http://llvm.org https://git.amat.us/llvm/atom/test/Transforms/BBVectorize?h=release_32 2012-11-29T00:38:51Z 2012-11-29T00:38:51Z Hal Finkel hfinkel@anl.gov 2012-11-29T00:38:51Z urn:sha1:0416bc140fb88b3e5a0d5ce3a278655b1a539324 git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_32@168839 91177308-0d34-0410-b5e6-96231b3b80d8 2012-11-14T22:58:30Z Hal Finkel hfinkel@anl.gov 2012-11-14T22:58:30Z urn: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 2012-11-01T21:57:27Z Hal Finkel hfinkel@anl.gov 2012-11-01T21:57:27Z urn:sha1:4ac300dfac53022e690fc9a60deb42976c369a6f git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167257 91177308-0d34-0410-b5e6-96231b3b80d8 2012-11-01T21:50:12Z Hal Finkel hfinkel@anl.gov 2012-11-01T21:50:12Z urn: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 2012-10-31T15:17:07Z Hal Finkel hfinkel@anl.gov 2012-10-31T15:17:07Z urn:sha1:72465ea23d010507d3746adc126d719005981e05 BBVectorize would, except for loads and stores, always fuse instructions so that the first instruction (in the current source order) would always represent the low part of the input vectors and the second instruction would always represent the high part. This lead to too many shuffles being produced because sometimes the opposite order produces fewer of them. With this change, BBVectorize tracks the kind of pair connections that form the DAG of candidate pairs, and uses that information to reorder the pairs to avoid excess shuffles. Using this information, a future commit will be able to add VTTI-based shuffle costs to the pair selection procedure. Importantly, the number of remaining shuffles can now be estimated during pair selection. There are some trivial instruction reorderings in the test cases, and one simple additional test where we certainly want to do a reordering to avoid an unnecessary shuffle. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167122 91177308-0d34-0410-b5e6-96231b3b80d8 2012-10-30T19:47:37Z Hal Finkel hfinkel@anl.gov 2012-10-30T19:47:37Z urn:sha1:8a69a95f913dadae92d2ce0b3ac551610bf868e0 We need to make sure that we take the correct load/store alignment when the inputs are flipped. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167044 91177308-0d34-0410-b5e6-96231b3b80d8 2012-10-30T02:41:57Z Hal Finkel hfinkel@anl.gov 2012-10-30T02:41:57Z urn:sha1:c588e0e162fa08c81558871fb0c50fb51569afe3 getCastInstrCost had an assert prohibiting scalar to vector casts. Such casts, however, are allowed. This should make the vectorizer buildbot happier. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166998 91177308-0d34-0410-b5e6-96231b3b80d8 2012-10-26T19:38:09Z Hal Finkel hfinkel@anl.gov 2012-10-26T19:38:09Z urn:sha1:ecc69a1d99963b2b1bba92bdcefbfa8a8f1c497a git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166802 91177308-0d34-0410-b5e6-96231b3b80d8 2012-10-26T00:05:26Z Hal Finkel hfinkel@anl.gov 2012-10-26T00:05:26Z urn:sha1:822ab00847da841a63be4e3883cb5f442dc69069 Once vector-of-pointer support works, then this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166741 91177308-0d34-0410-b5e6-96231b3b80d8 2012-10-25T21:31:51Z Hal Finkel hfinkel@anl.gov 2012-10-25T21:31:51Z urn:sha1:32ff623c8bd3b6c0751874bceeb3e9062b1a9726 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166720 91177308-0d34-0410-b5e6-96231b3b80d8