llvm/lib/Transforms/InstCombine, branch release_33http://llvm.orghttps://git.amat.us/llvm/atom/lib/Transforms/InstCombine?h=release_332013-05-06T21:21:31ZInstCombine: (X ^ signbit) + C -> X + (signbit ^ C)2013-05-06T21:21:31ZDavid Majnemerdavid.majnemer@gmail.com2013-05-06T21:21:31Zurn:sha1:8ec23cb07e22198a720c4e151241059cca215c08
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181249 91177308-0d34-0410-b5e6-96231b3b80d8
Provide InstCombines for the following 3 cases:2013-05-06T16:55:50ZJean-Luc Dupratjduprat@apple.com2013-05-06T16:55:50Zurn:sha1:c5cf6e536598a3b1e30fce616b771d66a071a54c
A * (1 - (uitofp i1 C)) -> select C, 0, A
B * (uitofp i1 C) -> select C, B, 0
select C, 0, A + select C, B, 0 -> select C, B, A
These come up in code that has been hand-optimized from a select to a linear blend,
on platforms where that may have mattered. We want to undo such changes
with the following transform:
A*(1 - uitofp i1 C) + B*(uitofp i1 C) -> select C, A, B
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181216 91177308-0d34-0410-b5e6-96231b3b80d8
Revert r164763 because it introduces new shuffles.2013-05-06T02:39:09ZNadav Rotemnrotem@apple.com2013-05-06T02:39:09Zurn:sha1:4ee312bac138a18c39353d4736eb5921532f9624
Thanks Nick Lewycky for pointing this out.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181177 91177308-0d34-0410-b5e6-96231b3b80d8
Add ArrayRef constructor from None, and do the cleanups that this constructor enables2013-05-05T00:40:33ZDmitri Gribenkogribozavr@gmail.com2013-05-05T00:40:33Zurn:sha1:5c332dbd30d9398ed25b30c3080506f7b8e92290
Patch by Robert Wilhelm.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181138 91177308-0d34-0410-b5e6-96231b3b80d8
Tabs to spaces. No functionality change.2013-05-04T01:08:15ZNick Lewyckynicholas@mxc.ca2013-05-04T01:08:15Zurn:sha1:cd4e5e9b34c1f155cf4495431f7cc6e10e2d390b
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181082 91177308-0d34-0410-b5e6-96231b3b80d8
This patch breaks up Wrap.h so that it does not have to include all of 2013-05-01T20:59:00ZFilip Pizlofpizlo@apple.com2013-05-01T20:59:00Zurn:sha1:40be1e85665d10f5444186f0e7106e368dd735b8
the things, and renames it to CBindingWrapping.h. I also moved
CBindingWrapping.h into Support/.
This new file just contains the macros for defining different wrap/unwrap
methods.
The calls to those macros, as well as any custom wrap/unwrap definitions
(like for array of Values for example), are put into corresponding C++
headers.
Doing this required some #include surgery, since some .cpp files relied
on the fact that including Wrap.h implicitly caused the inclusion of a
bunch of other things.
This also now means that the C++ headers will include their corresponding
C API headers; for example Value.h must include llvm-c/Core.h. I think
this is harmless, since the C API headers contain just external function
declarations and some C types, so I don't believe there should be any
nasty dependency issues here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180881 91177308-0d34-0410-b5e6-96231b3b80d8
Revert "InstCombine: Fold more shuffles of shuffles."2013-05-01T00:25:27ZJim Grosbachgrosbach@apple.com2013-05-01T00:25:27Zurn:sha1:10cc563bfe330cf4118b2f0db6706c244e77ebb3
This reverts commit r180802
There's ongoing discussion about whether this is the right place to make
this transformation. Reverting for now while we figure it out.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180834 91177308-0d34-0410-b5e6-96231b3b80d8
InstCombine: Fold more shuffles of shuffles.2013-04-30T20:43:52ZJim Grosbachgrosbach@apple.com2013-04-30T20:43:52Zurn:sha1:6548096a2e2b34e685680e6e1055b8e407c2c243
Always fold a shuffle-of-shuffle into a single shuffle when there's only one
input vector in the first place. Continue to be more conservative when there's
multiple inputs.
rdar://13402653
PR15866
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180802 91177308-0d34-0410-b5e6-96231b3b80d8
Fix a bug in foldSelectICmpAndOr.2013-04-30T10:36:33ZDavid Majnemerdavid.majnemer@gmail.com2013-04-30T10:36:33Zurn:sha1:527db3f26b2b5fc3855ff671344e9691cbd54d37
Differences in bitwidth between X and Y could exist even if C1 and C2 have
the same Log2 representation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180779 91177308-0d34-0410-b5e6-96231b3b80d8
Fix "Combine bit test + conditional or into simple math"2013-04-30T08:57:58ZDavid Majnemerdavid.majnemer@gmail.com2013-04-30T08:57:58Zurn:sha1:defce4cfd607e013169033164def7a3e558bfd54
This fixes the optimization introduced in r179748 and reverted in r179750.
While the optimization was sound, it did not properly respect differences in
bit-width.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180777 91177308-0d34-0410-b5e6-96231b3b80d8