- User Since
- Jun 17 2015, 7:07 AM (140 w, 2 d)
Wed, Feb 21
LGTM, however I am not the right person to approve patches to the LoopVectorizer.
Tue, Feb 20
Is it worthwhile to add a regression test that would trigger the assertion?
Mon, Feb 12
Fri, Feb 2
Waiting for confirmation this is ok before committing
Instead of adding -polly-stmt-granulairy=bb globally, only add the flag to the tests that need it.
Jan 23 2018
Jan 19 2018
Program leone_F30_statementnaming_indep leone_G30_epilogue_indep diff
Jan 18 2018
@bollu I agree DependenceInfo would be a good place to compute the conditional dependencies. However, as long we don't have another consumer for this data, it doesn't seem important to me.
Jan 17 2018
@bollu Alex added the code duplication because I suggested him to do so. IMHO the live-range reordering should not depend on ppcg at all. ppcg is not written as a library so in terms of software quality, we should not use it as one. On the llvm mailing list we currently discuss the inclusion of isl. I don't think we will not do the same for ppcg, meaning that if live-range reordering depends on it, we would have to remove it again should Polly be included into LLVM main.
Jan 16 2018
Fallback value can be discussed separately, it also exists in the current codebase.
Jan 15 2018
- Include isl conditionally using LLVM_INCLUDE_ISL (defaults to OFF)
- Fix include paths (llvm-config and install)
- Use -w flag only for gcc/clang
Jan 12 2018
Jan 3 2018
I'd rather remove the number of magic constants that make impactful decisions. If the goal is to reduce the computational overhead, I'd recommend to take the convex hull before lexmin/lexmin, not bail-out.
Dec 21 2017
Dec 11 2017
I filed a bug for this option: http://llvm.org/PR35623 .
LGTM, going to committ...
Dec 10 2017
Oops, embarrassing type by me. Thanks a lot.
Dec 6 2017
Dec 4 2017
After removing the isl memory annotations: LGTM.
Do you know which revision/changes in clang-format caused this?
Dec 1 2017
Thanks Alex for the patch. I just see it now. As Eli is mentioning already, could you separate the patch into separate differentials?
Nov 30 2017
Nov 27 2017
Nov 21 2017
Nov 19 2017
Remove duplicated empty lines.
Looks great (whatever you decide to to with the inline comment)
Nov 16 2017
Nov 14 2017
I cannot apply this patch, arcanist always tries to apply it on git sha1 85a1620 (current trunk without D39979).
LGTM, only nitpicks. Thanks for your work.
Nov 9 2017
Nov 6 2017
LGTM, fails for me too. Probably caused by change in clang-format: rL317325
Oct 31 2017
Abandon because the alternative was committed
Oct 30 2017
- Address Siddharth's remarks
- Fix one more TODO
- Introduce isNormalizable to resolve TODO
- Remove unused function attributes in test-case
- Add example to NormalizedPHI doxygen
Oct 27 2017
For me this is OK. We cannot reliably test undefined behavior. Tools like TypeSanitizer exist for this purpose.
Oct 26 2017
Add selfrefphi test case
Oct 24 2017
Oct 20 2017
Could you check whether this works for you:
Oct 17 2017
Thanks for improving our documentation!
Oct 6 2017
Please add at least one test case where a basic block is successfully split up and check that there are two statements now.