603 15 1048576 2050.766 511.309
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1445.462 49.932 2558.241 952.988 83.469 1372.907
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
16 4 0.094 42.582 16 4 0.092 43.464
16 8 0.109 73.423 16 8 0.070 114.520
16 16 0.116 137.801 16 16 0.056 285.570
16 32 0.172 185.897 16 32 0.073 438.620
16 64 0.126 508.400 16 64 0.101 633.102
16 128 0.163 784.899 16 128 0.153 836.248
16 256 0.239 1070.530 16 256 0.253 1012.009
16 512 0.392 1306.255 16 512 0.489 1047.042
16 1024 0.692 1479.493 16 1024 0.888 1153.011
16 2048 1.312 1560.955 16 2048 1.688 1213.268
16 4096 2.503 1636.490 16 4096 3.310 1237.475
16 8192 4.995 1640.083 16 8192 6.515 1257.447
16 16384 10.376 1579.032 16 16384 12.994 1260.885
16 32768 25.901 1265.121 16 32768 30.259 1082.921
16 65536 65.375 1002.461 16 65536 71.519 916.342
16 131072 141.552 925.964 16 131072 146.769 893.049
16 262144 292.107 897.424 16 262144 302.898 865.453
16 524288 610.461 858.839 16 524288 618.427 847.777
16 1048576 1353.084 774.953 16 1048576 1520.120 689.798
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
933.190 42.582 1640.083 833.053 43.464 1260.885
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
17 4 0.079 50.686 17 4 0.040 99.864
17 8 0.054 147.817 17 8 0.166 48.210
17 16 0.063 254.200 17 16 0.044 362.751
17 32 0.071 450.395 17 32 0.460 69.579
17 64 0.073 877.240 17 64 0.075 854.890
17 128 0.098 1306.255 17 128 0.124 1032.444
17 256 0.148 1729.053 17 256 0.220 1163.317
17 512 0.247 2072.861 17 512 0.459 1115.576
17 1024 0.451 2270.067 17 1024 0.807 1268.823
17 2048 0.859 2384.106 17 2048 1.583 1293.665
17 4096 1.698 2412.225 17 4096 3.136 1306.156
17 8192 3.370 2430.827 17 8192 6.253 1310.090
17 16384 6.630 2471.212 17 16384 12.585 1301.875
17 32768 17.129 1913.019 17 32768 27.963 1171.827
17 65536 39.869 1643.781 17 65536 60.414 1084.780
17 131072 90.382 1450.202 17 131072 122.092 1073.551
17 262144 215.820 1214.641 17 262144 260.138 1007.711
17 524288 393.345 1332.896 17 524288 509.908 1028.201
17 1048576 1345.493 779.325 17 1048576 1979.598 529.691
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
1431.095 50.686 2471.212 901.211 48.210 1310.090
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
18 4 0.098 40.820 18 4 0.069 57.852
18 8 0.081 98.690 18 8 0.084 95.325
18 16 0.084 190.650 18 16 0.059 271.696
18 32 0.095 336.385 18 32 0.076 420.745
18 64 0.111 576.042 18 64 0.113 566.320
18 128 0.147 870.131 18 128 0.186 688.296
18 256 0.220 1163.317 18 256 0.340 752.447
18 512 0.378 1354.879 18 512 0.633 808.845
18 1024 0.690 1484.094 18 1024 1.271 805.659
18 2048 1.316 1556.148 18 2048 2.380 860.456
18 4096 2.601 1574.834 18 4096 4.747 862.833
18 8192 5.135 1595.308 18 8192 9.417 869.911
18 16384 10.156 1613.209 18 16384 19.189 853.828
18 32768 27.494 1191.826 18 32768 43.777 748.521
18 65536 64.277 1019.584 18 65536 90.913 720.866
18 131072 145.292 902.128 18 131072 184.628 709.925
18 262144 289.841 904.441 18 262144 375.030 698.995
18 524288 631.107 830.743 18 524288 787.607 665.672
18 1048576 1943.711 539.471 18 1048576 1868.307 561.244
avg (B/µs) min (B/µs) max (B/µs) avg (B/µs) min (B/µs) max (B/µs)
939.090 40.820 1613.209 632.602 57.852 869.911
pack: unpack:
bits size (B) time (µs) speed (B/µs) bits size (B) time (µs) speed (B/µs)
19 4 0.063 63.550 19 4 0.097 41.222
19 8 0.082 97.542 19 8 0.084 95.325
19 16 0.084 190.650 19 16 0.059 271.696
19 32 0.095 336.385 19 32 0.079 405.492
19 64 0.110 581.029 19 64 0.115 555.767
19 128 0.145 881.561 19 128 0.196 652.334
19 256 0.307
gitextract_wmqt4bm2/
├── .gitattributes
├── .gitignore
├── .travis.yml
├── LICENSE
├── PackedArray.c
├── PackedArray.h
├── PackedArraySIMD.c
├── README.md
├── _gnu-make/
│ └── Makefile
├── _ios-xcode/
│ ├── .gitignore
│ ├── PackedArray-Info.plist
│ └── PackedArray.xcodeproj/
│ ├── project.pbxproj
│ └── project.xcworkspace/
│ └── contents.xcworkspacedata
├── _mac-xcode/
│ ├── .gitignore
│ └── PackedArray.xcodeproj/
│ ├── project.pbxproj
│ └── project.xcworkspace/
│ └── contents.xcworkspacedata
├── _win-vs11/
│ ├── .gitignore
│ ├── Common.props
│ ├── Debug.props
│ ├── PackedArray.sln
│ ├── PackedArraySIMDSelfBench.vcxproj
│ ├── PackedArraySIMDSelfTest.vcxproj
│ ├── PackedArraySelfBench.vcxproj
│ ├── PackedArraySelfTest.vcxproj
│ ├── Release.props
│ ├── x64.props
│ └── x86.props
└── benchmark/
├── PackedArraySIMDSelfBench-unrolled-galaxy-note-cortex-a9-1.4GHz.txt
├── PackedArraySIMDSelfBench-unrolled-ipad2-cortex-a9-1GHz.txt
├── PackedArraySIMDSelfBench-unrolled-iphone5-a6-1.3GHz.txt
├── PackedArraySIMDSelfBench-unrolled-mbp-corei7-M620-2.67GHz.txt
├── PackedArraySelfBench-reference-galaxy-note-cortex-a9-1.4GHz.txt
├── PackedArraySelfBench-reference-ipad2-cortex-a9-1GHz.txt
├── PackedArraySelfBench-reference-iphone5-a6-1.3GHz.txt
├── PackedArraySelfBench-reference-mbp-corei7-M620-2.67GHz.txt
├── PackedArraySelfBench-unrolled-galaxy-note-cortex-a9-1.4GHz.txt
├── PackedArraySelfBench-unrolled-ipad2-cortex-a9-1GHz.txt
├── PackedArraySelfBench-unrolled-iphone5-a6-1.3GHz.txt
└── PackedArraySelfBench-unrolled-mbp-corei7-M620-2.67GHz.txt
SYMBOL INDEX (44 symbols across 3 files) FILE: PackedArray.c function PackedArray (line 389) | PackedArray* PackedArray_create(uint32_t bitsPerItem, uint32_t count) function PackedArray_destroy (line 410) | void PackedArray_destroy(PackedArray* a) function PackedArray_pack (line 416) | void PackedArray_pack(PackedArray* a, const uint32_t offset, const uint3... function PackedArray_unpack (line 458) | void PackedArray_unpack(const PackedArray* a, const uint32_t offset, uin... function PackedArray_set (line 500) | void PackedArray_set(PackedArray* a, const uint32_t offset, const uint32... function PackedArray_get (line 538) | uint32_t PackedArray_get(const PackedArray* a, const uint32_t offset) function PackedArray_bufferSize (line 576) | uint32_t PackedArray_bufferSize(const PackedArray* a) function __PackedArray_log2 (line 584) | static uint32_t __PackedArray_log2(uint32_t v) function __PackedArray_highestBitSet (line 607) | static int __PackedArray_highestBitSet(uint32_t v) function PackedArray_computeBitsPerItem (line 619) | uint32_t PackedArray_computeBitsPerItem(const uint32_t* in, uint32_t count) function PackedArray_pack_reference (line 646) | static void PackedArray_pack_reference(PackedArray* a, const uint32_t of... function PackedArray_unpack_reference (line 717) | static void PackedArray_unpack_reference(const PackedArray* a, const uin... function main (line 780) | int main(void) function getChronometerTime (line 941) | static double getChronometerTime(void) function getChronometerTime (line 955) | static double getChronometerTime() function bench_memcpy (line 965) | static double bench_memcpy(uint32_t* in, uint32_t* out, uint32_t count) function bench_loopcpy (line 980) | static double bench_loopcpy(uint32_t* in, uint32_t* out, uint32_t count) function bench_pack (line 1000) | static double bench_pack(uint32_t* in, PackedArray* out, uint32_t count) function bench_unpack (line 1015) | static double bench_unpack(PackedArray* in, uint32_t* out, uint32_t count) function main (line 1032) | int main(void) FILE: PackedArray.h type _PackedArray (line 34) | struct _PackedArray type PackedArray (line 49) | typedef struct _PackedArray PackedArray; FILE: PackedArraySIMD.c function __PackedArray_pack_scalar (line 377) | static void __PackedArray_pack_scalar(uint32_t* buffer, const uint32_t b... function __PackedArray_unpack_scalar (line 418) | static void __PackedArray_unpack_scalar(const uint32_t* buffer, const ui... function PackedArray (line 606) | PackedArray* PackedArray_create(uint32_t bitsPerItem, uint32_t count) function PackedArray_destroy (line 629) | void PackedArray_destroy(PackedArray* a) function PackedArray_pack (line 635) | void PackedArray_pack(PackedArray* a, const uint32_t offset, const uint3... function PackedArray_unpack (line 677) | void PackedArray_unpack(const PackedArray* a, const uint32_t offset, uin... function PackedArray_set (line 719) | void PackedArray_set(PackedArray* a, const uint32_t offset, const uint32... function PackedArray_get (line 757) | uint32_t PackedArray_get(const PackedArray* a, const uint32_t offset) function PackedArray_bufferSize (line 795) | uint32_t PackedArray_bufferSize(const PackedArray* a) function __PackedArray_log2 (line 813) | static uint32_t __PackedArray_log2(uint32_t v) function __PackedArray_highestBitSet (line 836) | static int __PackedArray_highestBitSet(uint32_t v) function PackedArray_computeBitsPerItem (line 848) | uint32_t PackedArray_computeBitsPerItem(const uint32_t* in, uint32_t count) function PackedArray_pack_reference (line 875) | static void PackedArray_pack_reference(PackedArray* a, uint32_t offset, ... function PackedArray_unpack_reference (line 921) | static void PackedArray_unpack_reference(const PackedArray* a, uint32_t ... function main (line 964) | int main(void) function getChronometerTime (line 1125) | static double getChronometerTime(void) function getChronometerTime (line 1139) | static double getChronometerTime() function bench_memcpy (line 1149) | static double bench_memcpy(uint32_t* in, uint32_t* out, uint32_t count) function bench_loopcpy (line 1164) | static double bench_loopcpy(uint32_t* in, uint32_t* out, uint32_t count) function bench_pack (line 1184) | static double bench_pack(uint32_t* in, PackedArray* out, uint32_t count) function bench_unpack (line 1199) | static double bench_unpack(PackedArray* in, uint32_t* out, uint32_t count) function main (line 1216) | int main(void)
Condensed preview — 39 files, each showing path, character count, and a content snippet. Download the .json file or copy for the full structured content (978K chars).
[
{
"path": ".gitattributes",
"chars": 470,
"preview": "* text=auto\n\n# sources\n*.h text diff=cpp\n*.c text diff=cpp\n*.cpp text diff=cpp\n*.rb text diff=ruby\n*.html text diff=html"
},
{
"path": ".gitignore",
"chars": 103,
"preview": ".DS_Store\n*~\n*.swp\n\n/bin\nPackedArray.pp.c\nPackedArray.cut.c\nPackedArraySIMD.pp.c\nPackedArraySIMD.cut.c\n"
},
{
"path": ".travis.yml",
"chars": 176,
"preview": "language: cpp\ncompiler:\n - clang\n - gcc\nenv:\n - TARGET=build\n - TARGET=test\n - TARGET=preprocess\n - TARGET=cut\n -"
},
{
"path": "LICENSE",
"chars": 561,
"preview": " DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE \n Version 2, December 2004 \n\n Copyright (C) 2004 "
},
{
"path": "PackedArray.c",
"chars": 37878,
"preview": "// see README.md for usage instructions.\n// (‑●‑●)> released under the WTFPL v2 license, by Gregory Pakosz (@gpakosz)\n\n#"
},
{
"path": "PackedArray.h",
"chars": 2082,
"preview": "#ifndef PACKEDARRAY_H\n#define PACKEDARRAY_H\n\n#ifdef __cplusplus\nextern \"C\" {\n#endif\n\n#include <stdint.h>\n\n/*\n\nPackedArra"
},
{
"path": "PackedArraySIMD.c",
"chars": 45622,
"preview": "// see README.md for usage instructions.\n// (‑●‑●)> released under the WTFPL v2 license, by Gregory Pakosz (@gpakosz)\n\n#"
},
{
"path": "README.md",
"chars": 9358,
"preview": "# PackedArray: random access array of tightly packed unsigned integers\n[![Build Status](https://travis-ci.org/gpakosz/Pa"
},
{
"path": "_gnu-make/Makefile",
"chars": 2849,
"preview": ".PHONY: build test preprocess cut assembly clean\n\n# directories\nifeq ($(realpath .),)\n $(error your version of Make doe"
},
{
"path": "_ios-xcode/.gitignore",
"chars": 26,
"preview": "xcuserdata/\nxcshareddata/\n"
},
{
"path": "_ios-xcode/PackedArray-Info.plist",
"chars": 1457,
"preview": "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<!DOCTYPE plist PUBLIC \"-//Apple//DTD PLIST 1.0//EN\" \"http://www.apple.com/DTDs/P"
},
{
"path": "_ios-xcode/PackedArray.xcodeproj/project.pbxproj",
"chars": 15240,
"preview": "// !$*UTF8*$!\n{\n\tarchiveVersion = 1;\n\tclasses = {\n\t};\n\tobjectVersion = 46;\n\tobjects = {\n\n/* Begin PBXBuildFile section *"
},
{
"path": "_ios-xcode/PackedArray.xcodeproj/project.xcworkspace/contents.xcworkspacedata",
"chars": 156,
"preview": "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<Workspace\n version = \"1.0\">\n <FileRef\n location = \"self:PackedArray.xco"
},
{
"path": "_mac-xcode/.gitignore",
"chars": 26,
"preview": "xcuserdata/\nxcshareddata/\n"
},
{
"path": "_mac-xcode/PackedArray.xcodeproj/project.pbxproj",
"chars": 15082,
"preview": "// !$*UTF8*$!\n{\n\tarchiveVersion = 1;\n\tclasses = {\n\t};\n\tobjectVersion = 46;\n\tobjects = {\n\n/* Begin PBXBuildFile section *"
},
{
"path": "_mac-xcode/PackedArray.xcodeproj/project.xcworkspace/contents.xcworkspacedata",
"chars": 156,
"preview": "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<Workspace\n version = \"1.0\">\n <FileRef\n location = \"self:PackedArray.xco"
},
{
"path": "_win-vs11/.gitignore",
"chars": 59,
"preview": "/build\n/ipch\n\n*.suo\n*.sdf\n*.opensdf\n*.user\n*.sln.docstates\n"
},
{
"path": "_win-vs11/Common.props",
"chars": 1855,
"preview": "<?xml version=\"1.0\" encoding=\"utf-8\"?>\r\n<Project DefaultTargets=\"Build\" ToolsVersion=\"4.0\" xmlns=\"http://schemas.micros"
},
{
"path": "_win-vs11/Debug.props",
"chars": 825,
"preview": "<?xml version=\"1.0\" encoding=\"utf-8\"?>\r\n<Project DefaultTargets=\"Build\" ToolsVersion=\"4.0\" xmlns=\"http://schemas.micros"
},
{
"path": "_win-vs11/PackedArray.sln",
"chars": 3632,
"preview": "\r\nMicrosoft Visual Studio Solution File, Format Version 12.00\r\n# Visual Studio Express 2012 for Windows Desktop\r\nProjec"
},
{
"path": "_win-vs11/PackedArraySIMDSelfBench.vcxproj",
"chars": 5775,
"preview": "<?xml version=\"1.0\" encoding=\"utf-8\"?>\r\n<Project DefaultTargets=\"Build\" ToolsVersion=\"4.0\" xmlns=\"http://schemas.micros"
},
{
"path": "_win-vs11/PackedArraySIMDSelfTest.vcxproj",
"chars": 5745,
"preview": "<?xml version=\"1.0\" encoding=\"utf-8\"?>\r\n<Project DefaultTargets=\"Build\" ToolsVersion=\"4.0\" xmlns=\"http://schemas.micros"
},
{
"path": "_win-vs11/PackedArraySelfBench.vcxproj",
"chars": 5767,
"preview": "<?xml version=\"1.0\" encoding=\"utf-8\"?>\r\n<Project DefaultTargets=\"Build\" ToolsVersion=\"4.0\" xmlns=\"http://schemas.micros"
},
{
"path": "_win-vs11/PackedArraySelfTest.vcxproj",
"chars": 5737,
"preview": "<?xml version=\"1.0\" encoding=\"utf-8\"?>\r\n<Project DefaultTargets=\"Build\" ToolsVersion=\"4.0\" xmlns=\"http://schemas.micros"
},
{
"path": "_win-vs11/Release.props",
"chars": 1630,
"preview": "<?xml version=\"1.0\" encoding=\"utf-8\"?>\r\n<Project DefaultTargets=\"Build\" ToolsVersion=\"4.0\" xmlns=\"http://schemas.micros"
},
{
"path": "_win-vs11/x64.props",
"chars": 484,
"preview": "<?xml version=\"1.0\" encoding=\"utf-8\"?>\r\n<Project DefaultTargets=\"Build\" ToolsVersion=\"4.0\" xmlns=\"http://schemas.micros"
},
{
"path": "_win-vs11/x86.props",
"chars": 610,
"preview": "<?xml version=\"1.0\" encoding=\"utf-8\"?>\r\n<Project DefaultTargets=\"Build\" ToolsVersion=\"4.0\" xmlns=\"http://schemas.micros"
},
{
"path": "benchmark/PackedArraySIMDSelfBench-unrolled-galaxy-note-cortex-a9-1.4GHz.txt",
"chars": 59988,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
},
{
"path": "benchmark/PackedArraySIMDSelfBench-unrolled-ipad2-cortex-a9-1GHz.txt",
"chars": 60085,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
},
{
"path": "benchmark/PackedArraySIMDSelfBench-unrolled-iphone5-a6-1.3GHz.txt",
"chars": 59989,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
},
{
"path": "benchmark/PackedArraySIMDSelfBench-unrolled-mbp-corei7-M620-2.67GHz.txt",
"chars": 59987,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
},
{
"path": "benchmark/PackedArraySelfBench-reference-galaxy-note-cortex-a9-1.4GHz.txt",
"chars": 59988,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
},
{
"path": "benchmark/PackedArraySelfBench-reference-ipad2-cortex-a9-1GHz.txt",
"chars": 59989,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
},
{
"path": "benchmark/PackedArraySelfBench-reference-iphone5-a6-1.3GHz.txt",
"chars": 59989,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
},
{
"path": "benchmark/PackedArraySelfBench-reference-mbp-corei7-M620-2.67GHz.txt",
"chars": 59989,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
},
{
"path": "benchmark/PackedArraySelfBench-unrolled-galaxy-note-cortex-a9-1.4GHz.txt",
"chars": 59988,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
},
{
"path": "benchmark/PackedArraySelfBench-unrolled-ipad2-cortex-a9-1GHz.txt",
"chars": 59989,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
},
{
"path": "benchmark/PackedArraySelfBench-unrolled-iphone5-a6-1.3GHz.txt",
"chars": 59987,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
},
{
"path": "benchmark/PackedArraySelfBench-unrolled-mbp-corei7-M620-2.67GHz.txt",
"chars": 59989,
"preview": "-- PackedArray self bench ------------------------------------------------------\nmemcpy:\nbits\tsize (B)\ttime (µs)\tspeed ("
}
]
About this extraction
This page contains the full source code of the gpakosz/PackedArray GitHub repository, extracted and formatted as plain text for AI agents and large language models (LLMs). The extraction includes 39 files (862.6 KB), approximately 390.9k tokens, and a symbol index with 44 extracted functions, classes, methods, constants, and types. Use this with OpenClaw, Claude, ChatGPT, Cursor, Windsurf, or any other AI tool that accepts text input. You can copy the full output to your clipboard or download it as a .txt file.
Extracted by GitExtract — free GitHub repo to text converter for AI. Built by Nikandr Surkov.