Implementation notes: amd64, jasper3, crypto_aead/aeadaes128ocbtaglen128v1
Computer: jasper3
Microarchitecture: amd64; Tremont (906c0)
Architecture: amd64
CPU ID: GenuineIntel-000906c0-20-bfebfbff
SUPERCOP version: 20240625
Operation: crypto_aead
Primitive: aeadaes128ocbtaglen128v1
| Time | Object size | Test size | Implementation | Compiler | Benchmark date | SUPERCOP version |
|---|
| 4012 | 14714 21 0 | 36848 828 1080 | T:opt | gcc_-march=native_-mtune=native_-O3_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 4599 | 9083 0 0 | 31008 780 1080 | T:dolbeau/aesenc-int | gcc_-march=native_-mtune=native_-O3_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 4714 | 6290 0 0 | 25830 804 1016 | T:dolbeau/aesenc-int | clang_-march=native_-Os_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 4859 | 13828 21 0 | 35408 860 1016 | T:opt | clang_-march=native_-O2_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 4966 | 13828 21 0 | 37712 860 1016 | T:opt | clang_-march=native_-O3_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 5067 | 6347 0 0 | 27136 780 1080 | T:dolbeau/aesenc-int | gcc_-march=native_-mtune=native_-O2_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 5375 | 8300 0 0 | 29728 812 1016 | T:dolbeau/aesenc-int | clang_-march=native_-O2_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 5376 | 6439 0 0 | 26502 804 1016 | T:dolbeau/aesenc-int | clang_-march=native_-O_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 5409 | 8588 0 0 | 32320 812 1016 | T:dolbeau/aesenc-int | clang_-march=native_-O3_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 5693 | 6023 0 0 | 26255 772 1080 | T:dolbeau/aesenc-int | gcc_-march=native_-mtune=native_-O_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 5718 | 4077 0 0 | 22907 756 1048 | T:dolbeau/aesenc-int | gcc_-march=native_-mtune=native_-Os_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 9809 | 6720 18 0 | 30061 920 1016 | T:opt | clang_-mcpu=native_-O3_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 10198 | 7701 21 0 | 28656 828 1080 | T:opt | gcc_-march=native_-mtune=native_-O2_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 10354 | 7970 21 0 | 28375 820 1080 | T:opt | gcc_-march=native_-mtune=native_-O_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 11372 | 7055 21 0 | 26750 852 1016 | T:opt | clang_-march=native_-Os_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 11894 | 7617 21 0 | 27846 852 1016 | T:opt | clang_-march=native_-O_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 13177 | 5853 21 0 | 24827 804 1048 | T:opt | gcc_-march=native_-mtune=native_-Os_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 121024 | 5388 0 0 | 27174 844 1016 | T:ref | clang_-march=native_-O2_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 126989 | 5628 0 0 | 29718 844 1016 | T:ref | clang_-march=native_-O3_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 143471 | 9243 0 0 | 31559 828 1048 | T:ref | gcc_-march=native_-mtune=native_-O3_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 157217 | 9772 0 0 | 32494 844 1016 | T:ref | clang_-mcpu=native_-O3_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 166347 | 2217 0 0 | 22132 836 1016 | T:ref | clang_-march=native_-Os_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 168380 | 3659 0 0 | 24831 828 1048 | T:ref | gcc_-march=native_-mtune=native_-O2_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 177751 | 2305 0 0 | 22692 836 1016 | T:ref | clang_-march=native_-O_-fwrapv_-Qunused-arguments_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 182106 | 3209 0 0 | 23838 820 1048 | T:ref | gcc_-march=native_-mtune=native_-O_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
| 191483 | 1960 0 0 | 21146 804 1048 | T:ref | gcc_-march=native_-mtune=native_-Os_-fwrapv_-fPIC_-fPIE_-gdwarf-4_-Wall | 20240626 | 20240625 |
Compiler output
encrypt.c: encrypt.c:74:34: warning: incompatible pointer types passing 'const unsigned int *' to parameter of type 'const __m128i_u *' [-Wincompatible-pointer-types]
encrypt.c: __m128i key0 = _mm_loadu_si128((const unsigned int *)(key+0));
encrypt.c: ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
encrypt.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/emmintrin.h:3548:34: note: passing argument to parameter '__p' here
encrypt.c: _mm_loadu_si128(__m128i_u const *__p)
encrypt.c: ^
encrypt.c: encrypt.c:75:25: warning: unused variable 'temp2' [-Wunused-variable]
encrypt.c: __m128i temp0, temp1, temp2, temp4;
encrypt.c: ^
encrypt.c: encrypt.c:448:2: warning: misleading indentation; statement is not part of the previous 'for' [-Wmisleading-indentation]
encrypt.c: break;
encrypt.c: ^
encrypt.c: encrypt.c:446:7: note: previous statement is here
encrypt.c: for (i = 5; i < ntz ; i++)
encrypt.c: ^
encrypt.c: encrypt.c:617:25: warning: variable 'sum' is uninitialized when used here [-Wuninitialized]
encrypt.c: sum = _mm_xor_si128(sum,sum);
encrypt.c: ^~~
encrypt.c: encrypt.c:559:5: note: variable 'sum' is declared here
encrypt.c: __m128i lstar, ldollar, sum, offset, ktop, pad, nonce, tag, tmp, outv;
encrypt.c: ^
encrypt.c: encrypt.c:104:20: warning: unused function 'aes128ni_setkey_decrypt' [-Wunused-function]
encrypt.c: static inline void aes128ni_setkey_decrypt(const unsigned char* key, __m128i rkeys[11]) {
encrypt.c: ^
encrypt.c: encrypt.c:136:20: warning: unused function 'aes128ni_encrypt1' [-Wunused-function]
encrypt.c: ...
Number of similar (implementation,compiler) pairs: 5, namely:
| Implementation | Compiler |
| T:dolbeau/aesenc-int | clang -march=native -O2 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1) |
| T:dolbeau/aesenc-int | clang -march=native -O3 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1) |
| T:dolbeau/aesenc-int | clang -march=native -O -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1) |
| T:dolbeau/aesenc-int | clang -march=native -Os -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1) |
| T:dolbeau/aesenc-int | clang -mcpu=native -O3 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1) |
Compiler output
encrypt.c: encrypt.c: In function 'aes128ni_setkey_encrypt':
encrypt.c: encrypt.c:74:34: warning: passing argument 1 of '_mm_loadu_si128' from incompatible pointer type [-Wincompatible-pointer-types]
encrypt.c: 74 | __m128i key0 = _mm_loadu_si128((const unsigned int *)(key+0));
encrypt.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
encrypt.c: | |
encrypt.c: | const unsigned int *
encrypt.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/xmmintrin.h:1316,
encrypt.c: from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:29,
encrypt.c: from encrypt.c:45:
encrypt.c: /usr/lib/gcc/x86_64-linux-gnu/10/include/emmintrin.h:701:35: note: expected 'const __m128i_u *' but argument is of type 'const unsigned int *'
encrypt.c: 701 | _mm_loadu_si128 (__m128i_u const *__P)
encrypt.c: | ~~~~~~~~~~~~~~~~~^~~
encrypt.c: encrypt.c:75:25: warning: unused variable 'temp2' [-Wunused-variable]
encrypt.c: 75 | __m128i temp0, temp1, temp2, temp4;
encrypt.c: | ^~~~~
encrypt.c: encrypt.c: In function 'aes128ni_setkey_decrypt':
encrypt.c: encrypt.c:109: warning: ignoring '#pragma unroll ' [-Wunknown-pragmas]
encrypt.c: 109 | #pragma unroll(9)
encrypt.c: |
encrypt.c: encrypt.c: In function 'aes128ni_setkey_e2d':
encrypt.c: encrypt.c:118: warning: ignoring '#pragma unroll ' [-Wunknown-pragmas]
encrypt.c: 118 | #pragma unroll(9)
encrypt.c: |
encrypt.c: encrypt.c: In function 'aes128ni_encrypt1_si128':
encrypt.c: encrypt.c:129: warning: ignoring '#pragma unroll ' [-Wunknown-pragmas]
encrypt.c: ...
Number of similar (implementation,compiler) pairs: 4, namely:
| Implementation | Compiler |
| T:dolbeau/aesenc-int | gcc -march=native -mtune=native -O2 -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110) |
| T:dolbeau/aesenc-int | gcc -march=native -mtune=native -O3 -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110) |
| T:dolbeau/aesenc-int | gcc -march=native -mtune=native -O -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110) |
| T:dolbeau/aesenc-int | gcc -march=native -mtune=native -Os -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110) |
Compiler output
ocb_vaes.c: ocb_vaes.c:668:19: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'ae_encrypt' that is compiled without support for 'avx2'
ocb_vaes.c: k256[i] = _mm256_broadcastsi128_si256(load128(ctx->encrypt_key.rd_key+i));
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:668:19: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
ocb_vaes.c: ocb_vaes.c:669:14: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'ae_encrypt' that is compiled without support for 'avx2'
ocb_vaes.c: m[M01] = _mm256_broadcastsi128_si256(xor128(load128(ctx->L+0), load128(ctx->L+1)));
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:669:14: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
ocb_vaes.c: ocb_vaes.c:670:14: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'ae_encrypt' that is compiled without support for 'avx2'
ocb_vaes.c: m[M02] = _mm256_broadcastsi128_si256(xor128(load128(ctx->L+0), load128(ctx->L+2)));
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:670:14: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
ocb_vaes.c: ocb_vaes.c:671:14: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'ae_encrypt' that is compiled without support for 'avx2'
ocb_vaes.c: m[M03] = _mm256_broadcastsi128_si256(xor128(load128(ctx->L+0), load128(ctx->L+3)));
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:671:14: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
ocb_vaes.c: ocb_vaes.c:672:16: error: always_inline function '_mm256_set_m128i' requires target feature 'avx', but would be inlined into function 'ae_encrypt' that is compiled without support for 'avx'
ocb_vaes.c: m[M0_01] = _mm256_set_m128i(xor128(load128(ctx->L+0), load128(ctx->L+1)), load128(ctx->L+0));
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:672:16: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
ocb_vaes.c: ocb_vaes.c:677:20: error: always_inline function '_mm256_setzero_si256' requires target feature 'avx', but would be inlined into function 'ae_encrypt' that is compiled without support for 'avx'
ocb_vaes.c: checksum = zero256();
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:150:27: note: expanded from macro 'zero256'
ocb_vaes.c: #define zero256 _mm256_setzero_si256
ocb_vaes.c: ...
Number of similar (implementation,compiler) pairs: 4, namely:
| Implementation | Compiler |
| T:vaes | clang -march=native -O2 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1) |
| T:vaes | clang -march=native -O3 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1) |
| T:vaes | clang -march=native -O -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1) |
| T:vaes | clang -march=native -Os -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1) |
Compiler output
ocb_vaes.c: ocb_vaes.c:476:15: error: always_inline function '_mm_shuffle_epi8' requires target feature 'ssse3', but would be inlined into function 'ae_init' that is compiled without support for 'ssse3'
ocb_vaes.c: tmp_blk = reverse_bytes(load128(&ctx->Lstar));
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:155:5: note: expanded from macro 'reverse_bytes'
ocb_vaes.c: _mm_shuffle_epi8(b,_mm_set_epi8(0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15))
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:478:29: error: always_inline function '_mm_shuffle_epi8' requires target feature 'ssse3', but would be inlined into function 'ae_init' that is compiled without support for 'ssse3'
ocb_vaes.c: store128(&ctx->Ldollar, reverse_bytes(tmp_blk));
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:155:5: note: expanded from macro 'reverse_bytes'
ocb_vaes.c: _mm_shuffle_epi8(b,_mm_set_epi8(0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15))
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:480:24: error: always_inline function '_mm_shuffle_epi8' requires target feature 'ssse3', but would be inlined into function 'ae_init' that is compiled without support for 'ssse3'
ocb_vaes.c: store128(ctx->L+0, reverse_bytes(tmp_blk));
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:155:5: note: expanded from macro 'reverse_bytes'
ocb_vaes.c: _mm_shuffle_epi8(b,_mm_set_epi8(0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15))
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:483:25: error: always_inline function '_mm_shuffle_epi8' requires target feature 'ssse3', but would be inlined into function 'ae_init' that is compiled without support for 'ssse3'
ocb_vaes.c: store128(ctx->L+i, reverse_bytes(tmp_blk));
ocb_vaes.c: ^
ocb_vaes.c: ocb_vaes.c:155:5: note: expanded from macro 'reverse_bytes'
ocb_vaes.c: _mm_shuffle_epi8(b,_mm_set_epi8(0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15))
ocb_vaes.c: ^
ocb_vaes.c: 4 errors generated.
Number of similar (implementation,compiler) pairs: 1, namely:
| Implementation | Compiler |
| T:vaes | clang -mcpu=native -O3 -fwrapv -Qunused-arguments -fPIC -fPIE -gdwarf-4 -Wall (Debian_Clang_11.0.1) |
Compiler output
ocb_vaes.c: ocb_vaes.c: In function 'ae_encrypt':
ocb_vaes.c: ocb_vaes.c:668:17: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
ocb_vaes.c: 668 | k256[i] = _mm256_broadcastsi128_si256(load128(ctx->encrypt_key.rd_key+i));
ocb_vaes.c: | ~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
ocb_vaes.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:51,
ocb_vaes.c: from ocb_vaes.c:71:
ocb_vaes.c: ocb_vaes.c: In function 'process_ad':
ocb_vaes.c: /usr/lib/gcc/x86_64-linux-gnu/10/include/avxintrin.h:1521:1: error: inlining failed in call to 'always_inline' '_mm256_set_m128i': target specific option mismatch
ocb_vaes.c: 1521 | _mm256_set_m128i (__m128i __H, __m128i __L)
ocb_vaes.c: | ^~~~~~~~~~~~~~~~
ocb_vaes.c: ocb_vaes.c:554:13: note: called from here
ocb_vaes.c: 554 | oa[7] = _mm256_set_m128i(hi, lo);
ocb_vaes.c: | ^~~~~~~~~~~~~~~~~~~~~~~~
ocb_vaes.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:51,
ocb_vaes.c: from ocb_vaes.c:71:
ocb_vaes.c: /usr/lib/gcc/x86_64-linux-gnu/10/include/avxintrin.h:1460:1: error: inlining failed in call to 'always_inline' '_mm256_castsi256_si128': target specific option mismatch
ocb_vaes.c: 1460 | _mm256_castsi256_si128 (__m256i __A)
ocb_vaes.c: | ^~~~~~~~~~~~~~~~~~~~~~
ocb_vaes.c: ocb_vaes.c:139:27: note: called from here
ocb_vaes.c: 139 | #define xor128(x,y) _mm_xor_si128(x,y)
ocb_vaes.c: | ^~~~~~~~~~~~~~~~~~
ocb_vaes.c: ocb_vaes.c:552:10: note: in expansion of macro 'xor128'
ocb_vaes.c: 552 | lo = xor128(_mm256_castsi256_si128(oa[6]), _mm256_castsi256_si128(m[M01]));
ocb_vaes.c: | ^~~~~~
ocb_vaes.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:51,
ocb_vaes.c: ...
Number of similar (implementation,compiler) pairs: 4, namely:
| Implementation | Compiler |
| T:vaes | gcc -march=native -mtune=native -O2 -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110) |
| T:vaes | gcc -march=native -mtune=native -O3 -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110) |
| T:vaes | gcc -march=native -mtune=native -O -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110) |
| T:vaes | gcc -march=native -mtune=native -Os -fwrapv -fPIC -fPIE -gdwarf-4 -Wall (10.2.1_20210110) |