Implementation notes: amd64, h3neo, crypto_decode/653x4621

Computer: h3neo
Microarchitecture: amd64; K10 45nm (100f63)
Architecture: amd64
CPU ID: AuthenticAMD-00100f63-078bfbff
SUPERCOP version: 20240107
Operation: crypto_decode
Primitive: 653x4621

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
14099	3527 0 0	15435 844 896	`int16`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
14135	1874 0 0	12867 844 896	`int16`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
14209	1791 0 0	11221 836 896	`int16`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
14329	3561 0 0	15955 844 896	`int16`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
14621	2003 0 0	12051 844 896	`int16`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
16826	2139 0 0	11920 780 928	`int16`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
16976	2262 0 0	13044 796 960	`int16`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
17433	3066 0 0	15941 804 960	`int16`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
17492	2241 0 0	13789 804 960	`int16`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
21359	2646 0 0	12715 844 896	`portable`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
21400	2516 0 0	11949 836 896	`portable`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
24302	3516 0 0	16397 804 960	`portable`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
24591	2646 0 0	14205 804 960	`portable`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
25612	8479 0 0	20875 844 896	`portable`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
26518	8329 0 0	20259 844 896	`portable`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
26601	6037 0 0	17051 844 896	`portable`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
89438	3646 0 0	16027 844 896	`ref`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
92068	1718 0 0	14557 804 960	`ref`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
92504	3649 0 0	15571 844 896	`ref`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
92772	2070 0 0	13083 844 896	`ref`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
93780	1681 0 0	11488 780 928	`portable`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
94714	1616 0 0	13125 804 960	`ref`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
95755	1183 0 0	11251 844 896	`ref`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
96560	1785 0 0	12556 796 960	`portable`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
101556	1614 0 0	11021 836 896	`ref`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
106000	1655 0 0	12356 796 960	`ref`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
106716	1524 0 0	11200 780 928	`ref`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212

Compiler output

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

decode.c: decode.c:214:15: error: always_inline function '_mm256_loadu_si256' requires target feature 'avx', but would be inlined into function 'crypto_decode_653x4621_avx_constbranchindex' that is compiled without support for 'avx'
decode.c: A2 = A0 = _mm256_loadu_si256((__m256i *) &R5[i]);
decode.c: ^
decode.c: decode.c:214:15: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
decode.c: decode.c:215:10: error: always_inline function '_mm256_cvtepu8_epi16' requires target feature 'avx2', but would be inlined into function 'crypto_decode_653x4621_avx_constbranchindex' that is compiled without support for 'avx2'
decode.c: S0 = _mm256_cvtepu8_epi16(_mm_loadu_si128((__m128i *) (s+i)));
decode.c: ^
decode.c: decode.c:215:10: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
decode.c: decode.c:216:14: warning: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI [-Wpsabi]
decode.c: A0 = sub(mulhiconst(A0,-468),mulhiconst(mulloconst(A0,-9394),1786)); /* -1010...893 */
decode.c: ^
decode.c: decode.c:216:45: warning: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI [-Wpsabi]
decode.c: A0 = sub(mulhiconst(A0,-468),mulhiconst(mulloconst(A0,-9394),1786)); /* -1010...893 */
decode.c: ^
decode.c: decode.c:216:34: warning: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI [-Wpsabi]
decode.c: A0 = sub(mulhiconst(A0,-468),mulhiconst(mulloconst(A0,-9394),1786)); /* -1010...893 */
decode.c: ^
decode.c: decode.c:216:10: warning: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI [-Wpsabi]
decode.c: A0 = sub(mulhiconst(A0,-468),mulhiconst(mulloconst(A0,-9394),1786)); /* -1010...893 */
decode.c: ^
decode.c: decode.c:217:10: warning: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI [-Wpsabi]
decode.c: A0 = add(A0,S0); /* -1010...1148 */
decode.c: ^
decode.c: decode.c:218:10: warning: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI [-Wpsabi]
decode.c: A0 = ifnegaddconst(A0,1786); /* 0...1785 */
decode.c: ...

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

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

Compiler output

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

decode.c: decode.c: In function 'add':
decode.c: decode.c:21:1: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
decode.c: 21 | {
decode.c: | ^
decode.c: decode.c: In function 'signedshiftrightconst':
decode.c: decode.c:35:23: note: the ABI for passing parameters with 32-byte alignment has changed in GCC 4.6
decode.c: 35 | static inline __m256i signedshiftrightconst(__m256i x,int16 y)
decode.c: | ^~~~~~~~~~~~~~~~~~~~~
decode.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
decode.c: from decode.c:3:
decode.c: decode.c: In function 'add':
decode.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h:112:1: error: inlining failed in call to 'always_inline' '_mm256_add_epi16': target specific option mismatch
decode.c: 112 | _mm256_add_epi16 (__m256i __A, __m256i __B)
decode.c: | ^~~~~~~~~~~~~~~~
decode.c: decode.c:22:10: note: called from here
decode.c: 22 | return _mm256_add_epi16(x,y);
decode.c: | ^~~~~~~~~~~~~~~~~~~~~

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

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