Implementation notes: amd64, hydra7, crypto_sort/int32

Computer: hydra7
Microarchitecture: amd64; Sandy Bridge+AES (206a7)
Architecture: amd64
CPU ID: GenuineIntel-000206a7-bfebfbff
SUPERCOP version: 20240425
Operation: crypto_sort
Primitive: int32

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
23938	1525 0 0	14093 804 960	`T:radix256ml`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
24612	1303 0 0	11224 780 928	`T:radix256ml`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
24703	1525 0 0	12845 804 960	`T:radix256ml`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
26429	1946 0 0	11872 780 928	`T:radix256sml`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
27067	2205 0 0	13533 804 960	`T:radix256sml`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
27261	2264 0 0	13260 796 960	`T:radix256sml`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
27680	2205 0 0	14781 804 960	`T:radix256sml`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
27842	1535 0 0	12524 796 960	`T:radix256ml`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
31231	1328 0 0	12719 812 992	`T:stdsort`	`g++_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
31726	1141 0 0	12238 804 992	`T:stdsort`	`g++_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
32565	632 0 0	10536 780 928	`T:herf`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
33147	710 0 0	12005 804 960	`T:herf`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
33995	709 0 0	13253 804 960	`T:herf`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
35008	1364 0 0	14023 812 992	`T:stdsort`	`g++_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
38671	704 0 0	11684 796 960	`T:herf`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
39256	930 0 0	10946 788 960	`T:stdsort`	`g++_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
74647	8840 0 0	20141 804 960	`x863`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
76190	9258 0 0	21813 804 960	`x863`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
77998	9360 0 0	20348 796 960	`x863`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
88328	1315 0 0	13861 804 960	`x86`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
89980	1315 0 0	12613 804 960	`x86`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
91337	1115 0 0	11024 780 928	`x86`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
91647	1416 0 0	12396 796 960	`x86`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
95740	3935 0 0	13872 780 928	`x863`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
108039	1664 0 0	14205 804 960	`portable4`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
116806	731 0 0	10632 780 928	`portable4`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
118030	887 0 0	12189 804 960	`portable4`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
125839	947 0 0	11916 796 960	`portable4`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
126220	1333 0 0	11248 780 928	`portable5`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
128304	469 0 0	11765 804 960	`compact`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
128462	445 0 0	11749 804 960	`portable3`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
128779	445 0 0	12997 804 960	`portable3`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
128832	445 0 0	12997 804 960	`compact`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
129615	2328 0 0	14877 804 960	`portable5`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
130171	1555 0 0	12861 804 960	`portable5`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
135695	474 0 0	11444 796 960	`portable3`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
135705	385 0 0	10296 780 928	`portable3`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
137061	1651 0 0	12620 796 960	`portable5`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
185169	460 0 0	10376 780 928	`compact`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425
205283	612 0 0	11588 796 960	`compact`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240426	20240425

Test failure

Implementation: T:krasnov
Security model: timingleaks
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

error 111

Number of similar (compiler,implementation) pairs: 4, namely:

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:krasnov
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:krasnov
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:krasnov
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:krasnov

Compiler output

Implementation: T:aspas
Security model: timingleaks
Compiler: g++ -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

api.cpp: In file included from sorter.h:123,
api.cpp: from aspas.tcc:31,
api.cpp: from aspas.h:204,
api.cpp: from api.cpp:3:
api.cpp: sorter_avx.tcc:95:47: warning: ignoring attributes on template argument '__m256' [-Wignored-attributes]
api.cpp: 95 | typename std::enable_if<std::is_same<T, __m256>::value>::type
api.cpp: | ^
api.cpp: sorter_avx.tcc:149:47: warning: ignoring attributes on template argument '__m256' [-Wignored-attributes]
api.cpp: 149 | typename std::enable_if<std::is_same<T, __m256>::value>::type
api.cpp: | ^
api.cpp: sorter_avx.tcc:317:47: warning: ignoring attributes on template argument '__m256' [-Wignored-attributes]
api.cpp: 317 | typename std::enable_if<std::is_same<T, __m256>::value>::type
api.cpp: | ^
api.cpp: sorter_avx.tcc:647:48: warning: ignoring attributes on template argument '__m256i' [-Wignored-attributes]
api.cpp: 647 | typename std::enable_if<std::is_same<T, __m256i>::value>::type
api.cpp: | ^
api.cpp: sorter_avx.tcc:701:48: warning: ignoring attributes on template argument '__m256i' [-Wignored-attributes]
api.cpp: 701 | typename std::enable_if<std::is_same<T, __m256i>::value>::type
api.cpp: | ^
api.cpp: sorter_avx.tcc:854:48: warning: ignoring attributes on template argument '__m256i' [-Wignored-attributes]
api.cpp: 854 | typename std::enable_if<std::is_same<T, __m256i>::value>::type
api.cpp: | ^
api.cpp: sorter_avx.tcc:1184:48: warning: ignoring attributes on template argument '__m256d' [-Wignored-attributes]
api.cpp: 1184 | typename std::enable_if<std::is_same<T, __m256d>::value>::type
api.cpp: | ^
api.cpp: ...

Number of similar (compiler,implementation) pairs: 4, namely:

Compiler	Implementations
g++ -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:aspas
g++ -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:aspas
g++ -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:aspas
g++ -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:aspas

Compiler output

Implementation: avx2
Security model: constbranchindex
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

sort.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
sort.c: from sort.c:4:
sort.c: sort.c: In function 'minmax_vector':
sort.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h:363:1: error: inlining failed in call to 'always_inline' '_mm256_max_epi32': target specific option mismatch
sort.c: 363 | _mm256_max_epi32 (__m256i __A, __m256i __B)
sort.c: | ^~~~~~~~~~~~~~~~
sort.c: sort.c:12:21: note: called from here
sort.c: 12 | #define int32x8_max _mm256_max_epi32
sort.c: | ^
sort.c: sort.c:17:7: note: in expansion of macro 'int32x8_max'
sort.c: 17 | b = int32x8_max(a,b); \
sort.c: | ^~~~~~~~~~~
sort.c: sort.c:36:5: note: in expansion of macro 'int32x8_MINMAX'
sort.c: 36 | int32x8_MINMAX(x0,y0);
sort.c: | ^~~~~~~~~~~~~~
sort.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
sort.c: from sort.c:4:
sort.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h:405:1: error: inlining failed in call to 'always_inline' '_mm256_min_epi32': target specific option mismatch
sort.c: 405 | _mm256_min_epi32 (__m256i __A, __m256i __B)
sort.c: | ^~~~~~~~~~~~~~~~
sort.c: sort.c:11:21: note: called from here
sort.c: 11 | #define int32x8_min _mm256_min_epi32
sort.c: | ^
sort.c: sort.c:16:15: note: in expansion of macro 'int32x8_min'
sort.c: 16 | int32x8 c = int32x8_min(a,b); \
sort.c: ...

Number of similar (compiler,implementation) pairs: 4, namely:

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	avx2
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	avx2
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	avx2
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	avx2

Compiler output

Implementation: oldavx2
Security model: constbranchindex
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

int32_sort.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
int32_sort.c: from int32_sort.c:3:
int32_sort.c: int32_sort.c: In function 'minmax02through1315':
int32_sort.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h:878:1: error: inlining failed in call to 'always_inline' '_mm256_unpackhi_epi64': target specific option mismatch
int32_sort.c: 878 | _mm256_unpackhi_epi64 (__m256i __A, __m256i __B)
int32_sort.c: | ^~~~~~~~~~~~~~~~~~~~~
int32_sort.c: int32_sort.c:23:7: note: called from here
int32_sort.c: 23 | b = _mm256_unpackhi_epi64(g,h);
int32_sort.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~
int32_sort.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
int32_sort.c: from int32_sort.c:3:
int32_sort.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h:906:1: error: inlining failed in call to 'always_inline' '_mm256_unpacklo_epi64': target specific option mismatch
int32_sort.c: 906 | _mm256_unpacklo_epi64 (__m256i __A, __m256i __B)
int32_sort.c: | ^~~~~~~~~~~~~~~~~~~~~
int32_sort.c: int32_sort.c:22:7: note: called from here
int32_sort.c: 22 | a = _mm256_unpacklo_epi64(g,h);
int32_sort.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~
int32_sort.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
int32_sort.c: from int32_sort.c:3:
int32_sort.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h:363:1: error: inlining failed in call to 'always_inline' '_mm256_max_epi32': target specific option mismatch
int32_sort.c: 363 | _mm256_max_epi32 (__m256i __A, __m256i __B)
int32_sort.c: | ^~~~~~~~~~~~~~~~
int32_sort.c: int32_sort.c:21:15: note: called from here
int32_sort.c: 21 | __m256i h = _mm256_max_epi32(c,d);
int32_sort.c: | ^~~~~~~~~~~~~~~~~~~~~
int32_sort.c: ...

Number of similar (compiler,implementation) pairs: 4, namely:

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	oldavx2
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	oldavx2
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	oldavx2
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	oldavx2

Compiler output

Implementation: T:sid1607
Security model: timingleaks
Compiler: g++ -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

merge_sort.cpp: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
merge_sort.cpp: from merge_sort.h:1,
merge_sort.cpp: from merge_sort.cpp:1:
merge_sort.cpp: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h: In function '__m256i reverse(__m256i&)':
merge_sort.cpp: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h:1044:1: error: inlining failed in call to 'always_inline' '__m256i _mm256_permutevar8x32_epi32(__m256i, __m256i)': target specific option mismatch
merge_sort.cpp: 1044 | _mm256_permutevar8x32_epi32 (__m256i __X, __m256i __Y)
merge_sort.cpp: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~
merge_sort.cpp: merge_sort.cpp:11:37: note: called from here
merge_sort.cpp: 11 | return _mm256_permutevar8x32_epi32(v, global_masks.rev_idx_mask);
merge_sort.cpp: | ~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Number of similar (compiler,implementation) pairs: 4, namely:

Compiler	Implementations
g++ -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:sid1607
g++ -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:sid1607
g++ -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:sid1607
g++ -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:sid1607