- User Since
- Jun 28 2018, 9:57 PM (64 w, 3 d)
Wed, Sep 18
Mon, Sep 16
remove one useless comment.
Sorry for the delay. Ready for further review.
Sun, Sep 15
address Stefan and Amy Comments
Tue, Sep 10
Thanks for your confirm @hfinkel .
I am still working on the final patch. I will request code review later after I work out final patch.
revert file name changing to get a better diff shown.
Mon, Sep 9
@hfinkel Hi Hal, sorry for pushing you on this issue. Could you give some comments for this patch? As I tested on Power9, the patch gains 9.6% on cpu2017 benchmark exchanges, 2.0% on benchmark x264 for base rate, for peak rate, gains 2.1% on cactuBSSN and 1.9% on xalancbmk. Thanks.
Sun, Sep 8
address @nikic comments
Tue, Sep 3
new implementation not based on makeGuaranteedNoWrapRegion
Mon, Sep 2
delete incorrectly added file
@hfinkel Hi Hal, could you help to have a glance at this patch and give comment about if it is the right direction to implement such transform based on PPCPreIncPrep pass you have implemented. Thanks a lot.
address @stefanp comments
Aug 20 2019
Agree with your comment. @nikic Now we may get a smaller range than its real one with this patch, it may leads to wrong opt. Get a bigger range like trunk is conservative, but right. I will seek for another change for this issue.
Aug 18 2019
Hi @reames Could you help to have another look at this? Thanks
Aug 15 2019
Aug 12 2019
Aug 11 2019
address @reames comments
Aug 6 2019
Jul 25 2019
Jul 24 2019
Please fix the minor comments when commit this.
I will commit another NFC patch to merge addWithNoSignedWrap and new added addWithNoWrap.
address @reames comments
Jul 23 2019
Jul 18 2019
Jul 17 2019
Jul 16 2019
Jul 10 2019
Thanks very much for doing this, this is exactly what I want to do as a follow-up patch for https://reviews.llvm.org/D64422.
Jul 9 2019
Thanks for reverting @jsji . I will see how to fix it.
address @samparker comments.
Jul 8 2019
Jul 4 2019
Jul 3 2019
address review comments about test case.
@samparker Thanks for your comment, Sam. Updated the patch. I use the same source in both opt test and llc test, since the testing point is not the same, hope this is ok.
Jul 2 2019
Jun 26 2019
call target hook function outside the loop, it should be called once per loop.
Thanks for your comments @hfinkel .
codes moved to class TargetTransformInfo can not be called by both class HardwareLoops and class PPCTargetTransformInfo.
Move the code to class HardwareLoopInfo in NFC patch https://reviews.llvm.org/rL364415. FYI @samparker
committed in https://reviews.llvm.org/rL364397
Jun 25 2019
This patch causes a lot of cases fail except platform X86 and ARM.