Implementation notes: amd64, hunsnivy, crypto_hash/blake2b

Computer: hunsnivy
Microarchitecture: amd64; Ivy Bridge+AES (306a9)
Architecture: amd64
CPU ID: GenuineIntel-000306a9-bfebfbff
SUPERCOP version: 20240107
Operation: crypto_hash
Primitive: blake2b

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
5337	3973 0 0	14574 804 888	`T:moon/avx/64`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
5340	3973 0 0	18416 812 888	`T:moon/avx/64`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
5352	3971 0 0	14102 804 888	`T:moon/avx/64`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
5382	3973 0 0	16744 812 888	`T:moon/avx/64`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
5383	3941 0 0	14431 772 952	`T:moon/avx/64`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
5397	3904 0 0	15864 780 952	`T:moon/avx/64`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
5400	3953 0 0	16752 812 888	`T:moon/avx/64`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
5404	3883 0 0	13459 756 920	`T:moon/avx/64`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
5432	3954 0 0	14888 780 952	`T:moon/avx/64`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
5717	11616 0 0	22183 804 952	`T:avxicc`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
5742	11616 0 0	22592 812 952	`T:avxicc`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
5899	11616 0 0	24400 812 888	`T:avxicc`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
5913	11616 0 0	22230 804 888	`T:avxicc`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
5917	11616 0 0	23584 812 952	`T:avxicc`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
5924	11616 0 0	21758 804 888	`T:avxicc`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
5928	11616 0 0	21187 788 920	`T:avxicc`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
5936	11616 0 0	26072 812 888	`T:avxicc`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
5941	11616 0 0	24392 812 888	`T:avxicc`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
6115	11731 0 0	26176 812 888	`T:xmm`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
6136	11751 0 0	26192 812 888	`T:ymm`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
6214	11731 0 0	22334 804 888	`T:xmm`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
6236	11663 0 0	21790 804 888	`T:ymm`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
6265	11751 0 0	24520 812 888	`T:ymm`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
6281	11641 0 0	21758 804 888	`T:xmm`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
6302	11858 0 0	23792 780 952	`T:ymm`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
6312	11751 0 0	22350 804 888	`T:ymm`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
6334	11858 0 0	22800 780 952	`T:ymm`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
6350	11731 0 0	24504 812 888	`T:xmm`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
6361	11141 0 0	20667 756 920	`T:xmm`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
6388	11287 0 0	22216 780 952	`T:xmm`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
6520	11287 0 0	23208 780 952	`T:xmm`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
6540	11756 0 0	21275 756 920	`T:ymm`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
6740	12056 0 0	22567 772 952	`T:ymm`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
6823	11477 0 0	21983 772 952	`T:xmm`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
7075	13738 0 0	23838 804 888	`T:regs`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
7108	13705 0 0	26464 812 888	`T:regs`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
7178	13705 0 0	28136 812 888	`T:regs`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
7525	12746 0 0	25496 812 888	`T:regs`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
7568	12125 0 0	23048 780 952	`T:regs`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
7577	16893 0 0	29656 812 888	`T:xmm`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
7729	12112 0 0	24040 780 952	`T:regs`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
7740	11619 0 0	21131 756 920	`T:regs`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
8090	12170 0 0	22671 772 952	`T:regs`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
8353	14116 0 0	24718 804 888	`T:regs`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
8513	11242 0 0	25744 812 888	`T:ref`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
8530	11090 0 0	23920 812 888	`T:ref`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
8581	7442 0 0	18424 780 952	`T:ref`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
8650	9700 0 0	19878 804 888	`T:ref`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
8691	8066 0 0	20072 780 952	`T:ref`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
9048	10298 0 0	23120 812 888	`T:ref`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212
10367	7772 0 0	18400 780 952	`T:ref`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
10465	7023 0 0	16595 756 920	`T:ref`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20231213	20231212
28916	20577 0 0	31302 804 888	`T:ref`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20231213	20231212

Test failure

Implementation: T:moon/avx2/64
Security model: timingleaks
Compiler: clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

error 111

Number of similar (compiler,implementation) pairs: 18, namely:

Compiler	Implementations
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:moon/avx2/64
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:moon/avx2/64
clang -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:moon/avx2/64
clang -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:moon/avx2/64
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:moon/avx2/64
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:moon/avx2/64
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:moon/avx2/64
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:moon/avx2/64
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:moon/avx2/64
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:moon/xop/64
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:moon/xop/64
clang -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:moon/xop/64
clang -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:moon/xop/64
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:moon/xop/64
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:moon/xop/64
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:moon/xop/64
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:moon/xop/64
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:moon/xop/64

Compiler output

Implementation: T:avx2-1
Security model: timingleaks
Compiler: clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

blake2b.c: blake2b.c:100:14: error: '__builtin_ia32_pblendd256' needs target feature avx2
blake2b.c: _mm256_blend_epi32(_mm256_castsi128_si256(_mm_cvtsi64_si128(ctr)),
blake2b.c: ^
blake2b.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/avx2intrin.h:750:12: note: expanded from macro '_mm256_blend_epi32'
blake2b.c: (__m256i)__builtin_ia32_pblendd256((__v8si)(__m256i)(V1), \
blake2b.c: ^
blake2b.c: blake2b.c:96:12: error: always_inline function '_mm256_xor_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_1_timingleaks' that is compiled without support for 'avx2'
blake2b.c: ymm3 = _mm256_xor_si256
blake2b.c: ^
blake2b.c: blake2b.c:109:5: error: always_inline function '_mm256_add_epi64' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_1_timingleaks' that is compiled without support for 'avx2'
blake2b.c: ROUNDS(in);
blake2b.c: ^
blake2b.c: ./blake2b-round.h:105:5: note: expanded from macro 'ROUNDS'
blake2b.c: ROUND( 0 ); \
blake2b.c: ^
blake2b.c: ./blake2b-round.h:74:3: note: expanded from macro 'ROUND'
blake2b.c: G1(ymm0,ymm1,ymm2,ymm3,b0); \
blake2b.c: ^
blake2b.c: ./blake2b-round.h:40:27: note: expanded from macro 'G1'
blake2b.c: row1 = _mm256_add_epi64(_mm256_add_epi64(row1, b0), row2); \
blake2b.c: ^
blake2b.c: blake2b.c:109:5: error: always_inline function '_mm256_add_epi64' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_1_timingleaks' that is compiled without support for 'avx2'
blake2b.c: ./blake2b-round.h:105:5: note: expanded from macro 'ROUNDS'
blake2b.c: ROUND( 0 ); \
blake2b.c: ^
blake2b.c: ...

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

Compiler	Implementations
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-1
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-1
clang -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-1
clang -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-1

Compiler output

Implementation: T:avx2-1
Security model: timingleaks
Compiler: clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

blake2b.c: blake2b.c:37:23: error: always_inline function '_mm256_setr_epi8' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_1_timingleaks' that is compiled without support for 'avx'
blake2b.c: const __m256i r16 = _mm256_setr_epi8( 2, 3, 4, 5, 6, 7, 0, 1, 10, 11, 12, 13, 14, 15, 8, 9,
blake2b.c: ^
blake2b.c: blake2b.c:37:23: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:39:23: error: always_inline function '_mm256_setr_epi8' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_1_timingleaks' that is compiled without support for 'avx'
blake2b.c: const __m256i r24 = _mm256_setr_epi8( 3, 4, 5, 6, 7, 0, 1, 2, 11, 12, 13, 14, 15, 8, 9, 10,
blake2b.c: ^
blake2b.c: blake2b.c:39:23: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:56:10: error: always_inline function '_mm256_set_epi64x' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_1_timingleaks' that is compiled without support for 'avx'
blake2b.c: ymm0 = _mm256_set_epi64x(0xA54FF53A5F1D36F1ULL, 0x3C6EF372FE94F82BULL, 0xBB67AE8584CAA73BULL, 0x6A09E667F3BCC908ULL ^ 0x0000000001010040ULL);
blake2b.c: ^
blake2b.c: blake2b.c:56:10: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:57:10: error: always_inline function '_mm256_set_epi64x' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_1_timingleaks' that is compiled without support for 'avx'
blake2b.c: ymm1 = _mm256_set_epi64x(0x5BE0CD19137E2179ULL, 0x1F83D9ABFB41BD6BULL, 0x9B05688C2B3E6C1FULL, 0x510E527FADE682D1ULL);
blake2b.c: ^
blake2b.c: blake2b.c:57:10: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:95:12: error: always_inline function '_mm256_set_epi64x' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_1_timingleaks' that is compiled without support for 'avx'
blake2b.c: ymm2 = _mm256_set_epi64x(0xA54FF53A5F1D36F1ULL, 0x3C6EF372FE94F82BULL, 0xBB67AE8584CAA73BULL, 0x6a09e667f3bcc908ULL);
blake2b.c: ^
blake2b.c: blake2b.c:95:12: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:98:14: error: always_inline function '_mm256_set_epi64x' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_1_timingleaks' that is compiled without support for 'avx'
blake2b.c: _mm256_set_epi64x(0x5BE0CD19137E2179ULL, 0x1F83D9ABFB41BD6BULL, 0x9B05688C2B3E6C1FULL, 0x510E527FADE682D1ULL),
blake2b.c: ^
blake2b.c: blake2b.c:98:14: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:100:14: error: '__builtin_ia32_pblendd256' needs target feature avx2
blake2b.c: ...

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-1

Compiler output

Implementation: T:avx2-1
Security model: timingleaks
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

blake2b.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:53,
blake2b.c: from blake2b.c:11:
blake2b.c: blake2b.c: In function 'crypto_hash_blake2b_avx2_1_timingleaks':
blake2b.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
blake2b.c: 913 | _mm256_xor_si256 (__m256i __A, __m256i __B)
blake2b.c: | ^~~~~~~~~~~~~~~~
blake2b.c: blake2b.c:114:12: note: called from here
blake2b.c: 114 | ymm1 = _mm256_xor_si256(ymm1, iv1);
blake2b.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~
blake2b.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:53,
blake2b.c: from blake2b.c:11:
blake2b.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
blake2b.c: 913 | _mm256_xor_si256 (__m256i __A, __m256i __B)
blake2b.c: | ^~~~~~~~~~~~~~~~
blake2b.c: blake2b.c:113:12: note: called from here
blake2b.c: 113 | ymm0 = _mm256_xor_si256(ymm0, iv0);
blake2b.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~
blake2b.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:53,
blake2b.c: from blake2b.c:11:
blake2b.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
blake2b.c: 913 | _mm256_xor_si256 (__m256i __A, __m256i __B)
blake2b.c: | ^~~~~~~~~~~~~~~~
blake2b.c: blake2b.c:112:12: note: called from here
blake2b.c: 112 | ymm1 = _mm256_xor_si256(ymm1, ymm3);
blake2b.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
blake2b.c: ...

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

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-1
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-1
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-1
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-1

Compiler output

Implementation: T:avx2-2
Security model: timingleaks
Compiler: clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

blake2b.c: blake2b.c:64:24: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx2'
blake2b.c: const __m256i m0 = _mm256_broadcastsi128_si256(LOADU128( in + 00 ));
blake2b.c: ^
blake2b.c: blake2b.c:65:24: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx2'
blake2b.c: const __m256i m1 = _mm256_broadcastsi128_si256(LOADU128( in + 16 ));
blake2b.c: ^
blake2b.c: blake2b.c:66:24: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx2'
blake2b.c: const __m256i m2 = _mm256_broadcastsi128_si256(LOADU128( in + 32 ));
blake2b.c: ^
blake2b.c: blake2b.c:67:24: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx2'
blake2b.c: const __m256i m3 = _mm256_broadcastsi128_si256(LOADU128( in + 48 ));
blake2b.c: ^
blake2b.c: blake2b.c:68:24: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx2'
blake2b.c: const __m256i m4 = _mm256_broadcastsi128_si256(LOADU128( in + 64 ));
blake2b.c: ^
blake2b.c: blake2b.c:69:24: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx2'
blake2b.c: const __m256i m5 = _mm256_broadcastsi128_si256(LOADU128( in + 80 ));
blake2b.c: ^
blake2b.c: blake2b.c:70:24: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx2'
blake2b.c: const __m256i m6 = _mm256_broadcastsi128_si256(LOADU128( in + 96 ));
blake2b.c: ^
blake2b.c: blake2b.c:71:24: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx2'
blake2b.c: const __m256i m7 = _mm256_broadcastsi128_si256(LOADU128( in + 112 ));
blake2b.c: ^
blake2b.c: blake2b.c:101:14: error: '__builtin_ia32_pblendd256' needs target feature avx2
blake2b.c: ...

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

Compiler	Implementations
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-2
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-2
clang -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-2
clang -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-2

Compiler output

Implementation: T:avx2-2
Security model: timingleaks
Compiler: clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

blake2b.c: blake2b.c:38:23: error: always_inline function '_mm256_setr_epi8' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx'
blake2b.c: const __m256i r16 = _mm256_setr_epi8( 2, 3, 4, 5, 6, 7, 0, 1, 10, 11, 12, 13, 14, 15, 8, 9,
blake2b.c: ^
blake2b.c: blake2b.c:38:23: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:40:23: error: always_inline function '_mm256_setr_epi8' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx'
blake2b.c: const __m256i r24 = _mm256_setr_epi8( 3, 4, 5, 6, 7, 0, 1, 2, 11, 12, 13, 14, 15, 8, 9, 10,
blake2b.c: ^
blake2b.c: blake2b.c:40:23: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:57:10: error: always_inline function '_mm256_set_epi64x' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx'
blake2b.c: ymm0 = _mm256_set_epi64x(0xA54FF53A5F1D36F1ULL, 0x3C6EF372FE94F82BULL, 0xBB67AE8584CAA73BULL, 0x6A09E667F3BCC908ULL ^ 0x0000000001010040ULL);
blake2b.c: ^
blake2b.c: blake2b.c:57:10: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:58:10: error: always_inline function '_mm256_set_epi64x' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx'
blake2b.c: ymm1 = _mm256_set_epi64x(0x5BE0CD19137E2179ULL, 0x1F83D9ABFB41BD6BULL, 0x9B05688C2B3E6C1FULL, 0x510E527FADE682D1ULL);
blake2b.c: ^
blake2b.c: blake2b.c:58:10: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:64:24: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx2'
blake2b.c: const __m256i m0 = _mm256_broadcastsi128_si256(LOADU128( in + 00 ));
blake2b.c: ^
blake2b.c: blake2b.c:64:24: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:65:24: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx2'
blake2b.c: const __m256i m1 = _mm256_broadcastsi128_si256(LOADU128( in + 16 ));
blake2b.c: ^
blake2b.c: blake2b.c:65:24: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:66:24: error: always_inline function '_mm256_broadcastsi128_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_2_timingleaks' that is compiled without support for 'avx2'
blake2b.c: ...

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-2

Compiler output

Implementation: T:avx2-2
Security model: timingleaks
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

blake2b.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:53,
blake2b.c: from blake2b.c:12:
blake2b.c: blake2b.c: In function 'crypto_hash_blake2b_avx2_2_timingleaks':
blake2b.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
blake2b.c: 913 | _mm256_xor_si256 (__m256i __A, __m256i __B)
blake2b.c: | ^~~~~~~~~~~~~~~~
blake2b.c: blake2b.c:115:12: note: called from here
blake2b.c: 115 | ymm1 = _mm256_xor_si256(ymm1, iv1);
blake2b.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~
blake2b.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:53,
blake2b.c: from blake2b.c:12:
blake2b.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
blake2b.c: 913 | _mm256_xor_si256 (__m256i __A, __m256i __B)
blake2b.c: | ^~~~~~~~~~~~~~~~
blake2b.c: blake2b.c:114:12: note: called from here
blake2b.c: 114 | ymm0 = _mm256_xor_si256(ymm0, iv0);
blake2b.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~
blake2b.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:53,
blake2b.c: from blake2b.c:12:
blake2b.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
blake2b.c: 913 | _mm256_xor_si256 (__m256i __A, __m256i __B)
blake2b.c: | ^~~~~~~~~~~~~~~~
blake2b.c: blake2b.c:113:12: note: called from here
blake2b.c: 113 | ymm1 = _mm256_xor_si256(ymm1, ymm3);
blake2b.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
blake2b.c: ...

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

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-2
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-2
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-2
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-2

Compiler output

Implementation: T:avx2-3
Security model: timingleaks
Compiler: clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

blake2b.c: blake2b.c:101:14: error: '__builtin_ia32_pblendd256' needs target feature avx2
blake2b.c: _mm256_blend_epi32(_mm256_castsi128_si256(_mm_cvtsi64_si128(ctr)),
blake2b.c: ^
blake2b.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/avx2intrin.h:750:12: note: expanded from macro '_mm256_blend_epi32'
blake2b.c: (__m256i)__builtin_ia32_pblendd256((__v8si)(__m256i)(V1), \
blake2b.c: ^
blake2b.c: blake2b.c:97:12: error: always_inline function '_mm256_xor_si256' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_3_timingleaks' that is compiled without support for 'avx2'
blake2b.c: ymm3 = _mm256_xor_si256
blake2b.c: ^
blake2b.c: blake2b.c:110:5: error: '__builtin_ia32_gatherd_q256' needs target feature avx2
blake2b.c: ROUNDS(in);
blake2b.c: ^
blake2b.c: ./blake2b-round.h:100:23: note: expanded from macro 'ROUNDS'
blake2b.c: #define ROUNDS(msg) GATHER_ROUND(msg)
blake2b.c: ^
blake2b.c: ./blake2b-round.h:87:10: note: expanded from macro 'GATHER_ROUND'
blake2b.c: b0 = _mm256_i32gather_epi64((void *)msg, LOAD128(&indices[16*i + 0]), 8); \
blake2b.c: ^
blake2b.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/avx2intrin.h:1128:12: note: expanded from macro '_mm256_i32gather_epi64'
blake2b.c: (__m256i)__builtin_ia32_gatherd_q256((__v4di)_mm256_undefined_si256(), \
blake2b.c: ^
blake2b.c: blake2b.c:110:5: error: always_inline function '_mm256_add_epi64' requires target feature 'avx2', but would be inlined into function 'crypto_hash_blake2b_avx2_3_timingleaks' that is compiled without support for 'avx2'
blake2b.c: ./blake2b-round.h:100:23: note: expanded from macro 'ROUNDS'
blake2b.c: #define ROUNDS(msg) GATHER_ROUND(msg)
blake2b.c: ^
blake2b.c: ...

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

Compiler	Implementations
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-3
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-3
clang -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-3
clang -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-3

Compiler output

Implementation: T:avx2-3
Security model: timingleaks
Compiler: clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

blake2b.c: blake2b.c:38:23: error: always_inline function '_mm256_setr_epi8' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_3_timingleaks' that is compiled without support for 'avx'
blake2b.c: const __m256i r16 = _mm256_setr_epi8( 2, 3, 4, 5, 6, 7, 0, 1, 10, 11, 12, 13, 14, 15, 8, 9,
blake2b.c: ^
blake2b.c: blake2b.c:38:23: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:40:23: error: always_inline function '_mm256_setr_epi8' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_3_timingleaks' that is compiled without support for 'avx'
blake2b.c: const __m256i r24 = _mm256_setr_epi8( 3, 4, 5, 6, 7, 0, 1, 2, 11, 12, 13, 14, 15, 8, 9, 10,
blake2b.c: ^
blake2b.c: blake2b.c:40:23: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:57:10: error: always_inline function '_mm256_set_epi64x' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_3_timingleaks' that is compiled without support for 'avx'
blake2b.c: ymm0 = _mm256_set_epi64x(0xA54FF53A5F1D36F1ULL, 0x3C6EF372FE94F82BULL, 0xBB67AE8584CAA73BULL, 0x6A09E667F3BCC908ULL ^ 0x0000000001010040ULL);
blake2b.c: ^
blake2b.c: blake2b.c:57:10: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:58:10: error: always_inline function '_mm256_set_epi64x' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_3_timingleaks' that is compiled without support for 'avx'
blake2b.c: ymm1 = _mm256_set_epi64x(0x5BE0CD19137E2179ULL, 0x1F83D9ABFB41BD6BULL, 0x9B05688C2B3E6C1FULL, 0x510E527FADE682D1ULL);
blake2b.c: ^
blake2b.c: blake2b.c:58:10: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:96:12: error: always_inline function '_mm256_set_epi64x' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_3_timingleaks' that is compiled without support for 'avx'
blake2b.c: ymm2 = _mm256_set_epi64x(0xA54FF53A5F1D36F1ULL, 0x3C6EF372FE94F82BULL, 0xBB67AE8584CAA73BULL, 0x6a09e667f3bcc908ULL);
blake2b.c: ^
blake2b.c: blake2b.c:96:12: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:99:14: error: always_inline function '_mm256_set_epi64x' requires target feature 'avx', but would be inlined into function 'crypto_hash_blake2b_avx2_3_timingleaks' that is compiled without support for 'avx'
blake2b.c: _mm256_set_epi64x(0x5BE0CD19137E2179ULL, 0x1F83D9ABFB41BD6BULL, 0x9B05688C2B3E6C1FULL, 0x510E527FADE682D1ULL),
blake2b.c: ^
blake2b.c: blake2b.c:99:14: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
blake2b.c: blake2b.c:101:14: error: '__builtin_ia32_pblendd256' needs target feature avx2
blake2b.c: ...

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2-3

Compiler output

Implementation: T:avx2-3
Security model: timingleaks
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

blake2b.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:53,
blake2b.c: from blake2b.c:12:
blake2b.c: blake2b.c: In function 'crypto_hash_blake2b_avx2_3_timingleaks':
blake2b.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
blake2b.c: 913 | _mm256_xor_si256 (__m256i __A, __m256i __B)
blake2b.c: | ^~~~~~~~~~~~~~~~
blake2b.c: blake2b.c:97:12: note: called from here
blake2b.c: 97 | ymm3 = _mm256_xor_si256
blake2b.c: | ^~~~~~~~~~~~~~~~
blake2b.c: 98 | (
blake2b.c: | ~
blake2b.c: 99 | _mm256_set_epi64x(0x5BE0CD19137E2179ULL, 0x1F83D9ABFB41BD6BULL, 0x9B05688C2B3E6C1FULL, 0x510E527FADE682D1ULL),
blake2b.c: | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
blake2b.c: 100 | #if defined(__x86_64__)
blake2b.c: | ~~~~~~~~~~~~~~~~~~~~~~~
blake2b.c: 101 | _mm256_blend_epi32(_mm256_castsi128_si256(_mm_cvtsi64_si128(ctr)),
blake2b.c: | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
blake2b.c: 102 | _mm256_setzero_si256(),
blake2b.c: | ~~~~~~~~~~~~~~~~~~~~~~~
blake2b.c: 103 | 0xF0
blake2b.c: | ~~~~
blake2b.c: 104 | )
blake2b.c: | ~
blake2b.c: 105 | #else
blake2b.c: | ~~~~~
blake2b.c: ...

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

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-3
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-3
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-3
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2-3

Compiler output

Implementation: T:avxicc
Security model: timingleaks
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

try.c: /usr/bin/ld: libcrypto_hash_blake2b.a(blake2b.o): warning: relocation in read-only section `.eh_frame'
try.c: /usr/bin/ld: warning: creating DT_TEXTREL in a PIE
try.c: /usr/bin/ld: libcrypto_hash_blake2b.a(blake2b.o): warning: relocation in read-only section `.eh_frame'
try.c: /usr/bin/ld: warning: creating DT_TEXTREL in a PIE
measure.c: /usr/bin/ld: libcrypto_hash_blake2b.a(blake2b.o): warning: relocation in read-only section `.eh_frame'
measure.c: /usr/bin/ld: warning: creating DT_TEXTREL in a PIE

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

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

Compiler output

Implementation: T:ymm
Security model: timingleaks
Compiler: clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

blake2b.c: blake2b.c:15:2: error: "This code requires at least AVX."
blake2b.c: #error "This code requires at least AVX."
blake2b.c: ^
blake2b.c: 1 error generated.

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:ymm