Implementation notes: amd64, par, crypto_stream/chacha8

Computer: par
Architecture: amd64
CPU ID: GenuineIntel-000406c3-bfebfbff
SUPERCOP version: 20161026
Operation: crypto_stream
Primitive: chacha8

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
3800	`moon/sse2/64`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
3800	`moon/sse2/64`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
3800	`moon/sse2/64`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
3820	`moon/sse2/64`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
3820	`moon/sse2/64`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
3820	`moon/sse2/64`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
4940	`krovetz/vec128`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
4980	`krovetz/vec128`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
5220	`krovetz/vec128`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
5220	`krovetz/vec128`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
5240	`krovetz/vec128`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
5240	`krovetz/vec128`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
5400	`dolbeau/amd64-avx2`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
5540	`dolbeau/amd64-avx2`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
5540	`e/amd64-xmm6`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
5540	`e/amd64-xmm6`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
5540	`dolbeau/amd64-avx2`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
5560	`e/amd64-xmm6`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
5560	`e/amd64-xmm6`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
5560	`e/amd64-xmm6`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
5560	`e/amd64-xmm6`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
5960	`moon/ssse3/64`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
5960	`moon/ssse3/64`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
5960	`moon/ssse3/64`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
5960	`moon/ssse3/64`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
5960	`moon/ssse3/64`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
5960	`moon/ssse3/64`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
6100	`dolbeau/amd64-avx2`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
6120	`dolbeau/amd64-avx2`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
6200	`amd64-ssse3`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
6200	`dolbeau/amd64-avx2`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
6200	`amd64-ssse3`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
6200	`amd64-ssse3`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
6200	`amd64-ssse3`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
6200	`amd64-ssse3`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
6200	`amd64-ssse3`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
8400	`e/amd64-3`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
8400	`e/amd64-3`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
8400	`e/amd64-3`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
8400	`e/amd64-3`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
8420	`e/amd64-3`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
8420	`e/amd64-3`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
10160	`e/merged`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
10220	`e/ref`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
10240	`e/merged`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
10240	`e/merged`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
10240	`e/ref`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
10240	`e/regs`	`gcc -march=native -mcpu=native -O3`	20161214	20161026
10340	`e/regs`	`gcc -funroll-loops -march=native -mcpu=native -O3`	20161214	20161026
10500	`e/merged`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
10700	`e/merged`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
11020	`e/merged`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
14120	`e/regs`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
14320	`e/regs`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
14540	`e/ref`	`gcc -funroll-loops -march=native -mcpu=native -O2`	20161214	20161026
18660	`e/ref`	`gcc -march=native -mcpu=native -O2`	20161214	20161026
19400	`e/regs`	`gcc -march=native -mcpu=native -Os`	20161214	20161026
19540	`e/regs`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
23520	`e/ref`	`gcc -funroll-loops -march=native -mcpu=native -Os`	20161214	20161026
28080	`e/ref`	`gcc -march=native -mcpu=native -Os`	20161214	20161026

Test failure

Implementation: crypto_stream/chacha8/moon/avx/64
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

error 111

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

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	moon/avx/64 moon/avx2/64 moon/xop/64
gcc -funroll-loops -march=native -mcpu=native -O3	moon/avx/64 moon/avx2/64 moon/xop/64
gcc -funroll-loops -march=native -mcpu=native -Os	moon/avx/64 moon/avx2/64 moon/xop/64
gcc -march=native -mcpu=native -O2	moon/avx/64 moon/avx2/64 moon/xop/64
gcc -march=native -mcpu=native -O3	moon/avx/64 moon/avx2/64 moon/xop/64
gcc -march=native -mcpu=native -Os	moon/avx/64 moon/avx2/64 moon/xop/64

Compiler output

Implementation: crypto_stream/chacha8/dolbeau/ppc-altivec
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

api.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
chacha.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
chacha.c: chacha.c:11:21: fatal error: altivec.h: No such file or directory
chacha.c: #include gt;
chacha.c: ^
chacha.c: compilation terminated.

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

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	dolbeau/ppc-altivec
gcc -funroll-loops -march=native -mcpu=native -O3	dolbeau/ppc-altivec
gcc -funroll-loops -march=native -mcpu=native -Os	dolbeau/ppc-altivec
gcc -march=native -mcpu=native -O2	dolbeau/ppc-altivec
gcc -march=native -mcpu=native -O3	dolbeau/ppc-altivec
gcc -march=native -mcpu=native -Os	dolbeau/ppc-altivec

Compiler output

Implementation: crypto_stream/chacha8/dolbeau/mipsel-msa
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

api.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
chacha.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
chacha.c: chacha.c:11:22: fatal error: arm_neon.h: No such file or directory
chacha.c: #include gt;
chacha.c: ^
chacha.c: compilation terminated.

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

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	dolbeau/mipsel-msa
gcc -funroll-loops -march=native -mcpu=native -O3	dolbeau/mipsel-msa
gcc -funroll-loops -march=native -mcpu=native -Os	dolbeau/mipsel-msa
gcc -march=native -mcpu=native -O2	dolbeau/mipsel-msa
gcc -march=native -mcpu=native -O3	dolbeau/mipsel-msa
gcc -march=native -mcpu=native -Os	dolbeau/mipsel-msa

Compiler output

Implementation: crypto_stream/chacha8/dolbeau/amd64-avx2
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

api.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
chacha.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: 24, namely:

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

Compiler output

Implementation: crypto_stream/chacha8/amd64-ssse3
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

api.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
chacha.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: 18, namely:

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	amd64-ssse3 e/amd64-3 e/amd64-xmm6
gcc -funroll-loops -march=native -mcpu=native -O3	amd64-ssse3 e/amd64-3 e/amd64-xmm6
gcc -funroll-loops -march=native -mcpu=native -Os	amd64-ssse3 e/amd64-3 e/amd64-xmm6
gcc -march=native -mcpu=native -O2	amd64-ssse3 e/amd64-3 e/amd64-xmm6
gcc -march=native -mcpu=native -O3	amd64-ssse3 e/amd64-3 e/amd64-xmm6
gcc -march=native -mcpu=native -Os	amd64-ssse3 e/amd64-3 e/amd64-xmm6

Compiler output

Implementation: crypto_stream/chacha8/moon/avx/64
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

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

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	moon/avx/64 moon/avx2/64 moon/xop/64
gcc -funroll-loops -march=native -mcpu=native -O3	moon/avx/64 moon/avx2/64 moon/xop/64
gcc -funroll-loops -march=native -mcpu=native -Os	moon/avx/64 moon/avx2/64 moon/xop/64
gcc -march=native -mcpu=native -O2	moon/avx/64 moon/avx2/64 moon/xop/64
gcc -march=native -mcpu=native -O3	moon/avx/64 moon/avx2/64 moon/xop/64
gcc -march=native -mcpu=native -Os	moon/avx/64 moon/avx2/64 moon/xop/64

Compiler output

Implementation: crypto_stream/chacha8/moon/sse2/64
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

crypto_stream.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
chacha.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: 12, namely:

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	moon/sse2/64 moon/ssse3/64
gcc -funroll-loops -march=native -mcpu=native -O3	moon/sse2/64 moon/ssse3/64
gcc -funroll-loops -march=native -mcpu=native -Os	moon/sse2/64 moon/ssse3/64
gcc -march=native -mcpu=native -O2	moon/sse2/64 moon/ssse3/64
gcc -march=native -mcpu=native -O3	moon/sse2/64 moon/ssse3/64
gcc -march=native -mcpu=native -Os	moon/sse2/64 moon/ssse3/64

Compiler output

Implementation: crypto_stream/chacha8/krovetz/avx2
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

stream.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
stream.c: stream.c: In function 'crypto_stream_chacha8_krovetz_avx2_xor':
stream.c: stream.c:58:13: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
stream.c: __m256i s0 = _mm256_broadcastsi128_si256(*(__m128i *)sigma);
stream.c: ^~
stream.c: In file included from /usr/lib/gcc/x86_64-pc-linux-gnu/6.2.1/include/immintrin.h:43:0,
stream.c: from stream.c:8:
stream.c: /usr/lib/gcc/x86_64-pc-linux-gnu/6.2.1/include/avx2intrin.h:574:1: error: inlining failed in call to always_inline '_mm256_or_si256': target specific option mismatch
stream.c: _mm256_or_si256 (__m256i __A, __m256i __B)
stream.c: ^~~~~~~~~~~~~~~
stream.c: stream.c:63:13: note: called from here
stream.c: __m256i s3 = _mm256_or_si256(
stream.c: ^~
stream.c: In file included from /usr/lib/gcc/x86_64-pc-linux-gnu/6.2.1/include/immintrin.h:43:0,
stream.c: from stream.c:8:
stream.c: /usr/lib/gcc/x86_64-pc-linux-gnu/6.2.1/include/avx2intrin.h:655:1: error: inlining failed in call to always_inline '_mm256_slli_si256': target specific option mismatch
stream.c: _mm256_slli_si256 (__m256i __A, const int __N)
stream.c: ^~~~~~~~~~~~~~~~~
stream.c: stream.c:63:18: note: called from here
stream.c: __m256i s3 = _mm256_or_si256(
stream.c: ^~~~~~~~~~~~~~~~
stream.c: _mm256_slli_si256(_mm256_broadcastq_epi64(*(__m128i *)n), 8),
stream.c: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
stream.c: _mm256_set_epi32(0,0,0,1,0,0,0,0)
stream.c: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
stream.c: ...

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

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

Compiler output

Implementation: crypto_stream/chacha8/goll_gueron
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

stream.c: gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
stream.c: stream.c:126:2: error: #error -- Implementation supports only microarchitectures with support for Advanced Vector Extensions (AVX2 or AVX512).
stream.c: #error -- Implementation supports only microarchitectures with support for Advanced Vector Extensions (AVX2 or AVX512).
stream.c: ^~~~~

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

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

Compiler output

Implementation: crypto_stream/chacha8/krovetz/vec128
Compiler: gcc -funroll-loops -march=native -mcpu=native -O2

stream.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: 6, namely:

Compiler	Implementations
gcc -funroll-loops -march=native -mcpu=native -O2	krovetz/vec128
gcc -funroll-loops -march=native -mcpu=native -O3	krovetz/vec128
gcc -funroll-loops -march=native -mcpu=native -Os	krovetz/vec128
gcc -march=native -mcpu=native -O2	krovetz/vec128
gcc -march=native -mcpu=native -O3	krovetz/vec128
gcc -march=native -mcpu=native -Os	krovetz/vec128