llvm/lib/Transforms/Vectorize, branch release_32http://llvm.orghttps://git.amat.us/llvm/atom/lib/Transforms/Vectorize?h=release_322012-11-29T00:38:51ZMerge in r168765 (BBVectorize bug fix)2012-11-29T00:38:51ZHal Finkelhfinkel@anl.gov2012-11-29T00:38:51Zurn:sha1:0416bc140fb88b3e5a0d5ce3a278655b1a539324
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_32@168839 91177308-0d34-0410-b5e6-96231b3b80d8
Merge r167942, r167966 to fix non-deterministic behavior in BBVectorize2012-11-15T00:34:28ZHal Finkelhfinkel@anl.gov2012-11-15T00:34:28Zurn:sha1:d0e51c6d1974636579f687f9454c85ebeacc7cd2
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_32@168012 91177308-0d34-0410-b5e6-96231b3b80d8
Merge BBVectorizer changes r167731, r167743, r167750, r167784, r167811, r167817.2012-11-14T22:58:30ZHal Finkelhfinkel@anl.gov2012-11-14T22:58:30Zurn:sha1:adea7bc349a18a1c8f9ba94bff61d5938075e600
These changes fix a serious interaction problem with the cost model on x86 that
could cause the vectorizer to enter an infinite loop (and sometimes crash in
other ways).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_32@167993 91177308-0d34-0410-b5e6-96231b3b80d8
Fix a comment typo and add comments.2012-11-11T05:15:00ZNadav Rotemnrotem@apple.com2012-11-11T05:15:00Zurn:sha1:9005b8d27d68caeb893d74aa75cd679972d86cef
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167684 91177308-0d34-0410-b5e6-96231b3b80d8
Add support for memory runtime check. When we can, we calculate array bounds.2012-11-09T07:09:44ZNadav Rotemnrotem@apple.com2012-11-09T07:09:44Zurn:sha1:b1bf1eeede72b8c93505dd80fdf21aed0e205c7d
If the arrays are found to be disjoint then we run the vectorized version of
the loop. If they are not, we run the scalar code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167608 91177308-0d34-0410-b5e6-96231b3b80d8
Fix sign compare warning. Patch by Mahesha HS.2012-11-02T05:24:00ZChandler Carruthchandlerc@gmail.com2012-11-02T05:24:00Zurn:sha1:6e48f0307758096d06d0e87875294c76df81dec1
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167282 91177308-0d34-0410-b5e6-96231b3b80d8
BBVectorize: Use target costs for incoming and outgoing values instead of the depth heuristic.2012-11-01T21:50:12ZHal Finkelhfinkel@anl.gov2012-11-01T21:50:12Zurn:sha1:78fd353d5e5daedc47ecc31b6193ca48793c249c
When target cost information is available, compute explicit costs of inserting and
extracting values from vectors. At this point, all costs are estimated using the
target information, and the chain-depth heuristic is not needed. As a result, it is now, by
default, disabled when using target costs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167256 91177308-0d34-0410-b5e6-96231b3b80d8
BBVectorize: Account for internal shuffle costs2012-11-01T06:26:34ZHal Finkelhfinkel@anl.gov2012-11-01T06:26:34Zurn:sha1:86ccc55c82651f91fd6a312c5f6a4b511bcd1aec
When target costs are available, use them to account for the costs of
shuffles on internal edges of the DAG of candidate pairs.
Because the shuffle costs here are currently for only the internal edges,
the current target cost model is trivial, and the chain depth requirement
is still in place, I don't yet have an easy test
case. Nevertheless, by looking at the debug output, it does seem to do the right
think to the effective "size" of each DAG of candidate pairs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167217 91177308-0d34-0410-b5e6-96231b3b80d8
LoopVectorize: Preserve NSW, NUW and IsExact flags.2012-10-31T21:40:39ZNadav Rotemnrotem@apple.com2012-10-31T21:40:39Zurn:sha1:e57b2cbce652d45ed5516e52ad82991bfa03cfd7
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167174 91177308-0d34-0410-b5e6-96231b3b80d8
Put the threshold magic number in a variable.2012-10-31T16:22:16ZNadav Rotemnrotem@apple.com2012-10-31T16:22:16Zurn:sha1:4c1b4b1fe794437cbb245b11650d9e4001c9605e
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167134 91177308-0d34-0410-b5e6-96231b3b80d8