Implementation notes: amd64, par, crypto_aead/hs1sivv2

Computer: par
Architecture: amd64
CPU ID: GenuineIntel-000406c3-bfebfbff
SUPERCOP version: 20161026
Operation: crypto_aead
Primitive: hs1sivv2

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
40900	`dolbeau/amd64-sse`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
41400	`dolbeau/amd64-sse`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
42360	`dolbeau/amd64-sse`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
42420	`dolbeau/amd64-sse`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
43480	`dolbeau/amd64-sse`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
44180	`dolbeau/amd64-sse`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
45960	`faster`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
46460	`faster`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
46740	`faster`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
46900	`faster`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
46920	`faster`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
47100	`faster`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
64240	`ref`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
67380	`ref`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
69100	`ref`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
89780	`ref`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
104380	`ref`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
105580	`ref`	`gcc -march=native -mcpu=native -Os`	20161214	20161026

Compiler output

Implementation: crypto_aead/hs1sivv2/dolbeau/amd64-avx2
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

encrypt.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
encrypt.c: encrypt.c:90:2: error: #error "This code requires AVX2 to work"
encrypt.c: #error "This code requires AVX2 to work"
encrypt.c: ^~~~~
encrypt.c: encrypt.c: In function 'prf_hash2_2':
encrypt.c: encrypt.c:425:19: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
encrypt.c: __m256i kv0 = _mm256_loadu_si256((const __m256i*)(nhkey+ 0)); // 1
encrypt.c: ^~~

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

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	dolbeau/amd64-avx2
gcc -funroll-loops -march=native -mcpu=native -O3	dolbeau/amd64-avx2
gcc -funroll-loops -march=native -mcpu=native -Os	dolbeau/amd64-avx2
gcc -march=native -mcpu=native -O2	dolbeau/amd64-avx2
gcc -march=native -mcpu=native -O3	dolbeau/amd64-avx2
gcc -march=native -mcpu=native -Os	dolbeau/amd64-avx2

Compiler output

Implementation: crypto_aead/hs1sivv2/dolbeau/amd64-avx512
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

encrypt.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
encrypt.c: encrypt.c:90:2: error: #error "This code requires AVX512F to work"
encrypt.c: #error "This code requires AVX512F to work"
encrypt.c: ^~~~~
encrypt.c: encrypt.c: In function '_mm512_reduce_add_epi64':
encrypt.c: encrypt.c:321:20: note: The ABI for passing parameters with 64-byte alignment has changed in GCC 4.6
encrypt.c: unsigned long long _mm512_reduce_add_epi64 (__m512i a) {
encrypt.c: ^~~~~~~~~~~~~~~~~~~~~~~
encrypt.c: encrypt.c: In function 'prf_hash2_2':
encrypt.c: encrypt.c:482:19: warning: AVX512F vector return without AVX512F enabled changes the ABI [-Wpsabi]
encrypt.c: __m512i kv0 = _mm512_loadu_si512((const __m512i*)(nhkey+ 0)); // 1
encrypt.c: ^~~
encrypt.c: encrypt.c:501:19: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
encrypt.c: __m256i inv0 = _mm256_inserti128_si256(_mm256_castsi128_si256(inv0lo), inv0lo, 1);
encrypt.c: ^~~~

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

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	dolbeau/amd64-avx512
gcc -funroll-loops -march=native -mcpu=native -O3	dolbeau/amd64-avx512
gcc -funroll-loops -march=native -mcpu=native -Os	dolbeau/amd64-avx512
gcc -march=native -mcpu=native -O2	dolbeau/amd64-avx512
gcc -march=native -mcpu=native -O3	dolbeau/amd64-avx512
gcc -march=native -mcpu=native -Os	dolbeau/amd64-avx512

Compiler output

Implementation: crypto_aead/hs1sivv2/dolbeau/amd64-sse
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

encrypt.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: 12, namely:

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	dolbeau/amd64-sse ref
gcc -funroll-loops -march=native -mcpu=native -O3	dolbeau/amd64-sse ref
gcc -funroll-loops -march=native -mcpu=native -Os	dolbeau/amd64-sse ref
gcc -march=native -mcpu=native -O2	dolbeau/amd64-sse ref
gcc -march=native -mcpu=native -O3	dolbeau/amd64-sse ref
gcc -march=native -mcpu=native -Os	dolbeau/amd64-sse ref

Compiler output

Implementation: crypto_aead/hs1sivv2/faster
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

hs1.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
chacha_moon.S: 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: 6, namely:

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