Implementation notes: amd64, gemini, crypto_hash/sha512

Computer: gemini
Microarchitecture: amd64; Goldmont Plus (706a8)
Architecture: amd64
CPU ID: GenuineIntel-000706a8-bfebfbff
SUPERCOP version: 20240107
Operation: crypto_hash
Primitive: sha512

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
18159	78 0 0	11296 804 920	`openssl`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
18159	78 0 0	12512 804 920	`openssl`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
18231	70 0 0	10705 812 888	`openssl`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
18231	81 0 0	10759 796 920	`openssl`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
18259	70 0 0	12043 820 888	`openssl`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
18282	70 0 0	12963 820 888	`openssl`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
18283	70 0 0	14347 820 888	`openssl`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
18283	70 0 0	10209 812 888	`openssl`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
18287	70 0 0	9747 780 920	`openssl`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
20604	793 0 0	18184 812 888	`ref`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
20617	793 0 0	15880 812 888	`ref`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
20618	967 0 0	16488 780 952	`ref`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
20621	799 0 0	16800 812 888	`ref`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
20623	678 0 0	13934 804 888	`ref`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
20711	622 0 0	13395 756 920	`ref`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
20856	960 0 0	15288 780 952	`ref`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
21137	? ? ?	295388 11193 1504	`rust_crypto`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
21172	? ? ?	294891 11185 1440	`rust_crypto`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
21172	? ? ?	296588 11193 1504	`rust_crypto`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
21189	? ? ?	297041 11193 1440	`rust_crypto`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
21192	? ? ?	294443 11185 1440	`rust_crypto`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
21196	? ? ?	298425 11193 1440	`rust_crypto`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
21418	? ? ?	291928 10880 1440	`rust_sha2`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
21423	? ? ?	291513 10880 1504	`rust_sha2`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
21440	? ? ?	295910 10888 1440	`rust_sha2`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
21444	? ? ?	293606 10888 1440	`rust_sha2`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
21457	? ? ?	292376 10880 1440	`rust_sha2`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
21460	? ? ?	292873 10888 1504	`rust_sha2`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
21801	8134 0 0	20504 780 952	`sphlib-small`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
21834	8134 0 0	20504 780 952	`sphlib`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
22429	703 0 0	14446 804 888	`ref`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
22615	4599 0 0	14235 756 920	`sphlib-small`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
22616	4599 0 0	14235 756 920	`sphlib`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
22931	4760 768 0	1366384 144380 10328	`cryptopp`	`clang++_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
22969	4926 992 0	1370291 144388 10328	`cryptopp`	`clang++_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
23090	4633 928 0	1658715 144612 10424	`cryptopp`	`g++_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
23110	4838 928 0	1659485 144612 10424	`cryptopp`	`g++_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
23146	4728 704 0	1659775 144380 10424	`cryptopp`	`g++_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
23282	629 0 0	14415 772 952	`ref`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
23286	6032 0 0	17216 780 952	`sphlib`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
23310	6032 0 0	17216 780 952	`sphlib-small`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
23362	5257 0 0	15911 772 952	`sphlib-small`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
23365	5257 0 0	15911 772 952	`sphlib`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
23730	6208 0 0	19040 812 888	`sphlib`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
23736	6208 0 0	19040 812 888	`sphlib-small`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
23779	8604 1576 0	1373845 144972 10328	`cryptopp`	`clang++_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
24321	6313 0 0	20536 812 888	`sphlib-small`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
24322	6313 0 0	18248 812 888	`sphlib-small`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
24396	6313 0 0	20536 812 888	`sphlib`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
24402	6313 0 0	18232 812 888	`sphlib`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
24786	5289 0 0	15390 804 888	`sphlib-small`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
24800	5289 0 0	15390 804 888	`sphlib`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
24971	2607 0 0	16776 812 888	`compact`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240113	20240107
25096	3781 992 0	1656897 144612 10392	`cryptopp`	`g++_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
25325	4976 992 0	1368051 144388 10328	`cryptopp`	`clang++_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
26346	2342 0 0	15128 812 888	`compact`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240113	20240107
27107	2332 0 0	14200 812 888	`compact`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240113	20240107
29186	5823 0 0	18128 780 952	`compact`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240113	20240107
29201	5493 0 0	16134 804 888	`sphlib-small`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
29202	5493 0 0	16134 804 888	`sphlib`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
31967	2081 0 0	13216 780 952	`compact`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240113	20240107
32544	1981 0 0	12030 804 888	`compact`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240113	20240107
34450	1760 0 0	12335 772 952	`compact`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240113	20240107
35894	1821 0 0	12374 804 888	`compact`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240113	20240107
35939	1642 0 0	11219 756 920	`compact`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240113	20240107

Compiler output

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

hash.c: hash.c:33:8: error: always_inline function '_mm256_loadu_si256' requires target feature 'avx', but would be inlined into function 'crypto_hash_sha512_blocksplusavx_constbranchindex' that is compiled without support for 'avx'
hash.c: X0 = load256(iv);
hash.c: ^
hash.c: hash.c:22:21: note: expanded from macro 'load256'
hash.c: #define load256(x) (_mm256_loadu_si256((void *) (x)))
hash.c: ^
hash.c: hash.c:33:8: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
hash.c: hash.c:22:21: note: expanded from macro 'load256'
hash.c: #define load256(x) (_mm256_loadu_si256((void *) (x)))
hash.c: ^
hash.c: hash.c:34:8: error: always_inline function '_mm256_loadu_si256' requires target feature 'avx', but would be inlined into function 'crypto_hash_sha512_blocksplusavx_constbranchindex' that is compiled without support for 'avx'
hash.c: X1 = load256(iv + 32);
hash.c: ^
hash.c: hash.c:22:21: note: expanded from macro 'load256'
hash.c: #define load256(x) (_mm256_loadu_si256((void *) (x)))
hash.c: ^
hash.c: hash.c:34:8: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
hash.c: hash.c:22:21: note: expanded from macro 'load256'
hash.c: #define load256(x) (_mm256_loadu_si256((void *) (x)))
hash.c: ^
hash.c: hash.c:36:3: error: always_inline function '_mm256_storeu_si256' requires target feature 'avx', but would be inlined into function 'crypto_hash_sha512_blocksplusavx_constbranchindex' that is compiled without support for 'avx'
hash.c: store256(h,X0);
hash.c: ^
hash.c: hash.c:23:24: note: expanded from macro 'store256'
hash.c: #define store256(x,y) (_mm256_storeu_si256((void *) (x),y))
hash.c: ...

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

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

Compiler output

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

hash.c: hash.c: In function 'crypto_hash_sha512_blocksplusavx_constbranchindex':
hash.c: hash.c:33:6: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
hash.c: 33 | X0 = load256(iv);
hash.c: | ^
hash.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:51,
hash.c: from hash.c:1:
hash.c: /usr/lib/gcc/x86_64-linux-gnu/10/include/avxintrin.h:926:1: error: inlining failed in call to 'always_inline' '_mm256_storeu_si256': target specific option mismatch
hash.c: 926 | _mm256_storeu_si256 (__m256i_u *__P, __m256i __A)
hash.c: | ^~~~~~~~~~~~~~~~~~~
hash.c: hash.c:23:24: note: called from here
hash.c: 23 | #define store256(x,y) (_mm256_storeu_si256((void *) (x),y))
hash.c: | ~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
hash.c: hash.c:37:3: note: in expansion of macro 'store256'
hash.c: 37 | store256(h + 32,X1);
hash.c: | ^~~~~~~~
hash.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:51,
hash.c: from hash.c:1:
hash.c: /usr/lib/gcc/x86_64-linux-gnu/10/include/avxintrin.h:926:1: error: inlining failed in call to 'always_inline' '_mm256_storeu_si256': target specific option mismatch
hash.c: 926 | _mm256_storeu_si256 (__m256i_u *__P, __m256i __A)
hash.c: | ^~~~~~~~~~~~~~~~~~~
hash.c: hash.c:23:24: note: called from here
hash.c: 23 | #define store256(x,y) (_mm256_storeu_si256((void *) (x),y))
hash.c: | ~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
hash.c: hash.c:36:3: note: in expansion of macro 'store256'
hash.c: 36 | store256(h,X0);
hash.c: ...

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

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

Compiler output

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

try.c: Alarm clock

Number of similar (compiler,implementation) pairs: 6, namely:

Compiler	Implementations
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	rust_crypto
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	rust_crypto
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	rust_crypto
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	rust_sha2
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	rust_sha2
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	rust_sha2