- User Since
- Sep 26 2016, 7:58 AM (151 w, 1 d)
Basically I assume that nobody really cares about formatting and comments in generated code (so in one sense this is stupd, and in another sense it won't hurt much either).
Sun, Aug 18
Sat, Aug 17
Code changes basically looks good to me, but I have no idea about:
- the expected gain
- if this is handled elsewhere (or if it should be)
- if there is a (high) cost related to "not preserving the CFG"
Fri, Aug 16
Thu, Aug 15
It seems like widening vector legalization for X86 can introduce fixed point multiplication of undef values. So that is one way that such operations could appear during ISel.
Update mulfix_combine.ll test case to show diffs compared to trunk.
Tue, Aug 13
Sun, Aug 11
Corrected a code comment typo.
Sat, Aug 10
Fri, Aug 9
Thu, Aug 8
Sun, Jul 28
Sat, Jul 27
I think the problem is related to the RUN line at line 76.
Since this patch we get
Jul 9 2019
Jul 8 2019
Jul 5 2019
Jul 3 2019
What is the idea here? Just a refactoring to get rid of recursive calls? Or are you going to use the histogram for some other purposes?
(the patch is lacking description, but perhaps it's just a temporary hack an not something that is subject for submitting?)
Jun 27 2019
LGTM as well.
Jun 24 2019
Fixup in description of llvm::isTriviallyVectorizable.
Jun 19 2019
Jun 18 2019
Jun 17 2019
Jun 16 2019
Jun 10 2019
Please add test cases for scale=0 and scale=width as I assume those need special handling (UB right now?).
And if scale=0 and scale=width needs special handling, then I guess scale=1 and scale=width-1 are new boundary values so I maybe it would be nice to have tests for those scales as well.
Looks good to me (matching my assumptions about why prolog/epilog is special).
Just a minor nit inline.