[FFmpeg-devel] Patch: Inline asm fixes for Intel compiler on Windows
michaelni at gmx.at
Sun Mar 30 23:41:19 CEST 2014
On Sun, Mar 30, 2014 at 12:45:42PM +1100, Matt Oliver wrote:
> > did you read what i wrote ? ;)
> Yeh, got what you meant, If this were to be applied then some more
> benchmarking would probably be in order first but if cdq is valid then as
> said this is rather all moot (I will make sure I include such results in
> the future though).
> Why not create a macro that expands to the correct instruction supported
> > for each compiler?
> Actually this is what I did in the first place (see previous patch in this
> email chain). But that patch wasnt applied so this patch was more just
> providing an alternative solution so people could decide which they
> preferred. I should point out that a patch such as thist one would not be
> my preferred approach it was just a different attempt to find a solution.
> > Maybe we should indeed just switch it and figure out a solution if and
> > when someone actually complains and we thus have a proper test-case?
> > yes, maybe
> > can someone provide a patch or push with matching commit message ?
> If cdq works for everyone then obviously the best and cleanest solution
> would be just changing it back. I have attached the necessary patch.
> cabac.h | 4 ++--
> mathops.h | 2 +-
> 2 files changed, 3 insertions(+), 3 deletions(-)
> 7c9da7074a9dc21159ec82db666e8d8124d5d809 Use-intel-compliant-CDQ-instead-of-CLTD-in-inline-as.patch
> From 7c40d0f7109cb45a09d274a7feaa33b71df71177 Mon Sep 17 00:00:00 2001
> From: Matt Oliver <protogonoi at gmail.com>
> Date: Sun, 30 Mar 2014 12:34:49 +1100
> Subject: [PATCH] Use intel compliant CDQ instead of CLTD in inline asm.
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
You can kill me, but you cannot change the truth.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 198 bytes
Desc: Digital signature
More information about the ffmpeg-devel