Implementation notes: amd64, cherry, crypto_kem/hqcrmrs128

Computer: cherry
Microarchitecture: amd64; Silvermont (406c4)
Architecture: amd64
CPU ID: GenuineIntel-000406c4-bfebfbff
SUPERCOP version: 20240625
Operation: crypto_kem
Primitive: hqcrmrs128

Compiler output

gf2x.c: gf2x.c:367:11: error: always_inline function '_mm256_lddqu_si256' requires target feature 'avx', but would be inlined into function 'TOOM3Mult' that is compiled without support for 'avx'
gf2x.c:                 U0[i] = _mm256_lddqu_si256((__m256i const *)(& A[i4]));
gf2x.c:                         ^
gf2x.c: gf2x.c:367:11: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
gf2x.c: gf2x.c:368:11: error: always_inline function '_mm256_lddqu_si256' requires target feature 'avx', but would be inlined into function 'TOOM3Mult' that is compiled without support for 'avx'
gf2x.c:                 V0[i] = _mm256_lddqu_si256((__m256i const *)(& B[i4]));
gf2x.c:                         ^
gf2x.c: gf2x.c:368:11: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
gf2x.c: gf2x.c:369:11: error: always_inline function '_mm256_lddqu_si256' requires target feature 'avx', but would be inlined into function 'TOOM3Mult' that is compiled without support for 'avx'
gf2x.c:                 U1[i] = _mm256_lddqu_si256((__m256i const *)(& A[i42 + T_TM3_3W_64]));
gf2x.c:                         ^
gf2x.c: gf2x.c:369:11: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
gf2x.c: gf2x.c:370:11: error: always_inline function '_mm256_lddqu_si256' requires target feature 'avx', but would be inlined into function 'TOOM3Mult' that is compiled without support for 'avx'
gf2x.c:                 V1[i] = _mm256_lddqu_si256((__m256i const *)(& B[i42 + T_TM3_3W_64]));
gf2x.c:                         ^
gf2x.c: gf2x.c:370:11: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
gf2x.c: gf2x.c:371:11: error: always_inline function '_mm256_lddqu_si256' requires target feature 'avx', but would be inlined into function 'TOOM3Mult' that is compiled without support for 'avx'
gf2x.c:                 U2[i] = _mm256_lddqu_si256((__m256i const *)(& A[i4 + T2 - 4]));
gf2x.c:                         ^
gf2x.c: gf2x.c:371:11: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
gf2x.c: gf2x.c:372:11: error: always_inline function '_mm256_lddqu_si256' requires target feature 'avx', but would be inlined into function 'TOOM3Mult' that is compiled without support for 'avx'
gf2x.c:                 V2[i] = _mm256_lddqu_si256((__m256i const *)(& B[i4 + T2 - 4]));
gf2x.c:                         ^
gf2x.c: gf2x.c:372:11: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
gf2x.c: gf2x.c:416:11: error: always_inline function '_mm256_lddqu_si256' requires target feature 'avx', but would be inlined into function 'TOOM3Mult' that is compiled without support for 'avx'
gf2x.c: ...
Number of similar (implementation,compiler) pairs: 4, namely:
ImplementationCompiler
T:avxclang -march=native -O2 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1)
T:avxclang -march=native -O3 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1)
T:avxclang -march=native -O -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1)
T:avxclang -march=native -Os -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1)

Compiler output

gf.c: gf.c:116:16: error: '__builtin_ia32_pclmulqdq128' needs target feature pclmul
gf.c:         __m128i vab = _mm_clmulepi64_si128(va, vb, 0);
gf.c:                       ^
gf.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/__wmmintrin_pclmul.h:45:13: note: expanded from macro '_mm_clmulepi64_si128'
gf.c:   ((__m128i)__builtin_ia32_pclmulqdq128((__v2di)(__m128i)(X), \
gf.c:             ^
gf.c: 1 error generated.
Number of similar (implementation,compiler) pairs: 1, namely:
ImplementationCompiler
T:avxclang -mcpu=native -O3 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1)

Compiler output

gf2x.c: gf2x.c: In function 'TOOM3Mult':
gf2x.c: gf2x.c:367:9: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
gf2x.c:   367 |   U0[i] = _mm256_lddqu_si256((__m256i const *)(& A[i4]));
gf2x.c:       |   ~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
gf2x.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:53,
gf2x.c:                  from gf2x.c:11:
gf2x.c: gf2x.c: In function 'karat_mult_2':
gf2x.c: /usr/lib/gcc/x86_64-linux-gnu/10/include/avx2intrin.h:913:1: error: inlining failed in call to 'always_inline' '_mm256_xor_si256': target specific option mismatch
gf2x.c:   913 | _mm256_xor_si256 (__m256i __A, __m256i __B)
gf2x.c:       | ^~~~~~~~~~~~~~~~
gf2x.c: gf2x.c:162:19: note: called from here
gf2x.c:   162 |  __m256i middle = _mm256_xor_si256(D0[1], D2[0]);
gf2x.c:       |                   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Number of similar (implementation,compiler) pairs: 4, namely:
ImplementationCompiler
T:avxgcc -march=native -mtune=native -O2 -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110)
T:avxgcc -march=native -mtune=native -O3 -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110)
T:avxgcc -march=native -mtune=native -O -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110)
T:avxgcc -march=native -mtune=native -Os -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110)