Implementation notes: amd64, skylake, crypto_sign/3icp

Computer: skylake
Architecture: amd64
CPU ID: GenuineIntel-000506e3-bfebfbff
SUPERCOP version: 20161026
Operation: crypto_sign
Primitive: 3icp

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
274180	`ref`	`g++ -m64 -O2 -fomit-frame-pointer`	20161217	20161026
286826	`ref`	`g++ -m64 -march=core-avx2 -O2 -fomit-frame-pointer`	20161217	20161026
310864	`ref`	`g++ -m64 -march=corei7-avx -O2 -fomit-frame-pointer`	20161217	20161026
319796	`ref`	`clang++ -O3 -fomit-frame-pointer -Qunused-arguments`	20161217	20161026
326248	`ref`	`g++ -m64 -march=corei7-avx -O3 -fomit-frame-pointer`	20161217	20161026
326480	`ref`	`g++ -m64 -march=corei7 -O3 -fomit-frame-pointer`	20161217	20161026
351560	`ref`	`g++ -m64 -march=core-avx-i -O2 -fomit-frame-pointer`	20161217	20161026
370172	`ref`	`g++ -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20161217	20161026
378312	`ref`	`g++ -m64 -march=core2 -msse4 -O2 -fomit-frame-pointer`	20161217	20161026
380606	`ref`	`g++ -m64 -march=native -mtune=native -O3 -fomit-frame-pointer`	20161217	20161026
382760	`ref`	`g++ -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20161217	20161026
392482	`ref`	`g++ -m64 -march=corei7 -O2 -fomit-frame-pointer`	20161217	20161026
406092	`ref`	`g++ -m64 -march=core-avx2 -O3 -fomit-frame-pointer`	20161217	20161026
408524	`ref`	`g++ -m64 -march=core2 -msse4 -O3 -fomit-frame-pointer`	20161217	20161026
409818	`ref`	`clang++ -mcpu=cortex-a8 -mfpu=neon -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20161217	20161026
410198	`ref`	`g++ -m64 -march=native -mtune=native -O2 -fomit-frame-pointer`	20161217	20161026
411156	`ref`	`g++ -m64 -march=core2 -O2 -fomit-frame-pointer`	20161217	20161026
411268	`ref`	`g++ -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20161217	20161026
413560	`ref`	`clang++ -mcpu=cortex-a9 -mfpu=neon -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20161217	20161026
415164	`ref`	`g++ -m64 -march=core2 -msse4 -O -fomit-frame-pointer`	20161217	20161026
418564	`ref`	`g++ -funroll-loops -fno-schedule-insns -O2 -fomit-frame-pointer`	20161217	20161026
420456	`ref`	`g++ -m64 -O3 -fomit-frame-pointer`	20161217	20161026
422664	`ref`	`g++ -fno-schedule-insns -O3 -fomit-frame-pointer`	20161217	20161026
424398	`ref`	`g++ -fno-schedule-insns -O -fomit-frame-pointer`	20161217	20161026
424940	`ref`	`g++ -m64 -march=core2 -O3 -fomit-frame-pointer`	20161217	20161026
428372	`ref`	`g++ -m64 -march=core-avx-i -O -fomit-frame-pointer`	20161217	20161026
428658	`ref`	`g++ -O3 -fomit-frame-pointer`	20161217	20161026
429166	`ref`	`g++ -m64 -march=core2 -msse4.1 -O3 -fomit-frame-pointer`	20161217	20161026
431760	`ref`	`clang++ -mcpu=native -mfpu=neon -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20161217	20161026
432458	`ref`	`g++ -m64 -march=barcelona -O2 -fomit-frame-pointer`	20161217	20161026
434644	`ref`	`g++ -march=barcelona -O2 -fomit-frame-pointer`	20161217	20161026
435202	`ref`	`g++ -m64 -march=corei7-avx -Os -fomit-frame-pointer`	20161217	20161026
438114	`ref`	`g++ -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20161217	20161026
440266	`ref`	`g++ -m64 -march=core2 -msse4.1 -O2 -fomit-frame-pointer`	20161217	20161026
442810	`ref`	`g++ -m64 -march=core2 -O -fomit-frame-pointer`	20161217	20161026
445890	`ref`	`g++ -march=nocona -O2 -fomit-frame-pointer`	20161217	20161026
448284	`ref`	`g++ -march=k8 -O -fomit-frame-pointer`	20161217	20161026
453404	`ref`	`g++ -march=nocona -O3 -fomit-frame-pointer`	20161217	20161026
455038	`ref`	`g++ -m64 -march=k8 -O3 -fomit-frame-pointer`	20161217	20161026
455416	`ref`	`g++ -m64 -march=corei7-avx -O -fomit-frame-pointer`	20161217	20161026
455748	`ref`	`g++ -m64 -march=native -mtune=native -Os -fomit-frame-pointer`	20161217	20161026
458202	`ref`	`g++ -funroll-loops -fno-schedule-insns -O -fomit-frame-pointer`	20161217	20161026
458460	`ref`	`g++ -m64 -march=nocona -O3 -fomit-frame-pointer`	20161217	20161026
463496	`ref`	`g++ -m64 -march=core-avx-i -O3 -fomit-frame-pointer`	20161217	20161026
463700	`ref`	`g++ -m64 -march=k8 -O -fomit-frame-pointer`	20161217	20161026
466392	`ref`	`g++ -O -fomit-frame-pointer`	20161217	20161026
474592	`ref`	`g++ -m64 -march=nocona -O -fomit-frame-pointer`	20161217	20161026
476230	`ref`	`g++ -m64 -march=core-avx2 -O -fomit-frame-pointer`	20161217	20161026
482964	`ref`	`g++ -m64 -march=core-avx2 -Os -fomit-frame-pointer`	20161217	20161026
496188	`ref`	`g++ -march=nocona -Os -fomit-frame-pointer`	20161217	20161026
496344	`ref`	`g++ -m64 -march=core2 -msse4 -Os -fomit-frame-pointer`	20161217	20161026
499242	`ref`	`g++ -m64 -march=nocona -O2 -fomit-frame-pointer`	20161217	20161026
501346	`ref`	`g++ -fno-schedule-insns -Os -fomit-frame-pointer`	20161217	20161026
503434	`ref`	`g++ -m64 -march=core-avx-i -Os -fomit-frame-pointer`	20161217	20161026
509038	`ref`	`g++ -m64 -march=core2 -msse4.1 -O -fomit-frame-pointer`	20161217	20161026
513860	`ref`	`g++ -funroll-loops -fno-schedule-insns -O3 -fomit-frame-pointer`	20161217	20161026
515890	`ref`	`g++ -m64 -march=k8 -O2 -fomit-frame-pointer`	20161217	20161026
516154	`ref`	`g++ -m64 -march=core2 -Os -fomit-frame-pointer`	20161217	20161026
523254	`ref`	`g++ -Os -fomit-frame-pointer`	20161217	20161026
524212	`ref`	`g++ -march=nocona -O -fomit-frame-pointer`	20161217	20161026
527276	`ref`	`g++ -march=barcelona -O3 -fomit-frame-pointer`	20161217	20161026
527842	`ref`	`g++ -m64 -march=barcelona -Os -fomit-frame-pointer`	20161217	20161026
537300	`ref`	`g++ -O2 -fomit-frame-pointer`	20161217	20161026
542088	`ref`	`g++ -march=k8 -O3 -fomit-frame-pointer`	20161217	20161026
548418	`ref`	`g++ -m64 -march=corei7 -Os -fomit-frame-pointer`	20161217	20161026
567238	`ref`	`g++ -march=k8 -Os -fomit-frame-pointer`	20161217	20161026
575040	`ref`	`g++ -fno-schedule-insns -O2 -fomit-frame-pointer`	20161217	20161026
575382	`ref`	`g++ -march=barcelona -O -fomit-frame-pointer`	20161217	20161026
580372	`ref`	`g++ -m64 -march=corei7 -O -fomit-frame-pointer`	20161217	20161026
588422	`ref`	`g++ -march=k8 -O2 -fomit-frame-pointer`	20161217	20161026
593820	`ref`	`g++ -funroll-loops -fno-schedule-insns -Os -fomit-frame-pointer`	20161217	20161026
593992	`ref`	`g++ -m64 -march=barcelona -O3 -fomit-frame-pointer`	20161217	20161026
609314	`ref`	`g++ -m64 -O -fomit-frame-pointer`	20161217	20161026
611276	`ref`	`g++ -march=barcelona -Os -fomit-frame-pointer`	20161217	20161026
612518	`ref`	`g++ -m64 -march=native -mtune=native -O -fomit-frame-pointer`	20161217	20161026
620818	`ref`	`g++ -m64 -march=nocona -Os -fomit-frame-pointer`	20161217	20161026
624136	`ref`	`g++ -m64 -march=k8 -Os -fomit-frame-pointer`	20161217	20161026
627502	`ref`	`g++ -m64 -march=core2 -msse4.1 -Os -fomit-frame-pointer`	20161217	20161026
645426	`ref`	`g++ -m64 -march=barcelona -O -fomit-frame-pointer`	20161217	20161026
661406	`ref`	`g++ -m64 -Os -fomit-frame-pointer`	20161217	20161026
6338086	`ref`	`g++`	20161217	20161026

Compiler output

Implementation: crypto_sign/3icp/ref
Compiler: clang++ -O3 -fwrapv -march=native -fomit-frame-pointer -Qunused-arguments

ThreeIC_p-test.cpp: In file included from ThreeIC_p-test.cpp:16:
ThreeIC_p-test.cpp: In file included from ./ThreeIC_p.h:6:
ThreeIC_p-test.cpp: ./tower.h:364:9: error: argument to '__builtin_ia32_palignr128' must be a constant integer
ThreeIC_p-test.cpp: return _mm_alignr_epi8(a1, a0, i*2);
ThreeIC_p-test.cpp: ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
ThreeIC_p-test.cpp: /usr/bin/../lib/clang/3.9.0/include/tmmintrin.h:161:12: note: expanded from macro '_mm_alignr_epi8'
ThreeIC_p-test.cpp: (__m128i)__builtin_ia32_palignr128((__v16qi)(__m128i)(a), \
ThreeIC_p-test.cpp: ^
ThreeIC_p-test.cpp: In file included from ThreeIC_p-test.cpp:16:
ThreeIC_p-test.cpp: In file included from ./ThreeIC_p.h:6:
ThreeIC_p-test.cpp: ./tower.h:364:9: error: cannot initialize return object of type '__m128i' (vector of 2 'long long' values) with an rvalue of type 'void'
ThreeIC_p-test.cpp: return _mm_alignr_epi8(a1, a0, i*2);
ThreeIC_p-test.cpp: ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
ThreeIC_p-test.cpp: /usr/bin/../lib/clang/3.9.0/include/tmmintrin.h:160:34: note: expanded from macro '_mm_alignr_epi8'
ThreeIC_p-test.cpp: #define _mm_alignr_epi8(a, b, n) __extension__ ({ \
ThreeIC_p-test.cpp: ^~~~~~~~~~~~~~~~~~
ThreeIC_p-test.cpp: 2 errors generated.

Number of similar (compiler,implementation) pairs: 6, namely:

Compiler	Implementations
clang++ -O3 -fwrapv -march=native -fomit-frame-pointer -Qunused-arguments	ref
clang++ -O3 -fwrapv -march=x86-64 -mcpu=core-avx2 -mavx2 -maes -mpclmul -fomit-frame-pointer -Qunused-arguments	ref
clang++ -O3 -fwrapv -mavx2 -fomit-frame-pointer -Qunused-arguments	ref
clang++ -O3 -fwrapv -mavx -fomit-frame-pointer -Qunused-arguments	ref
clang++ -O3 -fwrapv -mavx -maes -mpclmul -fomit-frame-pointer -Qunused-arguments	ref
clang++ -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments	ref