| Commit message (Expand) | Author | Age | Files | Lines |
* | LoopVectorize: Vectorize all accesses in address space zero with unit stride | Arnold Schwaighofer | 2013-07-11 | 1 | -0/+61 |
* | LoopVectorize: Math functions only read rounding mode | Arnold Schwaighofer | 2013-07-01 | 1 | -0/+32 |
* | LoopVectorize: Preserve debug location info | Arnold Schwaighofer | 2013-06-28 | 1 | -0/+92 |
* | LoopVectorize: Cache edge masks created during if-conversion | Arnold Schwaighofer | 2013-06-27 | 1 | -0/+243 |
* | LoopVectorize: Use vectorized loop invariant gep index anchored in loop | Arnold Schwaighofer | 2013-06-27 | 1 | -0/+36 |
* | Fix spelling. | Arnold Schwaighofer | 2013-06-27 | 1 | -1/+1 |
* | LoopVectorize: Don't store a reversed value in the vectorized value map | Arnold Schwaighofer | 2013-06-27 | 1 | -0/+55 |
* | Reapply 184685 after the SetVector iteration order fix. | Arnold Schwaighofer | 2013-06-24 | 3 | -2/+224 |
* | Revert "LoopVectorize: Use the dependence test utility class" | Arnold Schwaighofer | 2013-06-24 | 3 | -224/+2 |
* | LoopVectorize: Use the dependence test utility class | Arnold Schwaighofer | 2013-06-24 | 3 | -2/+224 |
* | Fix for a regression caused by the LoopVectorizer when | Pekka Jaaskelainen | 2013-06-17 | 1 | -0/+47 |
* | LoopVectorize: PHIs with only outside users should prevent vectorization | Arnold Schwaighofer | 2013-05-31 | 1 | -0/+41 |
* | Add support for llvm.vectorizer metadata | Paul Redmond | 2013-05-28 | 6 | -10/+86 |
* | LoopVectorize: LoopSimplify can't canonicalize loops with an indirectbr in it... | Benjamin Kramer | 2013-05-24 | 1 | -0/+11 |
* | LoopVectorize: Make Value pointers that could be RAUW'ed a VH | Arnold Schwaighofer | 2013-05-22 | 1 | -0/+50 |
* | LoopVectorize: Handle single edge PHIs | Arnold Schwaighofer | 2013-05-18 | 1 | -0/+22 |
* | LoopVectorize: Hoist conditional loads if possible | Arnold Schwaighofer | 2013-05-15 | 1 | -0/+69 |
* | LoopVectorize: Handle loops with multiple forward inductions | Arnold Schwaighofer | 2013-05-14 | 1 | -0/+30 |
* | LoopVectorize: Use the widest induction variable type | Arnold Schwaighofer | 2013-05-11 | 1 | -0/+69 |
* | Add an additional testcase for PR15882. | Nadav Rotem | 2013-05-10 | 1 | -0/+45 |
* | LoopVectorizer: Don't assert on the absence of induction variables | Arnold Schwaighofer | 2013-05-09 | 1 | -0/+34 |
* | LoopVectorizer: Improve reduction variable identification | Arnold Schwaighofer | 2013-05-07 | 1 | -0/+119 |
* | LoopVectorize: getConsecutiveVector must respect signed arithmetic | Arnold Schwaighofer | 2013-05-07 | 1 | -0/+79 |
* | LoopVectorize: Add support for floating point min/max reductions | Arnold Schwaighofer | 2013-05-05 | 1 | -0/+480 |
* | LoopVectorize: We don't need an identity element for min/max reductions | Arnold Schwaighofer | 2013-05-05 | 1 | -1/+4 |
* | LoopVectorizer: Add support for if-conversion of PHINodes with 3+ incoming va... | Nadav Rotem | 2013-05-03 | 1 | -0/+48 |
* | TBAA: remove !tbaa from testing cases if not used. | Manman Ren | 2013-05-02 | 3 | -98/+84 |
* | TBAA: remove !tbaa from testing cases if not used. | Manman Ren | 2013-04-30 | 13 | -90/+44 |
* | LoopVectorizer: Calculate the number of pointers to disambiguate at runtime ... | Nadav Rotem | 2013-04-26 | 1 | -0/+84 |
* | LoopVectorizer: No need to generate pointer disambiguation checks between rea... | Nadav Rotem | 2013-04-25 | 1 | -0/+36 |
* | LoopVectorize: Scalarize padded types | Arnold Schwaighofer | 2013-04-24 | 1 | -0/+29 |
* | LoopVectorizer: Bail out if we don't have datalayout we need it | Arnold Schwaighofer | 2013-04-24 | 3 | -1/+8 |
* | LoopVectorizer: Fix 15830. When scalarizing and unrolling stores make sure th... | Nadav Rotem | 2013-04-23 | 1 | -0/+36 |
* | Call the potentially costly isAnnotatedParallel() only once. | Pekka Jaaskelainen | 2013-04-23 | 1 | -1/+2 |
* | Refuse to (even try to) vectorize loops which have uniform writes, | Pekka Jaaskelainen | 2013-04-23 | 1 | -0/+58 |
* | LoopVectorizer: Recognize min/max reductions | Arnold Schwaighofer | 2013-04-18 | 1 | -0/+399 |
* | LoopVectorize: Use a set to avoid longer cycles in the reduction chain too. | Benjamin Kramer | 2013-04-18 | 1 | -0/+18 |
* | LoopVectorizer: integer division is not a reduction operation | Arnold Schwaighofer | 2013-04-12 | 1 | -0/+24 |
* | LoopVectorizer: Pass OperandValueKind information to the cost model | Arnold Schwaighofer | 2013-04-04 | 1 | -0/+28 |
* | X86TTI: Add accurate costs for itofp operations, based on the actual instruct... | Benjamin Kramer | 2013-04-01 | 1 | -5/+4 |
* | LoopVectorizer: Insert some white space to make test case more readable | Arnold Schwaighofer | 2013-03-14 | 1 | -6/+10 |
* | Add missing asserts flag to test - it uses debug flags | Arnold Schwaighofer | 2013-03-14 | 1 | -1/+1 |
* | LoopVectorize: Invert case when we use a vector cmp value to query select cost | Arnold Schwaighofer | 2013-03-14 | 1 | -0/+62 |
* | Test case hygiene. | Benjamin Kramer | 2013-03-09 | 1 | -20/+20 |
* | LoopVectorizer: Ignore dbg.value instructions | Arnold Schwaighofer | 2013-03-09 | 1 | -0/+70 |
* | Force cpu in test. | Benjamin Kramer | 2013-03-08 | 1 | -1/+1 |
* | Insert the reduction start value into the first bypass block to preserve domi... | Benjamin Kramer | 2013-03-08 | 1 | -0/+35 |
* | X86 cost model: Adjust cost for custom lowered vector multiplies | Arnold Schwaighofer | 2013-03-02 | 1 | -2/+2 |
* | PR14448 - prevent the loop vectorizer from vectorizing the same loop twice. | Nadav Rotem | 2013-03-02 | 1 | -0/+75 |
* | LoopVectorize: Don't hang forever if a PHI only has skipped PHI uses. | Benjamin Kramer | 2013-03-01 | 1 | -0/+29 |