Implementation notes: amd64, h8atom, crypto_core/invhrss701

Computer: h8atom
Microarchitecture: amd64; Bonnell (30661)
Architecture: amd64
CPU ID: GenuineIntel-00030661-bfebfbff
SUPERCOP version: 20240107
Operation: crypto_core
Primitive: invhrss701

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
6180552	5322 0 0	19080 784 832	`simpler`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
19047896	4079 0 0	17710 776 832	`ref`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
20531161	3504 0 0	17014 824 760	`simpler`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
21923664	2561 0 0	16070 824 760	`simpler`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
22156547	2189 0 0	14526 824 760	`simpler`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
37904370	6460 0 0	19900 816 760	`ref`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
38076003	4068 0 0	16324 816 760	`ref`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
38080063	4804 0 0	18236 816 760	`ref`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
55832525	1017 0 0	13448 784 832	`simpler`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
59746288	814 0 0	12110 824 760	`simpler`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
60726792	781 0 0	11336 816 760	`simpler`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
70625128	783 0 0	11651 760 800	`simpler`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
73878532	895 0 0	12893 768 832	`simpler`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
104781873	1459 0 0	13774 776 832	`ref`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
106743434	1381 0 0	12596 816 760	`ref`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
111623316	1888 0 0	12382 808 760	`ref`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
116898691	1179 0 0	11961 752 800	`ref`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
134313851	1374 0 0	13349 768 832	`ref`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212

Test failure

Implementation: avx2
Security model: constbranchindex
Compiler: clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

error 111

Number of similar (compiler,implementation) pairs: 18, namely:

Compiler	Implementations
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	avx2
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	avx2
clang -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	avx2
clang -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	avx2
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	avx2
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
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	faster921
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	faster921
clang -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	faster921
clang -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	faster921
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	faster921
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	faster921
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	faster921
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	faster921
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	faster921

Compiler output

Implementation: faster
Security model: constbranchindex
Compiler: clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

core.c: core.c:353:19: error: always_inline function '_mm256_set_epi32' requires target feature 'avx', but would be inlined into function 'r3_recip' that is compiled without support for 'avx'
core.c: F0[0] = F0[1] = _mm256_set_epi32(-1,-1,-1,-1,-1,-1,-1,-1);
core.c: ^
core.c: core.c:353:19: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
core.c: core.c:354:11: error: always_inline function '_mm256_set_epi32' requires target feature 'avx', but would be inlined into function 'r3_recip' that is compiled without support for 'avx'
core.c: F0[2] = _mm256_set_epi32(32767,-1,32767,-1,32767,-1,65535,-1);
core.c: ^
core.c: core.c:354:11: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
core.c: core.c:355:27: error: always_inline function '_mm256_set1_epi32' requires target feature 'avx', but would be inlined into function 'r3_recip' that is compiled without support for 'avx'
core.c: F1[0] = F1[1] = F1[2] = _mm256_set1_epi32(0);
core.c: ^
core.c: core.c:355:27: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
core.c: core.c:357:27: error: always_inline function '_mm256_set1_epi32' requires target feature 'avx', but would be inlined into function 'r3_recip' that is compiled without support for 'avx'
core.c: V0[2] = V0[1] = V0[0] = _mm256_set1_epi32(0);
core.c: ^
core.c: core.c:357:27: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
core.c: core.c:358:27: error: always_inline function '_mm256_set1_epi32' requires target feature 'avx', but would be inlined into function 'r3_recip' that is compiled without support for 'avx'
core.c: V1[2] = V1[1] = V1[0] = _mm256_set1_epi32(0);
core.c: ^
core.c: core.c:358:27: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
core.c: core.c:360:11: error: always_inline function '_mm256_set_epi32' requires target feature 'avx', but would be inlined into function 'r3_recip' that is compiled without support for 'avx'
core.c: R0[0] = _mm256_set_epi32(0,0,0,0,0,0,0,1);
core.c: ^
core.c: core.c:360:11: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
core.c: core.c:361:19: error: always_inline function '_mm256_set1_epi32' requires target feature 'avx', but would be inlined into function 'r3_recip' that is compiled without support for 'avx'
core.c: ...

Number of similar (compiler,implementation) pairs: 5, namely:

Compiler	Implementations
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	faster
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	faster
clang -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	faster
clang -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	faster
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	faster

Compiler output

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

core.c: core.c: In function 'r3_recip':
core.c: core.c:353:17: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
core.c: 353 | F0[0] = F0[1] = _mm256_set_epi32(-1,-1,-1,-1,-1,-1,-1,-1);
core.c: | ~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
core.c: core.c: In function 'vec256_swap':
core.c: core.c:171:20: note: the ABI for passing parameters with 32-byte alignment has changed in GCC 4.6
core.c: 171 | static inline void vec256_swap(vec256 *f,vec256 *g,int len,vec256 mask)
core.c: | ^~~~~~~~~~~
core.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
core.c: from core.c:4:
core.c: core.c: In function 'vec256_frombits':
core.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h:597:1: error: inlining failed in call to 'always_inline' '_mm256_shuffle_epi32': target specific option mismatch
core.c: 597 | _mm256_shuffle_epi32 (__m256i __A, const int __mask)
core.c: | ^~~~~~~~~~~~~~~~~~~~
core.c: core.c:56:9: note: called from here
core.c: 56 | h = _mm256_shuffle_epi32(h,0xd8);
core.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
core.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
core.c: from core.c:4:
core.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h:1071:1: error: inlining failed in call to 'always_inline' '_mm256_permute4x64_epi64': target specific option mismatch
core.c: 1071 | _mm256_permute4x64_epi64 (__m256i __X, const int __M)
core.c: | ^~~~~~~~~~~~~~~~~~~~~~~~
core.c: core.c:55:9: note: called from here
core.c: 55 | h = _mm256_permute4x64_epi64(h,0xd8);
core.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
core.c: ...

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

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