Implementation notes: amd64, nucnuc, crypto_decode/653x3

Computer: nucnuc
Microarchitecture: amd64; Airmont (406c3)
Architecture: amd64
CPU ID: GenuineIntel-000406c3-bfebfbff
SUPERCOP version: 20230530
Operation: crypto_decode
Primitive: 653x3

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
390	2850 0 0	14536 780 952	`ref`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20221010	20221005
1862	147 0 0	12432 812 888	`ref`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20221010	20221005
1866	147 0 0	11232 812 888	`ref`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20221010	20221005
1866	147 0 0	12880 812 888	`ref`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20221010	20221005
1867	147 0 0	10222 804 888	`ref`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20221010	20221005
1903	150 0 0	10680 780 952	`ref`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20221010	20221005
2354	150 0 0	10255 772 952	`ref`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20221010	20221005
2526	126 0 0	9694 804 888	`ref`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20221010	20221005
2578	129 0 0	9259 756 920	`ref`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20221010	20221005

Compiler output

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

decode.c: decode.c:18:18: error: always_inline function '_mm256_loadu_si256' requires target feature 'avx', but would be inlined into function 'crypto_decode_653x3_avx_constbranchindex' that is compiled without support for 'avx'
decode.c: __m256i s0 = _mm256_loadu_si256((const __m256i *) s);
decode.c: ^
decode.c: decode.c:18:18: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
decode.c: decode.c:22:39: error: always_inline function '_mm256_set1_epi8' requires target feature 'avx', but would be inlined into function 'crypto_decode_653x3_avx_constbranchindex' that is compiled without support for 'avx'
decode.c: __m256i s1 = _mm256_srli_epi16(s0&_mm256_set1_epi8(-16),4);
decode.c: ^
decode.c: decode.c:22:39: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
decode.c: decode.c:22:18: error: always_inline function '_mm256_srli_epi16' requires target feature 'avx2', but would be inlined into function 'crypto_decode_653x3_avx_constbranchindex' that is compiled without support for 'avx2'
decode.c: __m256i s1 = _mm256_srli_epi16(s0&_mm256_set1_epi8(-16),4);
decode.c: ^
decode.c: decode.c:22:18: error: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
decode.c: decode.c:23:11: error: always_inline function '_mm256_set1_epi8' requires target feature 'avx', but would be inlined into function 'crypto_decode_653x3_avx_constbranchindex' that is compiled without support for 'avx'
decode.c: s0 &= _mm256_set1_epi8(15);
decode.c: ^
decode.c: decode.c:23:11: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
decode.c: decode.c:25:18: error: always_inline function '_mm256_unpacklo_epi8' requires target feature 'avx2', but would be inlined into function 'crypto_decode_653x3_avx_constbranchindex' that is compiled without support for 'avx2'
decode.c: __m256i a0 = _mm256_unpacklo_epi8(s0,s1);
decode.c: ^
decode.c: decode.c:25:18: error: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
decode.c: decode.c:28:18: error: always_inline function '_mm256_unpackhi_epi8' requires target feature 'avx2', but would be inlined into function 'crypto_decode_653x3_avx_constbranchindex' that is compiled without support for 'avx2'
decode.c: __m256i a1 = _mm256_unpackhi_epi8(s0,s1);
decode.c: ^
decode.c: decode.c:28:18: error: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
decode.c: decode.c:32:39: error: always_inline function '_mm256_set1_epi8' requires target feature 'avx', but would be inlined into function 'crypto_decode_653x3_avx_constbranchindex' that is compiled without support for 'avx'
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 'crypto_decode_653x3_avx_constbranchindex':
decode.c: decode.c:18:13: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
decode.c: 18 | __m256i s0 = _mm256_loadu_si256((const __m256i *) s);
decode.c: | ^~
decode.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:51,
decode.c: from decode.c:1:
decode.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
decode.c: 926 | _mm256_storeu_si256 (__m256i_u *__P, __m256i __A)
decode.c: | ^~~~~~~~~~~~~~~~~~~
decode.c: decode.c:61:5: note: called from here
decode.c: 61 | _mm256_storeu_si256((__m256i *) (f+96),f3);
decode.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
decode.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:51,
decode.c: from decode.c:1:
decode.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
decode.c: 926 | _mm256_storeu_si256 (__m256i_u *__P, __m256i __A)
decode.c: | ^~~~~~~~~~~~~~~~~~~
decode.c: decode.c:60:5: note: called from here
decode.c: 60 | _mm256_storeu_si256((__m256i *) (f+64),f2);
decode.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
decode.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:51,
decode.c: from decode.c:1:
decode.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
decode.c: 926 | _mm256_storeu_si256 (__m256i_u *__P, __m256i __A)
decode.c: | ^~~~~~~~~~~~~~~~~~~
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