diff options
| author | Michael Niedermayer <michaelni@gmx.at> | 2012-02-04 20:27:02 +0100 |
|---|---|---|
| committer | Michael Niedermayer <michaelni@gmx.at> | 2012-02-04 20:27:02 +0100 |
| commit | d3f912c1f06b90960659d59bb75d2eb98932fde6 (patch) | |
| tree | 8883bdb68632c4abb85dab4c7eea03bc89b51661 /libavutil | |
| parent | 26585d2a7f5f13bec3ed8623119bf625fdb728f4 (diff) | |
| download | ffmpeg-streaming-d3f912c1f06b90960659d59bb75d2eb98932fde6.zip ffmpeg-streaming-d3f912c1f06b90960659d59bb75d2eb98932fde6.tar.gz | |
adler32: add comment for the SIMD code
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Diffstat (limited to 'libavutil')
| -rw-r--r-- | libavutil/adler32.c | 5 |
1 files changed, 5 insertions, 0 deletions
diff --git a/libavutil/adler32.c b/libavutil/adler32.c index c5f96db..bc9b9a7 100644 --- a/libavutil/adler32.c +++ b/libavutil/adler32.c @@ -57,6 +57,11 @@ unsigned long av_adler32_update(unsigned long adler, const uint8_t * buf, len2 -= 8; buf+=8; } + + //We combine the 8 interleaved adler32 checksums without overflows + //Decreasing the number of iterations would allow below code to be + //simplified but would likely be slower due to the fewer iterations + //of the inner loop s1 += ((a1+b1)*0x1000100010001)>>48; s2 += ((((a2&0xFFFF0000FFFF)+(b2&0xFFFF0000FFFF)+((a2>>16)&0xFFFF0000FFFF)+((b2>>16)&0xFFFF0000FFFF))*0x800000008)>>32) #if HAVE_BIGENDIAN |
