Implementation notes: x86, thoth, crypto_hash/jh256

Computer: thoth
Architecture: x86
CPU ID: AuthenticAMD-00000622-0183f9ff
SUPERCOP version: 20160806
Operation: crypto_hash
Primitive: jh256

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
213024	`bitslice_opt32`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
214758	`bitslice_opt32`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
220369	`bitslice_opt32`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
227267	`bitslice_opt32`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
243191	`simple`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
246319	`bitslice_ref32`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
264304	`simple`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
270361	`simple`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
270429	`simple`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
280157	`simple`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
296107	`bitslice_ref32`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
310427	`bitslice_opt64`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
311043	`bitslice_ref32`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
311342	`bitslice_opt64`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
316800	`bitslice_ref32`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
317734	`bitslice_opt64`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
318265	`bitslice_opt64`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
321568	`bitslice_opt32`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
333411	`bitslice_ref32`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
354755	`bitslice_ref64`	`gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv`	20160726	20160724
372332	`bitslice_opt64`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724
388360	`bitslice_ref64`	`gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv`	20160726	20160724
388897	`bitslice_ref64`	`gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv`	20160726	20160724
397357	`bitslice_ref64`	`gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv`	20160726	20160724
450560	`bitslice_ref64`	`clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments`	20160726	20160724

Compiler output

Implementation: crypto_hash/jh256/bitslice_sse2_opt32
Compiler: clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments

nist.c: In file included from nist.c:2:
nist.c: ./jh_sse2_opt32.h:247:12: error: always_inline function '_mm_xor_si128' requires target feature 'sse2', but would be inlined into function 'F8' that is compiled without support for 'sse2'
nist.c: y0 = XOR(y0, LOAD(state->gt;buffer));
nist.c: ^
nist.c: ./jh_sse2_opt32.h:112:23: note: expanded from macro 'XOR'
nist.c: #define XOR(x,y) _mm_xor_si128((x),(y)) /*XOR(x,y) = x ^ y, where x and y are two 128-bit word*/
nist.c: ^
nist.c: ./jh_sse2_opt32.h:247:20: error: always_inline function '_mm_load_si128' requires target feature 'sse2', but would be inlined into function 'F8' that is compiled without support for 'sse2'
nist.c: y0 = XOR(y0, LOAD(state->gt;buffer));
nist.c: ^
nist.c: ./jh_sse2_opt32.h:142:23: note: expanded from macro 'LOAD'
nist.c: #define LOAD(p) _mm_load_si128((__m128i *)(p)) /*load 16 bytes from the memory address p, return a 128-bit word, where p is the multile of 16 bytes*/
nist.c: ^
nist.c: ./jh_sse2_opt32.h:248:12: error: always_inline function '_mm_xor_si128' requires target feature 'sse2', but would be inlined into function 'F8' that is compiled without support for 'sse2'
nist.c: y1 = XOR(y1, LOAD(state->gt;buffer+16));
nist.c: ^
nist.c: ./jh_sse2_opt32.h:112:23: note: expanded from macro 'XOR'
nist.c: #define XOR(x,y) _mm_xor_si128((x),(y)) /*XOR(x,y) = x ^ y, where x and y are two 128-bit word*/
nist.c: ^
nist.c: ./jh_sse2_opt32.h:248:20: error: always_inline function '_mm_load_si128' requires target feature 'sse2', but would be inlined into function 'F8' that is compiled without support for 'sse2'
nist.c: y1 = XOR(y1, LOAD(state->gt;buffer+16));
nist.c: ^
nist.c: ./jh_sse2_opt32.h:142:23: note: expanded from macro 'LOAD'
nist.c: #define LOAD(p) _mm_load_si128((__m128i *)(p)) /*load 16 bytes from the memory address p, return a 128-bit word, where p is the multile of 16 bytes*/
nist.c: ^
nist.c: ...

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments	bitslice_sse2_opt32

Compiler output

Implementation: crypto_hash/jh256/bitslice_sse2_opt64
Compiler: clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments

nist.c: In file included from nist.c:2:
nist.c: ./jh_sse2_opt64.h:255:12: error: always_inline function '_mm_xor_si128' requires target feature 'sse2', but would be inlined into function 'F8' that is compiled without support for 'sse2'
nist.c: y0 = XOR(y0, LOAD(state->gt;buffer));
nist.c: ^
nist.c: ./jh_sse2_opt64.h:112:23: note: expanded from macro 'XOR'
nist.c: #define XOR(x,y) _mm_xor_si128((x),(y)) /*XOR(x,y) = x ^ y, where x and y are two 128-bit word*/
nist.c: ^
nist.c: ./jh_sse2_opt64.h:255:20: error: always_inline function '_mm_load_si128' requires target feature 'sse2', but would be inlined into function 'F8' that is compiled without support for 'sse2'
nist.c: y0 = XOR(y0, LOAD(state->gt;buffer));
nist.c: ^
nist.c: ./jh_sse2_opt64.h:141:23: note: expanded from macro 'LOAD'
nist.c: #define LOAD(p) _mm_load_si128((__m128i *)(p)) /*load 16 bytes from the memory address p, return a 128-bit word, where p is the multile of 16 bytes*/
nist.c: ^
nist.c: ./jh_sse2_opt64.h:256:12: error: always_inline function '_mm_xor_si128' requires target feature 'sse2', but would be inlined into function 'F8' that is compiled without support for 'sse2'
nist.c: y1 = XOR(y1, LOAD(state->gt;buffer+16));
nist.c: ^
nist.c: ./jh_sse2_opt64.h:112:23: note: expanded from macro 'XOR'
nist.c: #define XOR(x,y) _mm_xor_si128((x),(y)) /*XOR(x,y) = x ^ y, where x and y are two 128-bit word*/
nist.c: ^
nist.c: ./jh_sse2_opt64.h:256:20: error: always_inline function '_mm_load_si128' requires target feature 'sse2', but would be inlined into function 'F8' that is compiled without support for 'sse2'
nist.c: y1 = XOR(y1, LOAD(state->gt;buffer+16));
nist.c: ^
nist.c: ./jh_sse2_opt64.h:141:23: note: expanded from macro 'LOAD'
nist.c: #define LOAD(p) _mm_load_si128((__m128i *)(p)) /*load 16 bytes from the memory address p, return a 128-bit word, where p is the multile of 16 bytes*/
nist.c: ^
nist.c: ...

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments	bitslice_sse2_opt64

Compiler output

Implementation: crypto_hash/jh256/bitslice_sse2_opt32
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv

nist.c: In file included from nist.c:2:0:
nist.c: jh_sse2_opt32.h: In function 'F8':
nist.c: jh_sse2_opt32.h:247:10: warning: SSE vector return without SSE enabled changes the ABI [-Wpsabi]
nist.c: y0 = XOR(y0, LOAD(state->gt;buffer));
nist.c: ^
nist.c: In file included from jh_sse2_opt32.h:32:0,
nist.c: from nist.c:2:
nist.c: /usr/lib/gcc/i686-linux-gnu/5/include/emmintrin.h:1286:1: error: inlining failed in call to always_inline '_mm_xor_si128': target specific option mismatch
nist.c: _mm_xor_si128 (__m128i __A, __m128i __B)
nist.c: ^
nist.c: In file included from nist.c:2:0:
nist.c: jh_sse2_opt32.h:250:10: error: called from here
nist.c: y3 = XOR(y3, LOAD(state->gt;buffer+48));
nist.c: ^
nist.c: In file included from jh_sse2_opt32.h:32:0,
nist.c: from nist.c:2:
nist.c: /usr/lib/gcc/i686-linux-gnu/5/include/emmintrin.h:690:1: error: inlining failed in call to always_inline '_mm_load_si128': target specific option mismatch
nist.c: _mm_load_si128 (__m128i const *__P)
nist.c: ^
nist.c: In file included from nist.c:2:0:
nist.c: jh_sse2_opt32.h:112:23: error: called from here
nist.c: #define XOR(x,y) _mm_xor_si128((x),(y)) /*XOR(x,y) = x ^ y, where x and y are two 128-bit word*/
nist.c: ^
nist.c: jh_sse2_opt32.h:250:12: note: in expansion of macro 'XOR'
nist.c: y3 = XOR(y3, LOAD(state->gt;buffer+48));
nist.c: ...

Number of similar (compiler,implementation) pairs: 4, namely:

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv	bitslice_sse2_opt32
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv	bitslice_sse2_opt32
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv	bitslice_sse2_opt32
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv	bitslice_sse2_opt32

Compiler output

Implementation: crypto_hash/jh256/bitslice_sse2_opt64
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv

nist.c: In file included from nist.c:2:0:
nist.c: jh_sse2_opt64.h: In function 'F8':
nist.c: jh_sse2_opt64.h:255:10: warning: SSE vector return without SSE enabled changes the ABI [-Wpsabi]
nist.c: y0 = XOR(y0, LOAD(state->gt;buffer));
nist.c: ^
nist.c: In file included from jh_sse2_opt64.h:32:0,
nist.c: from nist.c:2:
nist.c: /usr/lib/gcc/i686-linux-gnu/5/include/emmintrin.h:1286:1: error: inlining failed in call to always_inline '_mm_xor_si128': target specific option mismatch
nist.c: _mm_xor_si128 (__m128i __A, __m128i __B)
nist.c: ^
nist.c: In file included from nist.c:2:0:
nist.c: jh_sse2_opt64.h:258:10: error: called from here
nist.c: y3 = XOR(y3, LOAD(state->gt;buffer+48));
nist.c: ^
nist.c: In file included from jh_sse2_opt64.h:32:0,
nist.c: from nist.c:2:
nist.c: /usr/lib/gcc/i686-linux-gnu/5/include/emmintrin.h:690:1: error: inlining failed in call to always_inline '_mm_load_si128': target specific option mismatch
nist.c: _mm_load_si128 (__m128i const *__P)
nist.c: ^
nist.c: In file included from nist.c:2:0:
nist.c: jh_sse2_opt64.h:112:23: error: called from here
nist.c: #define XOR(x,y) _mm_xor_si128((x),(y)) /*XOR(x,y) = x ^ y, where x and y are two 128-bit word*/
nist.c: ^
nist.c: jh_sse2_opt64.h:258:12: note: in expansion of macro 'XOR'
nist.c: y3 = XOR(y3, LOAD(state->gt;buffer+48));
nist.c: ...

Number of similar (compiler,implementation) pairs: 4, namely:

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv	bitslice_sse2_opt64
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv	bitslice_sse2_opt64
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv	bitslice_sse2_opt64
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv	bitslice_sse2_opt64