Implementation notes: x86, thoth, crypto_hash/keccakc256

Computer: thoth
Architecture: x86
CPU ID: AuthenticAMD-00000622-0183f9ff
SUPERCOP version: 20160806
Operation: crypto_hash
Primitive: keccakc256

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
48014	`mmxu1`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
49233	`mmxu1`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
51662	`mmxu1`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
52287	`mmxu1`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
63387	`opt32bi-s2lcu4`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
64655	`opt32bi-rvku2`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
64696	`opt32bi-s2lcu4`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
65673	`opt32biT-s2lcu4`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
65935	`simple32bi`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
66376	`inplace32bi`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
66470	`simple32bi`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
66648	`opt32biT-s2lcu4`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
67491	`opt32biT-s2lcu4`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
68843	`opt32bi-rvku2`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
69776	`opt32bi-s2lcu4`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
70129	`opt32bi-rvku2`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
70190	`simple32bi`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
70911	`opt32biT-s2lcu4`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
70918	`opt32bi-s2lcu4`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
71620	`simple32bi`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
72761	`inplace32bi`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
72790	`opt32bi-s2lcu4`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
73093	`inplace32bi`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
73509	`inplace32bi`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
74412	`opt32biT-s2lcu4`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
75382	`opt32bi-rvku2`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
77144	`opt32bi-rvku2`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
78616	`simple32bi`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
78775	`opt64lcu6`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
84195	`opt64u6`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
84797	`simple`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
88704	`inplace`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
102681	`opt64lcu24`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
109689	`opt64lcu24`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
109713	`opt64lcu24`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
112729	`inplace`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
114168	`opt64lcu24`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
114264	`opt64lcu6`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
116801	`simple`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
117189	`opt64u6`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
118516	`inplace`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
118942	`inplace`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
121239	`inplace`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
121290	`inplace32bi`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
122085	`simple`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
122327	`simple`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
122634	`opt64lcu6`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
123945	`opt64lcu6`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
124913	`opt64u6`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
125259	`opt64lcu6`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
126156	`opt64u6`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
127993	`simple`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
128886	`opt64u6`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
134049	`opt64lcu24`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
217032	`compact`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
218962	`compact`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
281723	`compact`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
291500	`compact`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
291872	`mmxu1`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
305463	`compact`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
418680	`compact8`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
422331	`compact8`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
503653	`compact8`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
554817	`compact8`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
566431	`compact8`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724

Checksum failure

Implementation: crypto_hash/keccakc256/opt64lcu24shld
Compiler: clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments

a7b68ec7b736c4175d875f741758fc0968702c68bff870d3658e91326981120f

Number of similar (compiler,implementation) pairs: 5, namely:

Compiler	Implementations
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments	opt64lcu24shld
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv	opt64lcu24shld
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv	opt64lcu24shld
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv	opt64lcu24shld
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv	opt64lcu24shld

Compiler output

Implementation: crypto_hash/keccakc256/sseu2
Compiler: clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments

KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:184:5: error: always_inline function '_mm_load_si128' requires target feature 'sse2', but would be inlined into function 'KeccakPermutationOnWords' that is compiled without support for 'sse2'
KeccakF-1600-opt64.c: copyFromState(A, state)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: ./KeccakF-1600-simd128.macros:547:19: note: expanded from macro 'copyFromState'
KeccakF-1600-opt64.c: X##bae.v128 = LOAD128(state[ 0]); \
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:51:33: note: expanded from macro 'LOAD128'
KeccakF-1600-opt64.c: #define LOAD128(a) _mm_load_si128((const V128 *)&(a))
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:184:5: error: always_inline function '_mm_unpackhi_epi64' requires target feature 'sse2', but would be inlined into function 'KeccakPermutationOnWords' that is compiled without support for 'sse2'
KeccakF-1600-opt64.c: ./KeccakF-1600-simd128.macros:549:13: note: expanded from macro 'copyFromState'
KeccakF-1600-opt64.c: X##be = GET64HIHI(X##bae.v128, X##bae.v128); \
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:58:33: note: expanded from macro 'GET64HIHI'
KeccakF-1600-opt64.c: #define GET64HIHI(a, b) _mm_unpackhi_epi64(a, b)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:184:5: error: always_inline function '_mm_load_si128' requires target feature 'sse2', but would be inlined into function 'KeccakPermutationOnWords' that is compiled without support for 'sse2'
KeccakF-1600-opt64.c: ./KeccakF-1600-simd128.macros:551:19: note: expanded from macro 'copyFromState'
KeccakF-1600-opt64.c: X##bio.v128 = LOAD128(state[ 2]); \
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:51:33: note: expanded from macro 'LOAD128'
KeccakF-1600-opt64.c: #define LOAD128(a) _mm_load_si128((const V128 *)&(a))
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:184:5: error: always_inline function '_mm_unpackhi_epi64' requires target feature 'sse2', but would be inlined into function 'KeccakPermutationOnWords' that is compiled without support for 'sse2'
KeccakF-1600-opt64.c: ./KeccakF-1600-simd128.macros:553:13: note: expanded from macro 'copyFromState'
KeccakF-1600-opt64.c: ...

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

Compiler	Implementations
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments	sseu2

Compiler output

Implementation: crypto_hash/keccakc256/xopu24
Compiler: clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments

KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:184:5: error: always_inline function '_mm_load_si128' requires target feature 'sse2', but would be inlined into function 'KeccakPermutationOnWords' that is compiled without support for 'sse2'
KeccakF-1600-opt64.c: copyFromState(A, state)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: ./KeccakF-1600-xop.macros:482:14: note: expanded from macro 'copyFromState'
KeccakF-1600-opt64.c: X##bae = LOAD128(state[ 0]); \
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:87:33: note: expanded from macro 'LOAD128'
KeccakF-1600-opt64.c: #define LOAD128(a) _mm_load_si128((const V128 *)&(a))
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:184:5: error: always_inline function '_mm_unpackhi_epi64' requires target feature 'sse2', but would be inlined into function 'KeccakPermutationOnWords' that is compiled without support for 'sse2'
KeccakF-1600-opt64.c: ./KeccakF-1600-xop.macros:484:13: note: expanded from macro 'copyFromState'
KeccakF-1600-opt64.c: X##be = GET64HIHI(X##bae, X##bae); \
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:96:33: note: expanded from macro 'GET64HIHI'
KeccakF-1600-opt64.c: #define GET64HIHI(a, b) _mm_unpackhi_epi64(a, b)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:184:5: error: always_inline function '_mm_load_si128' requires target feature 'sse2', but would be inlined into function 'KeccakPermutationOnWords' that is compiled without support for 'sse2'
KeccakF-1600-opt64.c: ./KeccakF-1600-xop.macros:486:14: note: expanded from macro 'copyFromState'
KeccakF-1600-opt64.c: X##bio = LOAD128(state[ 2]); \
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:87:33: note: expanded from macro 'LOAD128'
KeccakF-1600-opt64.c: #define LOAD128(a) _mm_load_si128((const V128 *)&(a))
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:184:5: error: always_inline function '_mm_unpackhi_epi64' requires target feature 'sse2', but would be inlined into function 'KeccakPermutationOnWords' that is compiled without support for 'sse2'
KeccakF-1600-opt64.c: ./KeccakF-1600-xop.macros:488:13: note: expanded from macro 'copyFromState'
KeccakF-1600-opt64.c: ...

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

Compiler	Implementations
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments	xopu24

Compiler output

Implementation: crypto_hash/keccakc256/xopu24
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv

KeccakF-1600-opt64.c: In file included from KeccakF-1600-opt64.c:130:0:
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c: In function 'KeccakPermutationOnWords':
KeccakF-1600-opt64.c: KeccakF-1600-xop.macros:482:12: warning: SSE vector return without SSE enabled changes the ABI [-Wpsabi]
KeccakF-1600-opt64.c: X##bae = LOAD128(state[ 0]); \
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:184:5: note: in expansion of macro 'copyFromState'
KeccakF-1600-opt64.c: copyFromState(A, state)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: In file included from /usr/lib/gcc/i686-linux-gnu/5/include/xmmintrin.h:1249:0,
KeccakF-1600-opt64.c: from /usr/lib/gcc/i686-linux-gnu/5/include/x86intrin.h:31,
KeccakF-1600-opt64.c: from KeccakF-1600-opt64.c:74:
KeccakF-1600-opt64.c: /usr/lib/gcc/i686-linux-gnu/5/include/emmintrin.h:720:1: error: inlining failed in call to always_inline '_mm_storel_epi64': target specific option mismatch
KeccakF-1600-opt64.c: _mm_storel_epi64 (__m128i *__P, __m128i __B)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:80:33: error: called from here
KeccakF-1600-opt64.c: #define STORE64(a, b) _mm_storel_epi64((V64 *)&(a), b)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-xop.macros:557:5: note: in expansion of macro 'STORE64'
KeccakF-1600-opt64.c: STORE64(state[24], X##su); \
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-unrolling.macros:41:5: note: in expansion of macro 'copyToState'
KeccakF-1600-opt64.c: copyToState(state, A)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:185:5: note: in expansion of macro 'rounds'
KeccakF-1600-opt64.c: rounds
KeccakF-1600-opt64.c: ...

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

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv	xopu24
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv	xopu24
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv	xopu24
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv	xopu24

Compiler output

Implementation: crypto_hash/keccakc256/sseu2
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv

KeccakF-1600-opt64.c: In file included from KeccakF-1600-opt64.c:67:0:
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c: In function 'KeccakPermutationOnWords':
KeccakF-1600-opt64.c: KeccakF-1600-simd128.macros:547:17: warning: SSE vector return without SSE enabled changes the ABI [-Wpsabi]
KeccakF-1600-opt64.c: X##bae.v128 = LOAD128(state[ 0]); \
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:184:5: note: in expansion of macro 'copyFromState'
KeccakF-1600-opt64.c: copyFromState(A, state)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: In file included from /usr/lib/gcc/i686-linux-gnu/5/include/xmmintrin.h:1249:0,
KeccakF-1600-opt64.c: from /usr/lib/gcc/i686-linux-gnu/5/include/x86intrin.h:31,
KeccakF-1600-opt64.c: from KeccakF-1600-opt64.c:31:
KeccakF-1600-opt64.c: /usr/lib/gcc/i686-linux-gnu/5/include/emmintrin.h:1286:1: error: inlining failed in call to always_inline '_mm_xor_si128': target specific option mismatch
KeccakF-1600-opt64.c: _mm_xor_si128 (__m128i __A, __m128i __B)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:45:35: error: called from here
KeccakF-1600-opt64.c: #define XOReq64(a, b) a = _mm_xor_si128(a, b)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-simd128.macros:600:5: note: in expansion of macro 'XOReq64'
KeccakF-1600-opt64.c: XOReq64(Cu, X##su); \
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: KeccakF-1600-opt64.c:184:5: note: in expansion of macro 'copyFromState'
KeccakF-1600-opt64.c: copyFromState(A, state)
KeccakF-1600-opt64.c: ^
KeccakF-1600-opt64.c: In file included from /usr/lib/gcc/i686-linux-gnu/5/include/xmmintrin.h:1249:0,
KeccakF-1600-opt64.c: from /usr/lib/gcc/i686-linux-gnu/5/include/x86intrin.h:31,
KeccakF-1600-opt64.c: ...

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

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv	sseu2
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv	sseu2
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv	sseu2
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv	sseu2