Implementation notes: amd64, jasper3, crypto_kem/frodokem640aes

Computer: jasper3
Microarchitecture: amd64; Tremont (906c0)
Architecture: amd64
CPU ID: GenuineIntel-000906c0-20-bfebfbff
SUPERCOP version: 20240625
Operation: crypto_kem
Primitive: frodokem640aes
TimeObject sizeTest sizeImplementationCompilerBenchmark dateSUPERCOP version
506632220094 28 038902 916 1720T:optimizedclang_-march=native_-O2_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall2024070220240625
519041342155 42 061507 916 1752T:optimizedgcc_-march=native_-mtune=native_-O3_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall2024070220240625
574598423699 28 043438 916 1720T:optimizedclang_-mcpu=native_-O3_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall2024070220240625
577148421666 28 042774 916 1720T:optimizedclang_-march=native_-O3_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall2024070220240625
74228619849 28 026828 908 1720T:optimizedclang_-march=native_-Os_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall2024070220240625
2184754910561 42 028819 916 1752T:optimizedgcc_-march=native_-mtune=native_-O2_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall2024070220240625
268485258576 42 025291 908 1720T:optimizedgcc_-march=native_-mtune=native_-Os_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall2024070220240625
356117999883 28 027436 908 1720T:optimizedclang_-march=native_-O_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall2024070220240625
428435409582 42 027427 916 1752T:optimizedgcc_-march=native_-mtune=native_-O_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall2024070220240625

Compiler output

frodo640.c: In file included from frodo640.c:32:
frodo640.c: ./frodo_macrify.inc:105:26: error: always_inline function '_mm256_setzero_si256' requires target feature 'avx', but would be inlined into function 'frodo_mul_add_sa_plus_e' that is compiled without support for 'avx'
frodo640.c:                 acc[0] = _mm256_setzero_si256();
frodo640.c:                          ^
frodo640.c: ./frodo_macrify.inc:105:26: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
frodo640.c: ./frodo_macrify.inc:106:26: error: always_inline function '_mm256_setzero_si256' requires target feature 'avx', but would be inlined into function 'frodo_mul_add_sa_plus_e' that is compiled without support for 'avx'
frodo640.c:                 acc[1] = _mm256_setzero_si256();
frodo640.c:                          ^
frodo640.c: ./frodo_macrify.inc:106:26: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
frodo640.c: ./frodo_macrify.inc:107:26: error: always_inline function '_mm256_setzero_si256' requires target feature 'avx', but would be inlined into function 'frodo_mul_add_sa_plus_e' that is compiled without support for 'avx'
frodo640.c:                 acc[2] = _mm256_setzero_si256();
frodo640.c:                          ^
frodo640.c: ./frodo_macrify.inc:107:26: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
frodo640.c: ./frodo_macrify.inc:108:26: error: always_inline function '_mm256_setzero_si256' requires target feature 'avx', but would be inlined into function 'frodo_mul_add_sa_plus_e' that is compiled without support for 'avx'
frodo640.c:                 acc[3] = _mm256_setzero_si256();
frodo640.c:                          ^
frodo640.c: ./frodo_macrify.inc:108:26: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
frodo640.c: ./frodo_macrify.inc:110:25: error: always_inline function '_mm256_load_si256' requires target feature 'avx', but would be inlined into function 'frodo_mul_add_sa_plus_e' that is compiled without support for 'avx'
frodo640.c:                     b = _mm256_load_si256((__m256i*)&s[i*PARAMS_N + j]);
frodo640.c:                         ^
frodo640.c: ./frodo_macrify.inc:110:25: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
frodo640.c: ./frodo_macrify.inc:111:28: error: always_inline function '_mm256_load_si256' requires target feature 'avx', but would be inlined into function 'frodo_mul_add_sa_plus_e' that is compiled without support for 'avx'
frodo640.c:                     a[0] = _mm256_load_si256((__m256i*)&a_cols_t[(k+0)*PARAMS_N + j]);
frodo640.c:                            ^
frodo640.c: ./frodo_macrify.inc:111:28: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
frodo640.c: ...
Number of similar (implementation,compiler) pairs: 4, namely:
ImplementationCompiler
T:x64clang -march=native -O2 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1)
T:x64clang -march=native -O3 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1)
T:x64clang -march=native -O -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1)
T:x64clang -march=native -Os -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1)

Compiler output

aes_ni.c: aes_ni.c:35:16: error: '__builtin_ia32_aeskeygenassist128' needs target feature aes
aes_ni.c:         schedule[1] = key_exp(schedule[0], 0x01);
aes_ni.c:                       ^
aes_ni.c: aes_ni.c:28:40: note: expanded from macro 'key_exp'
aes_ni.c: #define key_exp(k, rcon) key_expand(k, _mm_aeskeygenassist_si128(k, rcon))
aes_ni.c:                                        ^
aes_ni.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/__wmmintrin_aes.h:136:12: note: expanded from macro '_mm_aeskeygenassist_si128'
aes_ni.c:   (__m128i)__builtin_ia32_aeskeygenassist128((__v2di)(__m128i)(C), (int)(R))
aes_ni.c:            ^
aes_ni.c: aes_ni.c:36:16: error: '__builtin_ia32_aeskeygenassist128' needs target feature aes
aes_ni.c:         schedule[2] = key_exp(schedule[1], 0x02);
aes_ni.c:                       ^
aes_ni.c: aes_ni.c:28:40: note: expanded from macro 'key_exp'
aes_ni.c: #define key_exp(k, rcon) key_expand(k, _mm_aeskeygenassist_si128(k, rcon))
aes_ni.c:                                        ^
aes_ni.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/__wmmintrin_aes.h:136:12: note: expanded from macro '_mm_aeskeygenassist_si128'
aes_ni.c:   (__m128i)__builtin_ia32_aeskeygenassist128((__v2di)(__m128i)(C), (int)(R))
aes_ni.c:            ^
aes_ni.c: aes_ni.c:37:16: error: '__builtin_ia32_aeskeygenassist128' needs target feature aes
aes_ni.c:         schedule[3] = key_exp(schedule[2], 0x04);
aes_ni.c:                       ^
aes_ni.c: aes_ni.c:28:40: note: expanded from macro 'key_exp'
aes_ni.c: #define key_exp(k, rcon) key_expand(k, _mm_aeskeygenassist_si128(k, rcon))
aes_ni.c:                                        ^
aes_ni.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/__wmmintrin_aes.h:136:12: note: expanded from macro '_mm_aeskeygenassist_si128'
aes_ni.c: ...
Number of similar (implementation,compiler) pairs: 1, namely:
ImplementationCompiler
T:x64clang -mcpu=native -O3 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1)

Compiler output

frodo640.c: In file included from frodo640.c:32:
frodo640.c: frodo_macrify.inc: In function 'frodo_mul_add_sa_plus_e':
frodo640.c: frodo_macrify.inc:105:24: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
frodo640.c:   105 |                 acc[0] = _mm256_setzero_si256();
frodo640.c:       |                 ~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~
frodo640.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:51,
frodo640.c:                  from frodo_macrify.inc:8,
frodo640.c:                  from frodo640.c:32:
frodo640.c: /usr/lib/gcc/x86_64-linux-gnu/10/include/avxintrin.h:1235:1: error: inlining failed in call to 'always_inline' '_mm256_setzero_si256': target specific option mismatch
frodo640.c:  1235 | _mm256_setzero_si256 (void)
frodo640.c:       | ^~~~~~~~~~~~~~~~~~~~
frodo640.c: In file included from frodo640.c:32:
frodo640.c: frodo_macrify.inc:108:26: note: called from here
frodo640.c:   108 |                 acc[3] = _mm256_setzero_si256();
frodo640.c:       |                          ^~~~~~~~~~~~~~~~~~~~~~
frodo640.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:51,
frodo640.c:                  from frodo_macrify.inc:8,
frodo640.c:                  from frodo640.c:32:
frodo640.c: /usr/lib/gcc/x86_64-linux-gnu/10/include/avxintrin.h:1235:1: error: inlining failed in call to 'always_inline' '_mm256_setzero_si256': target specific option mismatch
frodo640.c:  1235 | _mm256_setzero_si256 (void)
frodo640.c:       | ^~~~~~~~~~~~~~~~~~~~
frodo640.c: In file included from frodo640.c:32:
frodo640.c: frodo_macrify.inc:107:26: note: called from here
frodo640.c:   107 |                 acc[2] = _mm256_setzero_si256();
frodo640.c:       |                          ^~~~~~~~~~~~~~~~~~~~~~
frodo640.c: ...
Number of similar (implementation,compiler) pairs: 4, namely:
ImplementationCompiler
T:x64gcc -march=native -mtune=native -O2 -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110)
T:x64gcc -march=native -mtune=native -O3 -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110)
T:x64gcc -march=native -mtune=native -O -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110)
T:x64gcc -march=native -mtune=native -Os -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110)