| Commit message (Expand) | Author | Age | Files | Lines |
* | Add patterns to generate copies for extract_subvector instead of | Bruno Cardoso Lopes | 2011-07-28 | 1 | -3/+3 |
* | Add a few patterns to match allzeros without having to use the fp unit. | Bruno Cardoso Lopes | 2011-07-28 | 1 | -1/+1 |
* | Add SINT_TO_FP and FP_TO_SINT support for v8i32 types. Also move | Bruno Cardoso Lopes | 2011-07-28 | 1 | -0/+14 |
* | The vpermilps and vpermilpd have different behaviour regarding the | Bruno Cardoso Lopes | 2011-07-27 | 1 | -0/+16 |
* | It is quiet possible that inlined function body is split into multiple chunks... | Devang Patel | 2011-07-27 | 1 | -0/+140 |
* | Eliminate copies of undefined values during coalescing. | Jakob Stoklund Olesen | 2011-07-26 | 1 | -0/+27 |
* | Update test. | Benjamin Kramer | 2011-07-26 | 1 | -1/+1 |
* | Add a neat little two's complement hack for x86. | Benjamin Kramer | 2011-07-26 | 1 | -0/+11 |
* | Recognize unpckh* masks and match 256-bit versions. The new versions are | Bruno Cardoso Lopes | 2011-07-26 | 6 | -85/+58 |
* | Prevent x86-specific DAGCombine from creating nodes with illegal type (which ... | Eli Friedman | 2011-07-26 | 1 | -2/+1 |
* | XFAIL this test while I investigate it; it's failing for an unexpected reason. | Eli Friedman | 2011-07-26 | 1 | -0/+1 |
* | Add obvious missing case to switch. PR10497. | Eli Friedman | 2011-07-26 | 1 | -0/+12 |
* | Add 256-bit isel for movsldup/movshdup | Bruno Cardoso Lopes | 2011-07-26 | 1 | -0/+34 |
* | Codegen allonesvector better while using AVX: vpcmpeqd + vinsertf128 | Bruno Cardoso Lopes | 2011-07-25 | 1 | -0/+12 |
* | - Handle special scalar_to_vector case: splats. Using a native 128-bit | Bruno Cardoso Lopes | 2011-07-25 | 2 | -4/+23 |
* | Attempt to fix test failure reported on llvm-commits. | Eli Friedman | 2011-07-25 | 1 | -1/+1 |
* | Make sure this DAGCombine actually returns an UNDEF of the correct type; PR10... | Eli Friedman | 2011-07-25 | 1 | -2/+18 |
* | Get rid of an incorrect optimization for shuffles with PALIGNR and simplify i... | Eli Friedman | 2011-07-25 | 1 | -7/+22 |
* | Correctly handle <undef> tied uses when rewriting after a split. | Jakob Stoklund Olesen | 2011-07-24 | 1 | -0/+25 |
* | Fix test check! | Bruno Cardoso Lopes | 2011-07-22 | 1 | -1/+1 |
* | Fix PR10422 by adding the necessary AVX UCOMISD memory versions to | Bruno Cardoso Lopes | 2011-07-22 | 1 | -0/+28 |
* | Turn shuffles into unpacks for VT == MVT::v2i64 and MVT::v2f64 | Rafael Espindola | 2011-07-22 | 1 | -0/+26 |
* | -Inspected a AVX code block added by someone in early Feb. This was never used | Bruno Cardoso Lopes | 2011-07-22 | 2 | -2/+2 |
* | Although we already support this, add testcases for consistency | Bruno Cardoso Lopes | 2011-07-22 | 1 | -0/+21 |
* | Add a DAGCombine for transforming 128->256 casts into a simple | Bruno Cardoso Lopes | 2011-07-22 | 1 | -0/+26 |
* | - Register v16i16 as valid VR256 register class | Bruno Cardoso Lopes | 2011-07-21 | 1 | -1/+12 |
* | Add support for 256-bit versions of VPERMIL instruction. This is a new | Bruno Cardoso Lopes | 2011-07-21 | 1 | -0/+16 |
* | While emitting constant value, look through derived type and use underlying b... | Devang Patel | 2011-07-20 | 1 | -0/+61 |
* | PR10421: Fix a straightforward bug in the widening logic for CONCAT_VECTORS. | Eli Friedman | 2011-07-20 | 1 | -0/+8 |
* | New pointer rotate test. | Eric Christopher | 2011-07-20 | 1 | -0/+11 |
* | Fix an obvious typo that's preventing x86 (32-bit) from using .literal16. | Evan Cheng | 2011-07-19 | 1 | -1/+1 |
* | Revert r135423. | Devang Patel | 2011-07-19 | 1 | -120/+0 |
* | During bottom up fast-isel, instructions emitted to materalize registers are ... | Devang Patel | 2011-07-18 | 1 | -0/+120 |
* | Add AVX 128-bit sqrt versions | Bruno Cardoso Lopes | 2011-07-18 | 1 | -0/+18 |
* | Delete empty unused file. | Nick Lewycky | 2011-07-18 | 1 | -0/+0 |
* | Add AVX 128-bit patterns for sint_to_fp | Bruno Cardoso Lopes | 2011-07-16 | 1 | -1/+33 |
* | Fix a couple of things: | Bruno Cardoso Lopes | 2011-07-15 | 1 | -0/+0 |
* | Check register class matching instead of width of type matching | Eric Christopher | 2011-07-14 | 1 | -0/+13 |
* | Add 256-bit load/store recognition and matching in several places. | Bruno Cardoso Lopes | 2011-07-14 | 1 | -0/+24 |
* | Don't emit a bit test if there is only one case the test can yield false. A s... | Benjamin Kramer | 2011-07-14 | 1 | -0/+20 |
* | We already support 256-bit packed ADD, SUB, DIV, MUL. Add testcases. | Bruno Cardoso Lopes | 2011-07-13 | 1 | -0/+116 |
* | Make X86ISD::ANDNP more general and Codegen 256-bit VANDNP. A more | Bruno Cardoso Lopes | 2011-07-13 | 1 | -0/+45 |
* | Make sure we don't combine a large displacement and a frame index in the same... | Eli Friedman | 2011-07-13 | 1 | -0/+20 |
* | AVX Codegen support for 256-bit versions of vandps, vandpd, vorps, vorpd, vxo... | Bruno Cardoso Lopes | 2011-07-13 | 1 | -0/+116 |
* | Land the long talked about "type system rewrite" patch. This | Chris Lattner | 2011-07-09 | 4 | -16/+16 |
* | more tests not making the jump into the brave new world. | Chris Lattner | 2011-07-09 | 3 | -348/+0 |
* | test/CodeGen/X86/vector.ll: Tweak temporary output to appease Win32 hosts. | NAKAMURA Takumi | 2011-07-09 | 1 | -1/+1 |
* | Hoist spills within a basic block. | Jakob Stoklund Olesen | 2011-07-09 | 1 | -0/+35 |
* | Fix broken x86_64 tests which specify non-64-bit cpu's. | Evan Cheng | 2011-07-08 | 3 | -4/+4 |
* | Default 64-bit target features and SSE2 on when a triple specifies x86-64. C... | Eli Friedman | 2011-07-08 | 1 | -1/+1 |