Implementation notes: x86, katana, crypto_sign/mqqsig192

Computer: katana
Architecture: x86
CPU ID: GenuineIntel-000006f6-bfebfbff
SUPERCOP version: 20170105
Operation: crypto_sign
Primitive: mqqsig192

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
270592	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=core2_-O3_-fomit-frame-pointer`	20161030	20161009
271840	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=native_-mtune=native_-O3_-fomit-frame-pointer`	20161030	20161009
273272	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=core2_-O2_-fomit-frame-pointer`	20161030	20161009
273328	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=native_-mtune=native_-O2_-fomit-frame-pointer`	20161030	20161009
298400	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=prescott_-O3_-fomit-frame-pointer`	20161030	20161009
298760	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=nocona_-O3_-fomit-frame-pointer`	20161030	20161009
299256	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium-m_-O3_-fomit-frame-pointer`	20161030	20161009
307352	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=core2_-O_-fomit-frame-pointer`	20161030	20161009
307784	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=native_-mtune=native_-O_-fomit-frame-pointer`	20161030	20161009
313544	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=prescott_-O3_-fomit-frame-pointer`	20161030	20161009
315016	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=nocona_-O3_-fomit-frame-pointer`	20161030	20161009
360600	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium4_-O3_-fomit-frame-pointer`	20161030	20161009
361352	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium4_-O3_-fomit-frame-pointer`	20161030	20161009
363432	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=native_-mtune=native_-Os_-fomit-frame-pointer`	20161030	20161009
363544	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=core2_-Os_-fomit-frame-pointer`	20161030	20161009
364624	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium-m_-O3_-fomit-frame-pointer`	20161030	20161009
428840	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-O3_-fomit-frame-pointer`	20161030	20161009
431328	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=athlon_-O3_-fomit-frame-pointer`	20161030	20161009
453744	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium2_-O3_-fomit-frame-pointer`	20161030	20161009
454224	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentiumpro_-O3_-fomit-frame-pointer`	20161030	20161009
454576	? ? ?	? ? ?	`ref`	`gcc_-m32_-O3_-fomit-frame-pointer`	20161030	20161009
463856	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium-mmx_-O3_-fomit-frame-pointer`	20161030	20161009
465296	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium_-O3_-fomit-frame-pointer`	20161030	20161009
473472	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium2_-O3_-fomit-frame-pointer`	20161030	20161009
473592	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentiumpro_-O3_-fomit-frame-pointer`	20161030	20161009
480712	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=athlon_-O_-fomit-frame-pointer`	20161030	20161009
484864	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-O2_-fomit-frame-pointer`	20161030	20161009
486920	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=athlon_-O2_-fomit-frame-pointer`	20161030	20161009
487544	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6-2_-O3_-fomit-frame-pointer`	20161030	20161009
487640	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6-3_-O3_-fomit-frame-pointer`	20161030	20161009
488176	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=athlon_-O3_-fomit-frame-pointer`	20161030	20161009
490432	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6_-O3_-fomit-frame-pointer`	20161030	20161009
495816	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-O_-fomit-frame-pointer`	20161030	20161009
496840	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=nocona_-O2_-fomit-frame-pointer`	20161030	20161009
497648	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=nocona_-O_-fomit-frame-pointer`	20161030	20161009
499176	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium-m_-O_-fomit-frame-pointer`	20161030	20161009
499376	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=i386_-O3_-fomit-frame-pointer`	20161030	20161009
499600	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=prescott_-O_-fomit-frame-pointer`	20161030	20161009
499728	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentiumpro_-O_-fomit-frame-pointer`	20161030	20161009
499744	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium2_-O_-fomit-frame-pointer`	20161030	20161009
501352	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium3_-O_-fomit-frame-pointer`	20161030	20161009
503576	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium-m_-O2_-fomit-frame-pointer`	20161030	20161009
504352	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium3_-O3_-fomit-frame-pointer`	20161030	20161009
504472	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium2_-O2_-fomit-frame-pointer`	20161030	20161009
505032	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentiumpro_-O2_-fomit-frame-pointer`	20161030	20161009
505568	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=prescott_-O2_-fomit-frame-pointer`	20161030	20161009
506080	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6-3_-O3_-fomit-frame-pointer`	20161030	20161009
506528	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6-2_-O3_-fomit-frame-pointer`	20161030	20161009
506776	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium3_-O2_-fomit-frame-pointer`	20161030	20161009
507128	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6_-O3_-fomit-frame-pointer`	20161030	20161009
513656	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6-2_-O2_-fomit-frame-pointer`	20161030	20161009
514400	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=i386_-O3_-fomit-frame-pointer`	20161030	20161009
514400	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6_-O2_-fomit-frame-pointer`	20161030	20161009
514760	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6-3_-O2_-fomit-frame-pointer`	20161030	20161009
518000	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium_-O_-fomit-frame-pointer`	20161030	20161009
519592	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=i386_-O_-fomit-frame-pointer`	20161030	20161009
520472	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium-mmx_-O_-fomit-frame-pointer`	20161030	20161009
520744	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium_-O3_-fomit-frame-pointer`	20161030	20161009
521240	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium-mmx_-O3_-fomit-frame-pointer`	20161030	20161009
532888	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=i386_-O2_-fomit-frame-pointer`	20161030	20161009
534104	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6-3_-O_-fomit-frame-pointer`	20161030	20161009
534328	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=athlon_-O2_-fomit-frame-pointer`	20161030	20161009
536288	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6_-O_-fomit-frame-pointer`	20161030	20161009
536512	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6-2_-O_-fomit-frame-pointer`	20161030	20161009
537832	? ? ?	? ? ?	`ref`	`gcc_-m32_-O2_-fomit-frame-pointer`	20161030	20161009
538120	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium4_-O_-fomit-frame-pointer`	20161030	20161009
566408	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium4_-O2_-fomit-frame-pointer`	20161030	20161009
569824	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium3_-O2_-fomit-frame-pointer`	20161030	20161009
590192	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium-m_-O2_-fomit-frame-pointer`	20161030	20161009
590376	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium2_-O2_-fomit-frame-pointer`	20161030	20161009
590696	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium3_-O3_-fomit-frame-pointer`	20161030	20161009
591496	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentiumpro_-O2_-fomit-frame-pointer`	20161030	20161009
593600	? ? ?	? ? ?	`ref`	`gcc_-m32_-O_-fomit-frame-pointer`	20161030	20161009
596784	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6-2_-O2_-fomit-frame-pointer`	20161030	20161009
596928	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6-3_-O2_-fomit-frame-pointer`	20161030	20161009
597144	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium2_-O_-fomit-frame-pointer`	20161030	20161009
599784	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6_-O2_-fomit-frame-pointer`	20161030	20161009
602056	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentiumpro_-O_-fomit-frame-pointer`	20161030	20161009
605960	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=i386_-O_-fomit-frame-pointer`	20161030	20161009
609016	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6-2_-O_-fomit-frame-pointer`	20161030	20161009
609160	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6-3_-O_-fomit-frame-pointer`	20161030	20161009
609344	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6_-O_-fomit-frame-pointer`	20161030	20161009
609456	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium4_-O2_-fomit-frame-pointer`	20161030	20161009
614384	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium-mmx_-O_-fomit-frame-pointer`	20161030	20161009
615536	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium_-O_-fomit-frame-pointer`	20161030	20161009
627168	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium-mmx_-O2_-fomit-frame-pointer`	20161030	20161009
627520	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium_-O2_-fomit-frame-pointer`	20161030	20161009
630224	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=i386_-O2_-fomit-frame-pointer`	20161030	20161009
635960	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=prescott_-O2_-fomit-frame-pointer`	20161030	20161009
645272	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=nocona_-O2_-fomit-frame-pointer`	20161030	20161009
647048	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium-mmx_-O2_-fomit-frame-pointer`	20161030	20161009
648056	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium_-O2_-fomit-frame-pointer`	20161030	20161009
654744	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium3_-Os_-fomit-frame-pointer`	20161030	20161009
656232	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium4_-O_-fomit-frame-pointer`	20161030	20161009
658088	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=prescott_-O_-fomit-frame-pointer`	20161030	20161009
658568	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=nocona_-O_-fomit-frame-pointer`	20161030	20161009
658816	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=athlon_-O_-fomit-frame-pointer`	20161030	20161009
662904	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium-m_-O_-fomit-frame-pointer`	20161030	20161009
663352	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium3_-O_-fomit-frame-pointer`	20161030	20161009
668672	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k8_-Os_-fomit-frame-pointer`	20161030	20161009
669576	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=i486_-Os_-fomit-frame-pointer`	20161030	20161009
669800	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=barcelona_-Os_-fomit-frame-pointer`	20161030	20161009
671192	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium2_-Os_-fomit-frame-pointer`	20161030	20161009
671752	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium-m_-Os_-fomit-frame-pointer`	20161030	20161009
671840	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=prescott_-Os_-fomit-frame-pointer`	20161030	20161009
672640	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium4_-Os_-fomit-frame-pointer`	20161030	20161009
673112	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentiumpro_-Os_-fomit-frame-pointer`	20161030	20161009
673800	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=nocona_-Os_-fomit-frame-pointer`	20161030	20161009
673904	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-Os_-fomit-frame-pointer`	20161030	20161009
674160	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=athlon_-Os_-fomit-frame-pointer`	20161030	20161009
677680	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium_-Os_-fomit-frame-pointer`	20161030	20161009
678240	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium-mmx_-Os_-fomit-frame-pointer`	20161030	20161009
679128	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=i486_-Os_-fomit-frame-pointer`	20161030	20161009
680952	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=i386_-Os_-fomit-frame-pointer`	20161030	20161009
684096	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6-2_-Os_-fomit-frame-pointer`	20161030	20161009
684512	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=i386_-Os_-fomit-frame-pointer`	20161030	20161009
684600	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6_-Os_-fomit-frame-pointer`	20161030	20161009
685200	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6-3_-Os_-fomit-frame-pointer`	20161030	20161009
685664	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k6-3_-Os_-fomit-frame-pointer`	20161030	20161009
685848	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium_-Os_-fomit-frame-pointer`	20161030	20161009
685936	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6-2_-Os_-fomit-frame-pointer`	20161030	20161009
686656	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=k6_-Os_-fomit-frame-pointer`	20161030	20161009
687240	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium-mmx_-Os_-fomit-frame-pointer`	20161030	20161009
691928	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentiumpro_-Os_-fomit-frame-pointer`	20161030	20161009
693864	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium2_-Os_-fomit-frame-pointer`	20161030	20161009
699272	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium3_-Os_-fomit-frame-pointer`	20161030	20161009
700864	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=pentium-m_-Os_-fomit-frame-pointer`	20161030	20161009
738616	? ? ?	? ? ?	`ref`	`gcc_-m32_-Os_-fomit-frame-pointer`	20161030	20161009
739072	? ? ?	? ? ?	`ref`	`gcc_-m32_-march=athlon_-Os_-fomit-frame-pointer`	20161030	20161009
746464	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=pentium4_-Os_-fomit-frame-pointer`	20161030	20161009
746840	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=prescott_-Os_-fomit-frame-pointer`	20161030	20161009
746912	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=nocona_-Os_-fomit-frame-pointer`	20161030	20161009
746936	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=barcelona_-Os_-fomit-frame-pointer`	20161030	20161009
747008	? ? ?	? ? ?	`ref`	`gcc_-funroll-loops_-m32_-march=k8_-Os_-fomit-frame-pointer`	20161030	20161009

Test failure

Implementation: crypto_sign/mqqsig192/ref
Compiler: gcc -funroll-loops -m32 -march=barcelona -O2 -fomit-frame-pointer

error 111
crypto_sign_open does not handle c=t overlap

Number of similar (compiler,implementation) pairs: 15, namely:

Compiler	Implementations
gcc -funroll-loops -m32 -march=barcelona -O2 -fomit-frame-pointer	ref
gcc -funroll-loops -m32 -march=barcelona -O3 -fomit-frame-pointer	ref
gcc -funroll-loops -m32 -march=barcelona -O -fomit-frame-pointer	ref
gcc -funroll-loops -m32 -march=i486 -O2 -fomit-frame-pointer	ref
gcc -funroll-loops -m32 -march=i486 -O3 -fomit-frame-pointer	ref
gcc -funroll-loops -m32 -march=i486 -O -fomit-frame-pointer	ref
gcc -funroll-loops -m32 -march=k8 -O2 -fomit-frame-pointer	ref
gcc -funroll-loops -m32 -march=k8 -O3 -fomit-frame-pointer	ref
gcc -funroll-loops -m32 -march=k8 -O -fomit-frame-pointer	ref
gcc -m32 -march=i486 -O2 -fomit-frame-pointer	ref
gcc -m32 -march=i486 -O3 -fomit-frame-pointer	ref
gcc -m32 -march=i486 -O -fomit-frame-pointer	ref
gcc -m32 -march=k8 -O2 -fomit-frame-pointer	ref
gcc -m32 -march=k8 -O3 -fomit-frame-pointer	ref
gcc -m32 -march=k8 -O -fomit-frame-pointer	ref

Test failure

Implementation: crypto_sign/mqqsig192/ref
Compiler: gcc -m32 -march=barcelona -O2 -fomit-frame-pointer

error 111
crypto_sign_open does not handle c=t overlap
error 111
crypto_sign_open does not handle c=t overlap

Number of similar (compiler,implementation) pairs: 3, namely:

Compiler	Implementations
gcc -m32 -march=barcelona -O2 -fomit-frame-pointer	ref
gcc -m32 -march=barcelona -O3 -fomit-frame-pointer	ref
gcc -m32 -march=barcelona -O -fomit-frame-pointer	ref

Test failure

Implementation: crypto_sign/mqqsig192/ref
Compiler: gcc -m32 -march=core-avx-i -O2 -fomit-frame-pointer

error 111

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

Compiler	Implementations
gcc -m32 -march=core-avx-i -O2 -fomit-frame-pointer	ref
gcc -m32 -march=core-avx-i -O -fomit-frame-pointer	ref
gcc -m32 -march=core-avx-i -Os -fomit-frame-pointer	ref
gcc -m32 -march=core-avx2 -O2 -fomit-frame-pointer	ref
gcc -m32 -march=core-avx2 -O -fomit-frame-pointer	ref
gcc -m32 -march=core-avx2 -Os -fomit-frame-pointer	ref
gcc -m32 -march=core2 -msse4.1 -O2 -fomit-frame-pointer	ref
gcc -m32 -march=core2 -msse4.1 -O -fomit-frame-pointer	ref
gcc -m32 -march=core2 -msse4.1 -Os -fomit-frame-pointer	ref
gcc -m32 -march=core2 -msse4 -O2 -fomit-frame-pointer	ref
gcc -m32 -march=core2 -msse4 -O -fomit-frame-pointer	ref
gcc -m32 -march=core2 -msse4 -Os -fomit-frame-pointer	ref
gcc -m32 -march=corei7-avx -O2 -fomit-frame-pointer	ref
gcc -m32 -march=corei7-avx -O -fomit-frame-pointer	ref
gcc -m32 -march=corei7-avx -Os -fomit-frame-pointer	ref
gcc -m32 -march=corei7 -O2 -fomit-frame-pointer	ref
gcc -m32 -march=corei7 -O -fomit-frame-pointer	ref
gcc -m32 -march=corei7 -Os -fomit-frame-pointer	ref