[FFmpeg-devel] PATCH: av_strtod

[Pavel Pavlov]

> I tested again with attached patch, and it's not "much" slower like I
said before, but actually ~2% slower for x86_64 and ~6% slower for
x86_32. It adds 8k of object code.

> Besides, you don't "have to" make it work on intel compiler. It is
just an optimization. Optimizations can easily be disabled if the
compiler doesn't like them (for example x86 optimizations are disabled >
for ppc). Everything worked fine without them there, just a little bit
slower.

I reviewd your attached code and I did it differently, I didn't use c.
Could you bechmark my code and tell if it's actually faster or no?


See attached files (don't forget to #define __ICL inside that mlpdsp.c)

I include that mlpdsp_template.c multiple times, otherwise I had
recursion limit error. Initially I made it in a way that
mlpdsp_template.c would include iteself 44 times :)
I verified generated asm code and it's all good
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mlpdsp.c
Type: application/octet-stream
Size: 10003 bytes
Desc: mlpdsp.c
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20090601/a3fcb274/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mlpdsp_template.c
Type: application/octet-stream
Size: 5472 bytes
Desc: mlpdsp_template.c
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20090601/a3fcb274/attachment-0001.obj>