http://llvm.org https://git.amat.us/llvm/atom/lib/Transforms/InstCombine?h=testing 2013-03-14T18:08:26Z 2013-03-14T18:08:26Z Shuxin Yang shuxin.llvm@gmail.com 2013-03-14T18:08:26Z urn:sha1:a0c9939873c404f272b3e0abb102c335146764fe Rules include: 1)1 x*y +/- x*z => x*(y +/- z) (the order of operands dosen't matter) 2) y/x +/- z/x => (y +/- z)/x The transformation is disabled if the new add/sub expr "y +/- z" is a denormal/naz/inifinity. rdar://12911472 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@177088 91177308-0d34-0410-b5e6-96231b3b80d8 2013-03-13T14:40:37Z Arnaud A. de Grandmaison arnaud.adegm@gmail.com 2013-03-13T14:40:37Z urn:sha1:bdd2d981391bd9329aba669cffcdde0c39373cb1 Only combine when the shl is only used by the icmp git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176950 91177308-0d34-0410-b5e6-96231b3b80d8 2013-03-09T11:18:59Z Jakub Staszak kubastaszak@gmail.com 2013-03-09T11:18:59Z urn:sha1:e629a33d166f55a916f40a1feae47af8ab97feed git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176765 91177308-0d34-0410-b5e6-96231b3b80d8 2013-03-06T05:44:53Z Jim Grosbach grosbach@apple.com 2013-03-06T05:44:53Z urn:sha1:186d8a3d67ccd2b2401c5d7d4e2cc15c4d1fdeae When considering folding a bitcast of an alloca into the alloca itself, make sure we don't shrink the amount of memory being allocated, or things rapidly go sideways. rdar://13324424 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176547 91177308-0d34-0410-b5e6-96231b3b80d8 2013-02-28T21:12:40Z Quentin Colombet qcolombet@apple.com 2013-02-28T21:12:40Z urn:sha1:c5a4c25b8780434a00968ed93634974a0b796a06 The instcombine recognized pattern looks like: a = b * c d = a +/- Cst or a = b * c d = Cst +/- a When creating the new operands for fadd or fsub instruction following the related fmul, the first operand was created with the second original operand (M0 was created with C1) and the second with the first (M1 with Opnd0). The fix consists in creating the new operands with the appropriate original operand, i.e., M0 with Opnd0 and M1 with C1. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176300 91177308-0d34-0410-b5e6-96231b3b80d8 2013-02-16T23:41:36Z Bill Wendling isanbard@gmail.com 2013-02-16T23:41:36Z urn:sha1:cc54889cd58322b8766525f43cc1f7cb52e4692e (or (bool?A:B),(bool?C:D)) --> (bool?(or A,C):(or B,D)) By the time the OR is visited, both the SELECTs have been visited and not optimized and the OR itself hasn't been transformed so we do this transform in the hopes that the new ORs will be optimized. The transform is explicitly disabled for vector-selects until "codegen matures to handle them better". Patch by Muhammad Tauqir! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@175380 91177308-0d34-0410-b5e6-96231b3b80d8 2013-02-15T15:18:17Z Arnaud A. de Grandmaison arnaud.adegm@gmail.com 2013-02-15T15:18:17Z urn:sha1:ad079b274d6ecd3e36fb98099864baeeeb96ac82 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@175273 91177308-0d34-0410-b5e6-96231b3b80d8 2013-02-15T14:35:47Z Arnaud A. de Grandmaison arnaud.adegm@gmail.com 2013-02-15T14:35:47Z urn:sha1:7c5c9b39c91d5f53284011c0ddbf458d553740da It enables to work with a smaller constant, which is target friendly for those which can compare to immediates. It also avoids inserting a shift in favor of a trunc, which can be free on some targets. This used to work until LLVM-3.1, but regressed with the 3.2 release. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@175270 91177308-0d34-0410-b5e6-96231b3b80d8 2013-02-13T00:19:19Z Arnaud A. de Grandmaison arnaud.adegm@gmail.com 2013-02-13T00:19:19Z urn:sha1:66bff1eec9f45733ff12b8d2307326a389270874 visitSExt is an adapted copy of the related visitZExt method, so adapt the comment accordingly. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@175019 91177308-0d34-0410-b5e6-96231b3b80d8 2013-02-11T21:41:44Z Michael Ilseman milseman@apple.com 2013-02-11T21:41:44Z urn:sha1:9c213cc3c31da04a08a8e00a3395ac33d3d18536 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174905 91177308-0d34-0410-b5e6-96231b3b80d8