- c741ea4 Do not endian swap split vector loads. This fixes UnitTests/Vector/sumarray-dbl on PPC. by Chris Lattner · 19 years ago
- d9731af Do not endian swap the operands to a store if the operands came from a vector. by Chris Lattner · 19 years ago
- 29cd7db Remove dead *extloads. This allows us to codegen vector.ll:test_extract_elt to: by Chris Lattner · 19 years ago
- e4b9539 Delete dead loads in the dag. This allows us to compile by Chris Lattner · 19 years ago
- bf93414 Make this test harder, always check generic x86 and SSE-enabled x86. by Chris Lattner · 19 years ago
- 1597221 Implement PromoteOp for VEXTRACT_VECTOR_ELT. Thsi fixes by Chris Lattner · 19 years ago
- 2ae2e98 Scalarized vector stores need not be legal, e.g. if the vector element type by Chris Lattner · 19 years ago
- 4c5701d Fix build breakage. by Jeff Cohen · 19 years ago
- 33497cc note to self: *save* file, then check it in by Chris Lattner · 19 years ago
- 4468c22 Implement an item from the readme, folding vcmp/vcmp. instructions with by Chris Lattner · 19 years ago
- 2a8fd4a new testcase by Chris Lattner · 19 years ago
- 9492151 compactify some more instruction definitions by Chris Lattner · 19 years ago
- 5f7b019 Compactify comparisons. by Chris Lattner · 19 years ago
- ffa4fdc Allow bits init values to be used in patterns, turn them into ints. by Chris Lattner · 19 years ago
- a17b155 Lower vector compares to VCMP nodes, just like we lower vector comparison by Chris Lattner · 19 years ago
- 8f5d316 These are done by Chris Lattner · 19 years ago
- f124b46 Final bugfix for PR724. GCC won't inline varargs functions, so use one to by Chris Lattner · 19 years ago
- 536a9d5 Add a new method to verify intrinsic function prototypes. by Chris Lattner · 19 years ago
- c4d9b24 When emitting code for the verifier, instead of emitting each case statement by Chris Lattner · 19 years ago
- 7e02151 Make sure to pass enough values to phi nodes when we are dealing with by Chris Lattner · 19 years ago
- 70c2a61 Significantly improve handling of vectors that are live across basic blocks, by Chris Lattner · 19 years ago
- a6c9de4 Was returning the wrong type. by Chris Lattner · 19 years ago
- 36f4b0d Mark INSERT_VECTOR_ELT as expand by Chris Lattner · 19 years ago
- 33e85ca Expand all INSERT_VECTOR_ELT (obviously bad) for now. by Evan Cheng · 19 years ago
- eb0b461 Expand INSERT_VECTOR_ELT to store vec, sp; store elt, sp+k; vec = load sp; by Evan Cheng · 19 years ago
- 79227e2 Modify the TargetLowering::getPackedTypeBreakdown method to also return the by Chris Lattner · 19 years ago
- fb47a9b Typo by Evan Cheng · 19 years ago
- ef698ca Ok for vector_shuffle mask to contain undef elements. by Evan Cheng · 19 years ago
- dc87929 Implement TargetLowering::getPackedTypeBreakdown by Chris Lattner · 19 years ago
- 00cc494 Add a method useful for decimating vectors. by Chris Lattner · 19 years ago
- 8768bf6 Add the rest of the vmul instructions and the vmulsum* instructions. by Chris Lattner · 19 years ago
- e467829 fix incorrect prototypes by Chris Lattner · 19 years ago
- 3c4f4e9f Use a new tblgen feature to significantly shrinkify instruction definitions that by Chris Lattner · 19 years ago
- 30a6aba Add a bunch of new instructions for intrinsics. by Chris Lattner · 19 years ago
- 8177969 regenerate by Chris Lattner · 19 years ago
- 8c06318 Implement Regression/TableGen/DagDefSubst.ll by Chris Lattner · 19 years ago
- 824b46f Dag operator should be able to be template params. by Chris Lattner · 19 years ago
- df084ff Fix Transforms/InstCombine/2006-03-30-ExtractElement.ll by Chris Lattner · 19 years ago
- fc542a6 new testcase that crashes instcombine by Chris Lattner · 19 years ago
- 7d9061e Make sure all possible shuffles are matched. by Evan Cheng · 19 years ago
- 99e746f Add vector multiply, multiply sum, pack, unpack, and lvsl/lvsr intrinsics. by Chris Lattner · 19 years ago
- 1b32f22 More logical ops patterns by Evan Cheng · 19 years ago
- 0876aa5 Add support for _mm_cmp{cc}_ss and _mm_cmp{cc}_ps intrinsics by Evan Cheng · 19 years ago
- c5fb2b1 Add 128-bit pmovmskb intrinsic support. by Evan Cheng · 19 years ago
- 591f740 Change SSE pack operation definitions to fit what the intrinsics expected. by Evan Cheng · 19 years ago
- 4e27b8e Add SSE2 integer pack with saturation intrinsics. by Evan Cheng · 19 years ago
- 506d3df - Added some SSE2 128-bit packed integer ops. by Evan Cheng · 19 years ago
- 75658b9 Add a pshufhw test case. by Evan Cheng · 19 years ago
- 691c923 Need to special case splat after all. Make the second operand of splat by Evan Cheng · 19 years ago
- e4086e5 Use unpcklpd for v2f64 splat. by Evan Cheng · 19 years ago
- 5aa97b2 Floating point logical operation patterns should match bit_convert. Or else by Evan Cheng · 19 years ago
- 074fd57 Add more SSE intrinsics by Evan Cheng · 19 years ago
- 475aecf - More shuffle related bug fixes. by Evan Cheng · 19 years ago
- c999c74 Another entry about shuffles. by Evan Cheng · 19 years ago
- 4f56338 - Only use pshufd for v4i32 vector shuffles. - Other shuffle related fixes. by Evan Cheng · 19 years ago
- f3a6272 add a note by Chris Lattner · 19 years ago
- 246ce96 new testcase by Chris Lattner · 19 years ago
- 2bbd810 Bug fixes: handle constantexpr insert/extract element operations by Chris Lattner · 19 years ago
- c46349d Added aliases to scalar SSE instructions, e.g. addss, to match x86 intrinsics. by Evan Cheng · 19 years ago
- 36b27f3 Fixing buggy code. by Evan Cheng · 19 years ago
- f065a6f Don't sort the names before outputing the intrinsic name table. It causes a by Evan Cheng · 19 years ago
- cef896e When building a VVECTOR_SHUFFLE node from extract_element operations, make by Chris Lattner · 19 years ago
- 66445d3 Canonicalize VECTOR_SHUFFLE(X, X, Y) -> VECTOR_SHUFFLE(X,undef,Y') by Chris Lattner · 19 years ago
- bfc1a79 new testcase by Chris Lattner · 19 years ago
- d7648c8 Turn a series of extract_element's feeding a build_vector into a by Chris Lattner · 19 years ago
- 5b2316e Teach Legalize how to pack VVECTOR_SHUFFLE nodes into VECTOR_SHUFFLE nodes. by Chris Lattner · 19 years ago
- 354cde9 new node by Chris Lattner · 19 years ago
- eda6dfd Add a new node by Chris Lattner · 19 years ago
- 4fbdd59 Don't crash on X^X if X is a vector. Instead, produce a vector of zeros. by Chris Lattner · 19 years ago
- f35b297 Add an assertion by Chris Lattner · 19 years ago
- d732a29 add a note by Chris Lattner · 19 years ago
- b3e7be2 Refactor address attributes. Add base register to frame info. by Jim Laskey · 19 years ago
- a997918 Expose base register for DwarfWriter. Refactor code accordingly. by Jim Laskey · 19 years ago
- bf7637d More bulletproofing of llvm.dbg.declare. by Jim Laskey · 19 years ago
- 2d2a613 Added missing paren on behalf of Ramana Radhakrishnan. by Jim Laskey · 19 years ago
- ed4ca7f Missed X86::isUNPCKHMask by Evan Cheng · 19 years ago
- 9bbfd4f movlps and movlpd should be modeled as two address code. by Evan Cheng · 19 years ago
- 6a6d354 Update by Evan Cheng · 19 years ago
- be296ac Typo by Evan Cheng · 19 years ago
- 2064a2b * Prefer using operation of matching types. e.g unpcklpd rather than movlhps. by Evan Cheng · 19 years ago
- ec1ab44 Use movhpd is even better than movlhps. by Evan Cheng · 19 years ago
- 816cee2 Fix a couple typos by Nate Begeman · 19 years ago
- 98e70cc Add a few more altivec intrinsics by Nate Begeman · 19 years ago
- 5659b41 Keep Visual Studio informed. by Jeff Cohen · 19 years ago
- 9bade4d These don't directly map to gcc intrinsics any more. by Chris Lattner · 19 years ago
- 4c4a2e2 Added a couple of entries about movhps and movlhps. by Evan Cheng · 19 years ago
- 000e4dd All unpack cases are now being handled. by Evan Cheng · 19 years ago
- 4fcb922 - Clean up / consoladate various shuffle masks. by Evan Cheng · 19 years ago
- ecc219b implement a bunch more intrinsics. by Chris Lattner · 19 years ago
- 858b35e Add some more intrinsics: rotates, fp rounds, and random other fp instructions. by Chris Lattner · 19 years ago
- 78b9ec3 getVectorTyppe(MVT::i64, 2) ==> MVT::v2i64. by Evan Cheng · 19 years ago
- 572f974 Add lvxl by Chris Lattner · 19 years ago
- 7f20b13 Use normal lvx for scalar_to_vector instead of lve*x. They do the exact by Chris Lattner · 19 years ago
- 8156a0d Regression test for the handling of nulls as arguments to debug intrinsics. by Jim Laskey · 19 years ago
- aaa80eb More bulletproofing of DebugInfoDesc verify. by Jim Laskey · 19 years ago
- b026e70 Tblgen doesn't like multiple SDNode<> definitions that map to the sameenum value. Split them into separate enums. by Chris Lattner · 19 years ago
- 48b61a7 Tblgen doesn't like multiple SDNode<> definitions that map to the sameenum value. Split them into separate enums. by Chris Lattner · 19 years ago
- 0038e59 Model unpack lower and interleave as vector_shuffle so we can lower the by Evan Cheng · 19 years ago
- ef8ef91 Tblgen doesn't like multiple SDNode<> definitions that map to the same by Chris Lattner · 19 years ago
- c9e0b14 Reenable pointer intrinsics. by Chris Lattner · 19 years ago