Hello,
> + AV_WB64(y,KE);
> > +
> KE=SP[0][y[0]]^SP[1][y[1]]^SP[2][y[2]]^SP[3][y[3]]^SP[4][y[4]]^SP[5][y[5]]^SP[6][y[6]]^SP[7][y[7]];
> > + return KE;
>
> have you tried it without the y[] array ?
> storing and loading from memory may (or may not) be slower
I have tried without y[] array. And there seems to be some improvement in
the time.
lavu CAMELLIA size: 1048576 runs: 1024 time: 21.549 +- 0.176
I have attached a patch with the respective changes. Would you want me to
convert the F() into a macro ?
Thanks,
Supraja
From 95f089edee3cda76f7b5efe82869a42e841b7a3a Mon Sep 17 00:00:00 2001
From: Supraja Meedinti <supraja0...@gmail.com>
Date: Tue, 10 Feb 2015 21:24:07 +0530
Subject: [PATCH] libavutil: camellia: remove unwanted memory loads
Signed-off-by: Supraja Meedinti <supraja0...@gmail.com>
---
libavutil/camellia.c | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/libavutil/camellia.c b/libavutil/camellia.c
index 11e57a1..be2485a 100644
--- a/libavutil/camellia.c
+++ b/libavutil/camellia.c
@@ -142,11 +142,9 @@ static void LR128(uint64_t d[2], const uint64_t K[2], int x)
static uint64_t F(uint64_t F_IN, uint64_t KE)
{
- uint8_t y[8];
KE ^= F_IN;
- AV_WB64(y,KE);
- KE=SP[0][y[0]]^SP[1][y[1]]^SP[2][y[2]]^SP[3][y[3]]^SP[4][y[4]]^SP[5][y[5]]^SP[6][y[6]]^SP[7][y[7]];
- return KE;
+ F_IN=SP[0][KE >> 56]^SP[1][(KE >> 48) & MASK8]^SP[2][(KE >> 40) & MASK8]^SP[3][(KE >> 32) & MASK8]^SP[4][(KE >> 24) & MASK8]^SP[5][(KE >> 16) & MASK8]^SP[6][(KE >> 8) & MASK8]^SP[7][KE & MASK8];
+ return F_IN;
}
static uint64_t FL(uint64_t FL_IN, uint64_t KE)
--
1.8.3.2
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
