This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PATCH: PR target/46519: Missing vzeroupper
- From: Mark Mitchell <mark at codesourcery dot com>
- To: Uros Bizjak <ubizjak at gmail dot com>
- Cc: "H.J. Lu" <hjl dot tools at gmail dot com>, gcc-patches at gcc dot gnu dot org, Richard Guenther <rguenther at suse dot de>, Jakub Jelinek <jakub at redhat dot com>
- Date: Fri, 31 Dec 2010 17:05:18 -0800
- Subject: Re: PATCH: PR target/46519: Missing vzeroupper
- References: <20101217190347.GA2856@intel.com> <AANLkTinempMG4qbZsZ_tHMnyHqpXd+2JDZEYzqdASEWo@mail.gmail.com> <AANLkTi=ky0A+oH9e23zX=yWrwjvFKzkNG1qVrvQ1bcJO@mail.gmail.com> <AANLkTi=K3bPqru+joTk1fsjthVnSe0EMkoEF=tGr7SxF@mail.gmail.com> <AANLkTi=0tNb1FX7XweRCUQBh6OUjNW7f6+vyO0YuYk=z@mail.gmail.com> <AANLkTi=e2ZemY8AaWx6sGOBS_RMiS3=LG9z9pZ8SMbsq@mail.gmail.com>
On 12/30/2010 3:19 AM, Uros Bizjak wrote:
>> We will investigate LCM for 4.7. In the meantime, here is a small patch
>> on top of the current one. If the upper 128bits are never changed in a basic
>> block, we can skip it in the later passes. OK for trunk together with the
>> current patch?
For avoidance of doubt, since Uros explicitly asked for RM comments: I
have no objections to the patch, if the x86 maintainers are happy with it.
However, this comment:
>> I'm not sure that the algorithm is correct (and I don't have enough
>> experience in this area), so I'd rather leave the review to someone
>> else. AFAICS, there can be 20 passes, and from comments, it is
>> questionable if this is enough.
concerns me.
Do someone have confidence that this algorithm is correct, in the sense
that we will not generate wrong code? And are we talking about 20
passes over the complete set of basic blocks? That sounds pretty expensive.
Thank you,
--
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-3385 x713