Implementation notes: amd64, hydra5, crypto_scalarmult/kummer

Computer: hydra5
Microarchitecture: amd64; K10 32nm (300f10)
Architecture: amd64
CPU ID: AuthenticAMD-00300f10-178bfbff
SUPERCOP version: 20240107
Operation: crypto_scalarmult
Primitive: kummer

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
1658530	6834 0 0	28757 812 1792	`ref5u`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
1713108	13943 0 0	35312 868 1728	`ref5`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
1713628	13943 0 0	35312 868 1728	`ref5`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
1716558	14185 0 0	35528 868 1728	`ref5u`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
1716566	14185 0 0	35528 868 1728	`ref5u`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
1744339	6591 0 0	28485 812 1792	`ref5`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
1766289	4999 0 0	25621 812 1792	`ref5u`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
1769385	12077 0 0	32528 868 1728	`ref5`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
1769842	4334 0 0	22658 860 1728	`ref5`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
1775634	12471 0 0	32936 868 1728	`ref5u`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
1781124	12245 0 0	31272 868 1728	`ref5`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
1793576	4675 0 0	25293 812 1792	`ref5`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
1815919	4629 0 0	22962 860 1728	`ref5u`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
1829681	12479 0 0	31504 868 1728	`ref5u`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231215	20231212
1926673	4446 0 0	22773 804 1760	`ref5`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
1931385	4747 0 0	23125 804 1760	`ref5u`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
2009804	4768 0 0	24461 812 1792	`ref5`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212
2034297	5078 0 0	24789 812 1792	`ref5u`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231215	20231212

Test failure

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

error 111

Number of similar (compiler,implementation) pairs: 16, 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
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	avx
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	avx
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

Compiler output

Implementation: avx
Security model: constbranchindex
Compiler: gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE

try.c: /usr/bin/ld: /home/djb/benchmarking/supercop-20231212/supercop-data/hydra5/amd64/lib/libsupercop.a(crypto_rng_chacha20_ref_constbranchindex-rng.o) symbol number 4 references nonexistent SHT_SYMTAB_SHNDX section
try.c: /usr/bin/ld: failed to set dynamic section sizes: Success
try.c: collect2: error: ld returned 1 exit status

Number of similar (compiler,implementation) pairs: 1, namely:

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

Compiler output

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

try.c: /usr/bin/ld: error in /home/djb/benchmarking/supercop-20231212/supercop-data/hydra5/amd64/lib/libsupercop.a(crypto_rng_chacha20_ref_constbranchindex-rng.o)(.eh_frame); no .eh_frame_hdr table will be created
try.c: /usr/bin/ld: error in /home/djb/benchmarking/supercop-20231212/supercop-data/hydra5/amd64/lib/libsupercop.a(crypto_stream_chacha20_moon_sse2_64_constbranchindex-crypto_stream.o)(.eh_frame); no .eh_frame_hdr table will be created
try.c: /usr/bin/ld: /home/djb/benchmarking/supercop-20231212/supercop-data/hydra5/amd64/lib/libsupercop.a(crypto_rng_chacha20_ref_constbranchindex-rng.o): invalid string offset 1259409155 >= 124 for section `.strtab'
try.c: /usr/bin/ld: /home/djb/benchmarking/supercop-20231212/supercop-data/hydra5/amd64/lib/libsupercop.a(crypto_rng_chacha20_ref_constbranchindex-rng.o): invalid string offset 1259409155 >= 124 for section `.strtab'
try.c: collect2: error: ld returned 1 exit status

Number of similar (compiler,implementation) pairs: 1, namely:

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

Compiler output

Implementation: avx2int
Security model: constbranchindex
Compiler: clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

smult.c: smult.c:36:8: error: always_inline function '_mm256_mul_epi32' requires target feature 'avx2', but would be inlined into function 'crypto_scalarmult_kummer_avx2int_constbranchindex_gfe4x_mulconst' that is compiled without support for 'avx2'
smult.c: t0 = _mm256_mul_epi32(a->v[0],*b);
smult.c: ^
smult.c: smult.c:36:8: error: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
smult.c: smult.c:37:8: error: always_inline function '_mm256_mul_epi32' requires target feature 'avx2', but would be inlined into function 'crypto_scalarmult_kummer_avx2int_constbranchindex_gfe4x_mulconst' that is compiled without support for 'avx2'
smult.c: t1 = _mm256_mul_epi32(a->v[1],*b);
smult.c: ^
smult.c: smult.c:37:8: error: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
smult.c: smult.c:38:30: error: always_inline function '_mm256_srli_epi64' requires target feature 'avx2', but would be inlined into function 'crypto_scalarmult_kummer_avx2int_constbranchindex_gfe4x_mulconst' that is compiled without support for 'avx2'
smult.c: t1 = _mm256_add_epi64(t1,_mm256_srli_epi64(t0,26)); t0 &= mask26;
smult.c: ^
smult.c: smult.c:38:30: error: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
smult.c: smult.c:38:10: error: always_inline function '_mm256_add_epi64' requires target feature 'avx2', but would be inlined into function 'crypto_scalarmult_kummer_avx2int_constbranchindex_gfe4x_mulconst' that is compiled without support for 'avx2'
smult.c: t1 = _mm256_add_epi64(t1,_mm256_srli_epi64(t0,26)); t0 &= mask26;
smult.c: ^
smult.c: smult.c:38:10: error: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
smult.c: smult.c:39:8: error: always_inline function '_mm256_mul_epi32' requires target feature 'avx2', but would be inlined into function 'crypto_scalarmult_kummer_avx2int_constbranchindex_gfe4x_mulconst' that is compiled without support for 'avx2'
smult.c: t2 = _mm256_mul_epi32(a->v[2],*b);
smult.c: ^
smult.c: smult.c:39:8: error: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
smult.c: smult.c:40:8: error: always_inline function '_mm256_mul_epi32' requires target feature 'avx2', but would be inlined into function 'crypto_scalarmult_kummer_avx2int_constbranchindex_gfe4x_mulconst' that is compiled without support for 'avx2'
smult.c: t3 = _mm256_mul_epi32(a->v[3],*b);
smult.c: ^
smult.c: smult.c:40:8: error: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
smult.c: smult.c:41:8: error: always_inline function '_mm256_mul_epi32' requires target feature 'avx2', but would be inlined into function 'crypto_scalarmult_kummer_avx2int_constbranchindex_gfe4x_mulconst' that is compiled without support for 'avx2'
smult.c: ...

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

Compiler	Implementations
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	avx2int
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	avx2int
clang -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	avx2int
clang -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	avx2int
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	avx2int

Compiler output

Implementation: avx2int
Security model: constbranchindex
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

smult.c: smult.c: In function 'crypto_scalarmult_kummer_avx2int_constbranchindex_gfe4x_mulconst':
smult.c: smult.c:36:6: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
smult.c: 36 | t0 = _mm256_mul_epi32(a->v[0],*b);
smult.c: | ~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
smult.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
smult.c: from gfe4x.h:5,
smult.c: from smult.c:3:
smult.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h:126:1: error: inlining failed in call to 'always_inline' '_mm256_add_epi64': target specific option mismatch
smult.c: 126 | _mm256_add_epi64 (__m256i __A, __m256i __B)
smult.c: | ^~~~~~~~~~~~~~~~
smult.c: smult.c:45:10: note: called from here
smult.c: 45 | t2 = _mm256_add_epi64(t2,_mm256_srli_epi64(t1,25)); t1 &= mask25;
smult.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
smult.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
smult.c: from gfe4x.h:5,
smult.c: from smult.c:3:
smult.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/avx2intrin.h:787:1: error: inlining failed in call to 'always_inline' '_mm256_srli_epi64': target specific option mismatch
smult.c: 787 | _mm256_srli_epi64 (__m256i __A, int __B)
smult.c: | ^~~~~~~~~~~~~~~~~
smult.c: smult.c:45:10: note: called from here
smult.c: 45 | t2 = _mm256_add_epi64(t2,_mm256_srli_epi64(t1,25)); t1 &= mask25;
smult.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
smult.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:47,
smult.c: from gfe4x.h:5,
smult.c: from smult.c:3:
smult.c: ...

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

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	avx2int
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	avx2int
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	avx2int
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	avx2int