Implementation notes: amd64, par, crypto_hash/groestl512

Computer: par
Architecture: amd64
CPU ID: GenuineIntel-000406c3-bfebfbff
SUPERCOP version: 20161026
Operation: crypto_hash
Primitive: groestl512

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
82360	`aesni`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
82520	`aesni`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
82640	`aesni`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
83220	`aesni`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
84320	`aesni`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
84320	`aesni`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
87540	`aesni-intr`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
87780	`aesni-intr`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
87780	`aesni-intr`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
90100	`aesni-intr`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
92000	`opteron`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
92540	`opteron`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
92820	`opteron`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
93700	`opteron`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
94920	`aesni-intr`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
95900	`opteron`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
95980	`aesni-intr`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
96440	`opteron`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
122620	`sphlib`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
123220	`sphlib-adapted`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
124440	`sphlib`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
124880	`sphlib-adapted`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
126140	`sphlib`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
126780	`sphlib`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
129800	`sphlib-adapted`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
130860	`opt64`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
132820	`sphlib`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
134620	`core2duo`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
134640	`core2duo`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
134820	`core2duo`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
136180	`core2duo`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
136780	`sphlib-adapted`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
136840	`core2duo`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
137140	`core2duo`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
139260	`opt64`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
139440	`sphlib-adapted`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
139620	`opt64`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
140280	`opt64`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
143200	`opt64`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
160560	`sphlib-adapted`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
164440	`sphlib`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
166860	`opt64`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
193580	`sphlib-small`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
200040	`sphlib-small`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
200600	`sphlib-small`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
202680	`sphlib-small`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
221840	`sphlib-small`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
228400	`sphlib-small`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
230520	`opt32`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
237460	`opt32`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
237680	`opt32`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
291600	`mmx`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
292100	`mmx`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
301120	`mmx`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
302180	`mmx`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
304480	`mmx`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
312000	`mmx`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
332100	`vperm-intr`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
337380	`vperm`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
337400	`vperm`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
337980	`vperm`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
338460	`vperm`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
339340	`vperm`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
339540	`vperm`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
342540	`vperm-intr`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
342820	`vperm-intr`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
355560	`vperm-intr`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
356800	`vperm-intr`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
372660	`vperm-intr`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
396180	`opt32`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
400120	`opt32`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
419880	`opt32`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
539680	`32bit-bytesliced-c-small`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
542300	`32bit-bytesliced-c-small`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
593380	`32bit-bytesliced-c-small`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
1054940	`32bit-bytesliced-c-small`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
1145040	`32bit-bytesliced-c-small`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
1156140	`32bit-bytesliced-c-small`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026

Test failure

Implementation: crypto_hash/groestl512/avx
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

error 111

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

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	avx
gcc -funroll-loops -march=native -mcpu=native -O3	avx
gcc -funroll-loops -march=native -mcpu=native -Os	avx
gcc -march=native -mcpu=native -O2	avx
gcc -march=native -mcpu=native -O3	avx
gcc -march=native -mcpu=native -Os	avx

Compiler output

Implementation: crypto_hash/groestl512/sphlib
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

groestl.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
hash.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
try.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
try.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
measure.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead

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

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	sphlib sphlib-adapted sphlib-small
gcc -funroll-loops -march=native -mcpu=native -O3	sphlib sphlib-adapted sphlib-small
gcc -funroll-loops -march=native -mcpu=native -Os	sphlib sphlib-adapted sphlib-small
gcc -march=native -mcpu=native -O2	sphlib sphlib-adapted sphlib-small
gcc -march=native -mcpu=native -O3	sphlib sphlib-adapted sphlib-small
gcc -march=native -mcpu=native -Os	sphlib sphlib-adapted sphlib-small

Compiler output

Implementation: crypto_hash/groestl512/avx-intr
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

hash.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
hash.c: In file included from hash.c:31:0:
hash.c: groestl-intr-avx.h: In function 'TF1024':
hash.c: groestl-intr-avx.h:906:8: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
hash.c: ymm8 = insert_m128i_in_m256d(ymm8, xmm8, 0);
hash.c: ^
hash.c: In file included from /usr/lib/gcc/x86_64-pc-linux-gnu/6.2.1/include/immintrin.h:41:0,
hash.c: from groestl-intr-avx.h:12,
hash.c: from hash.c:31:
hash.c: /usr/lib/gcc/x86_64-pc-linux-gnu/6.2.1/include/avxintrin.h:1416:1: error: inlining failed in call to always_inline '_mm256_castsi256_pd': target specific option mismatch
hash.c: _mm256_castsi256_pd (__m256i __A)
hash.c: ^~~~~~~~~~~~~~~~~~~
hash.c: In file included from hash.c:31:0:
hash.c: groestl-intr-avx.h:33:47: note: called from here
hash.c: #define insert_m128i_in_m256d(ymm, xmm, pos) (_mm256_castsi256_pd(_mm256_insertf128_si256(_mm256_castpd_si256(ymm), xmm, pos)))
hash.c: ~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
hash.c: groestl-intr-avx.h:922:11: note: in expansion of macro 'insert_m128i_in_m256d'
hash.c: ymm15 = insert_m128i_in_m256d(ymm15, xmm7, 1);
hash.c: ^~~~~~~~~~~~~~~~~~~~~
hash.c: In file included from /usr/lib/gcc/x86_64-pc-linux-gnu/6.2.1/include/immintrin.h:41:0,
hash.c: from groestl-intr-avx.h:12,
hash.c: from hash.c:31:
hash.c: /usr/lib/gcc/x86_64-pc-linux-gnu/6.2.1/include/avxintrin.h:742:1: error: inlining failed in call to always_inline '_mm256_insertf128_si256': target specific option mismatch
hash.c: _mm256_insertf128_si256 (__m256i __X, __m128i __Y, const int __O)
hash.c: ^~~~~~~~~~~~~~~~~~~~~~~
hash.c: ...

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

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	avx-intr
gcc -funroll-loops -march=native -mcpu=native -O3	avx-intr
gcc -funroll-loops -march=native -mcpu=native -Os	avx-intr
gcc -march=native -mcpu=native -O2	avx-intr
gcc -march=native -mcpu=native -O3	avx-intr
gcc -march=native -mcpu=native -Os	avx-intr

Compiler output

Implementation: crypto_hash/groestl512/opt64
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

hash.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
hash.c: hash.c:194:14: warning: 'inP' is static but declared in inline function 'F1024' which is not static
hash.c: static u64 inP[COLS1024] __attribute__((aligned(16)));
hash.c: ^~~
hash.c: hash.c:193:14: warning: 'outQ' is static but declared in inline function 'F1024' which is not static
hash.c: static u64 outQ[COLS1024] __attribute__((aligned(16)));
hash.c: ^~~~
hash.c: hash.c:192:14: warning: 'z' is static but declared in inline function 'F1024' which is not static
hash.c: static u64 z[COLS1024] __attribute__((aligned(16)));
hash.c: ^
hash.c: hash.c:191:14: warning: 'y' is static but declared in inline function 'F1024' which is not static
hash.c: static u64 y[COLS1024] __attribute__((aligned(16)));
hash.c: ^
try.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
try.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
measure.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead

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

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	opt64
gcc -funroll-loops -march=native -mcpu=native -O3	opt64
gcc -funroll-loops -march=native -mcpu=native -Os	opt64
gcc -march=native -mcpu=native -O2	opt64
gcc -march=native -mcpu=native -O3	opt64
gcc -march=native -mcpu=native -Os	opt64

Compiler output

Implementation: crypto_hash/groestl512/avx
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

hash.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
try.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead

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

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	avx
gcc -funroll-loops -march=native -mcpu=native -O3	avx
gcc -funroll-loops -march=native -mcpu=native -Os	avx
gcc -march=native -mcpu=native -O2	avx
gcc -march=native -mcpu=native -O3	avx
gcc -march=native -mcpu=native -Os	avx

Compiler output

Implementation: crypto_hash/groestl512/32bit-bytesliced-c-small
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

hash.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
try.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
try.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
measure.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead

Number of similar (compiler,implementation) pairs: 54, namely:

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	32bit-bytesliced-c-small aesni aesni-intr core2duo mmx opt32 opteron vperm vperm-intr
gcc -funroll-loops -march=native -mcpu=native -O3	32bit-bytesliced-c-small aesni aesni-intr core2duo mmx opt32 opteron vperm vperm-intr
gcc -funroll-loops -march=native -mcpu=native -Os	32bit-bytesliced-c-small aesni aesni-intr core2duo mmx opt32 opteron vperm vperm-intr
gcc -march=native -mcpu=native -O2	32bit-bytesliced-c-small aesni aesni-intr core2duo mmx opt32 opteron vperm vperm-intr
gcc -march=native -mcpu=native -O3	32bit-bytesliced-c-small aesni aesni-intr core2duo mmx opt32 opteron vperm vperm-intr
gcc -march=native -mcpu=native -Os	32bit-bytesliced-c-small aesni aesni-intr core2duo mmx opt32 opteron vperm vperm-intr