please can you add some vector tests?
Jul 31 2020
Jul 30 2020
LGTM - thanks for taking this on!
Jul 29 2020
Tried to use the inline code editor - not sure if it worked that well.....
Jul 28 2020
rebase on D83948 ?
Please can you add unit test coverage at llvm/unittest/ADT/BitVectorTest.cpp
Jul 27 2020
simplify lane-crossing logic
Jul 26 2020
LGTM - sorry for the delay
LGTM - cheers!
Jul 25 2020
Always permit lane crossing on AVX2 targets, and on non-AVX2 targets, integer hops will always be split to 128-bits so permit lane crossing there as well.
These appear to be red due to timeouts:
Relax lane crossing limitations for AVX2+ targets
Jul 24 2020
@aqjune Please can you take a look at https://bugs.llvm.org/show_bug.cgi?id=46831 (https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=24278) - it seems related to this patch
Jul 23 2020
rebase with new tests
@craig.topper Maybe add something to the 12.00 release notes?
Jul 22 2020
Jul 21 2020
I haven't done any perf measurements with this yet. Anything specific people would like to see?
Jul 20 2020
Do we need additional mask/maskz predicate test coverage?
Have you triaged whats going on in the mulfixsat test regressions?
Jul 19 2020
Maybe add langref docs entries as well?
Jul 16 2020
LGTM - no objections, I don't think we have much that specifically targets optimal codegen on p4 - a few of the TTI costs and thats about it.
Jul 15 2020
We've been making progress performing similar ops in the vector-combiner pass based on cost metrics - have you looked at performing it there?
Jul 14 2020
Jul 13 2020
LGTM - I'm curious whether rotates across more load/stores would be useful or not: https://gcc.godbolt.org/z/196ar9
Jul 12 2020
LGTM - cheers
one minor in SelectionDAG - but a wasm dev needs to looks at the target code
Jul 10 2020
reverse ping - any luck on the amdgpu regressions?