mirrors/librempeg
1
0
Fork 0
mirror of https://github.com/librempeg/librempeg synced 2024-11-22 18:49:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

lavu/sha512: Fully unroll the transform function loops

Browse Source 
crypto_bench SHA-512 results using an AMD Athlon X2 7750+, mingw32-w64 GCC 4.7.3 x86_64

Before:
lavu       SHA-512      size: 1048576  runs:   1024  time:   12.737 +- 0.147

After:
lavu       SHA-512      size: 1048576  runs:   1024  time:   11.670 +- 0.173

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This commit is contained in:
James Almer 2013-09-09 20:16:40 -03:00 committed by Michael Niedermayer
parent 7e4fe5162a
commit bbcaf25d4d
1 changed files with 25 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

45
libavutil/sha512.c
View File

@ -150,27 +150,32 @@ static void sha512_transform(uint64_t *state, const uint8_t buffer[128])
a = T1 + T2;
}
#else
for (i = 0; i < 16 - 7;) {
ROUND512_0_TO_15(a, b, c, d, e, f, g, h);
ROUND512_0_TO_15(h, a, b, c, d, e, f, g);
ROUND512_0_TO_15(g, h, a, b, c, d, e, f);
ROUND512_0_TO_15(f, g, h, a, b, c, d, e);
ROUND512_0_TO_15(e, f, g, h, a, b, c, d);
ROUND512_0_TO_15(d, e, f, g, h, a, b, c);
ROUND512_0_TO_15(c, d, e, f, g, h, a, b);
ROUND512_0_TO_15(b, c, d, e, f, g, h, a);
}
for (; i < 80 - 7;) {
ROUND512_16_TO_80(a, b, c, d, e, f, g, h);
ROUND512_16_TO_80(h, a, b, c, d, e, f, g);
ROUND512_16_TO_80(g, h, a, b, c, d, e, f);
ROUND512_16_TO_80(f, g, h, a, b, c, d, e);
ROUND512_16_TO_80(e, f, g, h, a, b, c, d);
ROUND512_16_TO_80(d, e, f, g, h, a, b, c);
ROUND512_16_TO_80(c, d, e, f, g, h, a, b);
ROUND512_16_TO_80(b, c, d, e, f, g, h, a);
}
#define R512_0 \
ROUND512_0_TO_15(a, b, c, d, e, f, g, h); \
ROUND512_0_TO_15(h, a, b, c, d, e, f, g); \
ROUND512_0_TO_15(g, h, a, b, c, d, e, f); \
ROUND512_0_TO_15(f, g, h, a, b, c, d, e); \
ROUND512_0_TO_15(e, f, g, h, a, b, c, d); \
ROUND512_0_TO_15(d, e, f, g, h, a, b, c); \
ROUND512_0_TO_15(c, d, e, f, g, h, a, b); \
ROUND512_0_TO_15(b, c, d, e, f, g, h, a)
i = 0;
R512_0; R512_0;
#define R512_16 \
ROUND512_16_TO_80(a, b, c, d, e, f, g, h); \
ROUND512_16_TO_80(h, a, b, c, d, e, f, g); \
ROUND512_16_TO_80(g, h, a, b, c, d, e, f); \
ROUND512_16_TO_80(f, g, h, a, b, c, d, e); \
ROUND512_16_TO_80(e, f, g, h, a, b, c, d); \
ROUND512_16_TO_80(d, e, f, g, h, a, b, c); \
ROUND512_16_TO_80(c, d, e, f, g, h, a, b); \
ROUND512_16_TO_80(b, c, d, e, f, g, h, a)
R512_16; R512_16; R512_16; R512_16;
R512_16; R512_16; R512_16; R512_16;
#endif
state[0] += a;
state[1] += b;