Implementation notes: amd64, margaux, crypto_kem/mceliece8192128

Computer: margaux
Microarchitecture: amd64; Core 2 65nm (6fb)
Architecture: amd64
CPU ID: GenuineIntel-000006fb-bfebfbff
SUPERCOP version: 20240107
Operation: crypto_kem
Primitive: mceliece8192128

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
2895244	113687 0 0	206764 820 1760	`vec`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231229	20231212
3379233	136464 0 0	230302 860 1728	`vec`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231229	20231212
3434561	135820 0 0	229854 860 1728	`vec`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231229	20231212
3654886	74107 0 0	165990 860 1728	`vec`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231229	20231212
3657185	94578 0 0	187534 860 1728	`vec`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231229	20231212
3776158	70544 0 0	161336 852 1728	`vec`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231229	20231212
3869727	77155 0 0	169676 820 1760	`vec`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231229	20231212
4002791	75934 0 0	167804 820 1760	`vec`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231229	20231212
4370353	67760 0 0	158508 812 1728	`vec`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231229	20231212
660801480	31919 0 0	125024 868 1728	`ref`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231229	20231212
661050709	32401 0 0	125878 828 1760	`ref`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231229	20231212
661384170	46991 0 0	141080 868 1728	`ref`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231229	20231212
661435129	46502 0 0	140760 868 1728	`ref`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231229	20231212
830322799	13837 0 0	104954 860 1728	`ref`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231229	20231212
839350469	14064 0 0	106032 868 1728	`ref`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231229	20231212
853581565	15725 0 0	107476 820 1760	`ref`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231229	20231212
889821792	16115 0 0	108638 828 1760	`ref`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231229	20231212
951707479	13231 0 0	104014 820 1728	`ref`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231229	20231212

Compiler output

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

bm.c: bm.c:90:37: warning: AVX vector return of type 'vec256' (aka '__m256i') without 'avx' enabled changes the ABI [-Wpsabi]
bm.c: for (i = 13; i < 16; i++) buf[i] = vec256_setzero();
bm.c: ^
bm.c: bm.c:92:15: warning: AVX vector return of type 'vec256' (aka '__m256i') without 'avx' enabled changes the ABI [-Wpsabi]
bm.c: mask[0][0] = vec256_set1_16b(0x5555);
bm.c: ^
bm.c: bm.c:93:15: warning: AVX vector return of type 'vec256' (aka '__m256i') without 'avx' enabled changes the ABI [-Wpsabi]
bm.c: mask[0][1] = vec256_set1_16b(0xAAAA);
bm.c: ^
bm.c: bm.c:94:15: warning: AVX vector return of type 'vec256' (aka '__m256i') without 'avx' enabled changes the ABI [-Wpsabi]
bm.c: mask[1][0] = vec256_set1_16b(0x3333);
bm.c: ^
bm.c: bm.c:95:15: warning: AVX vector return of type 'vec256' (aka '__m256i') without 'avx' enabled changes the ABI [-Wpsabi]
bm.c: mask[1][1] = vec256_set1_16b(0xCCCC);
bm.c: ^
bm.c: bm.c:96:15: warning: AVX vector return of type 'vec256' (aka '__m256i') without 'avx' enabled changes the ABI [-Wpsabi]
bm.c: mask[2][0] = vec256_set1_16b(0x0F0F);
bm.c: ^
bm.c: bm.c:97:15: warning: AVX vector return of type 'vec256' (aka '__m256i') without 'avx' enabled changes the ABI [-Wpsabi]
bm.c: mask[2][1] = vec256_set1_16b(0xF0F0);
bm.c: ^
bm.c: bm.c:98:15: warning: AVX vector return of type 'vec256' (aka '__m256i') without 'avx' enabled changes the ABI [-Wpsabi]
bm.c: mask[3][0] = vec256_set1_16b(0x00FF);
bm.c: ^
bm.c: bm.c:99:15: warning: AVX vector return of type 'vec256' (aka '__m256i') without 'avx' enabled changes the ABI [-Wpsabi]
bm.c: ...

Number of similar (compiler,implementation) pairs: 10, namely:

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

Compiler output

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

benes.c: In file included from transpose.h:12,
benes.c: from benes.c:11:
benes.c: vec256.h: In function 'vec256_set1_16b':
benes.c: vec256.h:21:1: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
benes.c: 21 | {
benes.c: | ^
bm.c: In file included from bm.h:11,
bm.c: from bm.c:13:
bm.c: vec256.h: In function 'vec256_set1_16b':
bm.c: vec256.h:21:1: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
bm.c: 21 | {
bm.c: | ^
bm.c: vec256.h: In function 'vec256_and':
bm.c: vec256.h:48:22: note: the ABI for passing parameters with 32-byte alignment has changed in GCC 4.6
bm.c: 48 | static inline vec256 vec256_and(vec256 a, vec256 b)
bm.c: | ^~~~~~~~~~
bm.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/11/include/immintrin.h:43,
bm.c: from vec256.h:16,
bm.c: from bm.h:11,
bm.c: from bm.c:13:
bm.c: vec256.h: In function 'vec256_set1_16b':
bm.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/avxintrin.h:1333:1: error: inlining failed in call to 'always_inline' '_mm256_set1_epi16': target specific option mismatch
bm.c: 1333 | _mm256_set1_epi16 (short __A)
bm.c: | ^~~~~~~~~~~~~~~~~
bm.c: In file included from bm.h:11,
bm.c: from bm.c:13:
bm.c: vec256.h:22:16: note: called from here
bm.c: 22 | return _mm256_set1_epi16(a);
bm.c: | ^~~~~~~~~~~~~~~~~~~~

Number of similar (compiler,implementation) pairs: 8, namely:

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

Compiler output

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

decrypt.c: In file included from decrypt.c:8:
decrypt.c: In file included from ./fft_tr.h:10:
decrypt.c: ./vec128.h:45:9: error: always_inline function '_mm_testz_si128' requires target feature 'sse4.1', but would be inlined into function 'vec128_testz' that is compiled without support for 'sse4.1'
decrypt.c: return _mm_testz_si128(a, a);
decrypt.c: ^
decrypt.c: 1 error generated.

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

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

Compiler output

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

bm.c: In file included from vec128.h:17,
bm.c: from bm.h:11,
bm.c: from bm.c:13:
bm.c: bm.c: In function 'get_coefs':
bm.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/smmintrin.h:455:1: error: inlining failed in call to 'always_inline' '_mm_extract_epi64': target specific option mismatch
bm.c: 455 | _mm_extract_epi64 (__m128i __X, const int __N)
bm.c: | ^~~~~~~~~~~~~~~~~
bm.c: In file included from bm.h:11,
bm.c: from bm.c:13:
bm.c: vec128.h:41:42: note: called from here
bm.c: 41 | #define vec128_extract(a, i) ((uint64_t) _mm_extract_epi64((vec128) (a), (i)))
bm.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
bm.c: bm.c:136:44: note: in expansion of macro 'vec128_extract'
bm.c: 136 | out[ (4*1 + k)*16 + i ] = (vec128_extract(buf[i], 1) >> (k*16)) & GFMASK;
bm.c: | ^~~~~~~~~~~~~~
bm.c: In file included from vec128.h:17,
bm.c: from bm.h:11,
bm.c: from bm.c:13:
bm.c: /usr/lib/gcc/x86_64-linux-gnu/11/include/smmintrin.h:455:1: error: inlining failed in call to 'always_inline' '_mm_extract_epi64': target specific option mismatch
bm.c: 455 | _mm_extract_epi64 (__m128i __X, const int __N)
bm.c: | ^~~~~~~~~~~~~~~~~
bm.c: In file included from bm.h:11,
bm.c: from bm.c:13:
bm.c: vec128.h:41:42: note: called from here
bm.c: 41 | #define vec128_extract(a, i) ((uint64_t) _mm_extract_epi64((vec128) (a), (i)))
bm.c: ...

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

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