2013-01-30 |
Evan Cheng | Restrict sin/cos optimization to 64-bit only for now... |
blob | commitdiff | raw |
2013-01-29 |
Evan Cheng | Remove dead code. |
blob | commitdiff | raw | diff to current |
2013-01-29 |
Evan Cheng | Teach SDISel to combine fsin / fcos into a fsincos... |
blob | commitdiff | raw | diff to current |
2013-01-28 |
Craig Topper | Fix inconsistent usage of PALIGN and PALIGNR when refer... |
blob | commitdiff | raw | diff to current |
2013-01-26 |
Benjamin Kramer | X86: Decode PALIGN operands so I don't have to do it... |
blob | commitdiff | raw | diff to current |
2013-01-26 |
Benjamin Kramer | X86: Do splat promotion later, so the optimizer can... |
blob | commitdiff | raw | diff to current |
2013-01-25 |
Eli Bendersky | In this patch, we teach X86_64TargetMachine that it... |
blob | commitdiff | raw | diff to current |
2013-01-22 |
Michael Liao | Fix an issue of pseudo atomic instruction DAG schedule |
blob | commitdiff | raw | diff to current |
2013-01-22 |
Tim Northover | Make APFloat constructor require explicit semantics. |
blob | commitdiff | raw | diff to current |
2013-01-21 |
Craig Topper | Use <0 checks in place of ==-1 because it results in... |
blob | commitdiff | raw | diff to current |
2013-01-21 |
Craig Topper | Use MVT instead of EVT in LowerVECTOR_SHUFFLEtoBlend. |
blob | commitdiff | raw | diff to current |
2013-01-21 |
Craig Topper | Remove trailing whitespace. |
blob | commitdiff | raw | diff to current |
2013-01-21 |
Craig Topper | Fix some 80 column violations. |
blob | commitdiff | raw | diff to current |
2013-01-21 |
Craig Topper | Make helper method static. |
blob | commitdiff | raw | diff to current |
2013-01-20 |
Craig Topper | Convert more EVT's to MVT's in the lowering methods. |
blob | commitdiff | raw | diff to current |
2013-01-20 |
Craig Topper | Capitalize lowerTRUNCATE so that it matches the other... |
blob | commitdiff | raw | diff to current |
2013-01-20 |
Craig Topper | Make LowerVSETCC a static function and use MVT instead... |
blob | commitdiff | raw | diff to current |
2013-01-20 |
Nadav Rotem | Revert 172708. |
blob | commitdiff | raw | diff to current |
2013-01-20 |
Craig Topper | Make some helper methods static. |
blob | commitdiff | raw | diff to current |
2013-01-20 |
Craig Topper | Remove DebugLoc argument from static function. It can... |
blob | commitdiff | raw | diff to current |
2013-01-20 |
Craig Topper | Use MVT instead of EVT in more instruction lowering... |
blob | commitdiff | raw | diff to current |
2013-01-19 |
Craig Topper | Use MVT instead of EVT in more of the shuffle lowering... |
blob | commitdiff | raw | diff to current |
2013-01-19 |
Craig Topper | Capitalize LowerVectorIntExtend to be consistent with... |
blob | commitdiff | raw | diff to current |
2013-01-19 |
Nadav Rotem | On Sandybridge split unaligned 256bit stores into two... |
blob | commitdiff | raw | diff to current |
2013-01-19 |
Craig Topper | Use MVT instead of EVT when computing shuffle immediate... |
blob | commitdiff | raw | diff to current |
2013-01-18 |
Nadav Rotem | On Sandybridge loading unaligned 256bits using two... |
blob | commitdiff | raw | diff to current |
2013-01-18 |
Craig Topper | Calculate vector element size more directly for VINSERT... |
blob | commitdiff | raw | diff to current |
2013-01-18 |
Craig Topper | Minor formatting fix. No functional change. |
blob | commitdiff | raw | diff to current |
2013-01-18 |
Craig Topper | Spelling fix: extened->extended. Trailing whitespace... |
blob | commitdiff | raw | diff to current |
2013-01-18 |
Craig Topper | Make more use of is128BitVector/is256BitVector in place... |
blob | commitdiff | raw | diff to current |
2013-01-17 |
Elena Demikhovsky | Optimization for the following SIGN_EXTEND pairs: |
blob | commitdiff | raw | diff to current |
2013-01-09 |
Nadav Rotem | Efficient lowering of vector sdiv when the divisor... |
blob | commitdiff | raw | diff to current |
2013-01-07 |
Chandler Carruth | Switch TargetTransformInfo from an immutable analysis... |
blob | commitdiff | raw | diff to current |
2013-01-04 |
Nadav Rotem | LoopVectorizer: |
blob | commitdiff | raw | diff to current |
2013-01-04 |
Nadav Rotem | Revert revision: 171467. This transformation is incorre... |
blob | commitdiff | raw | diff to current |
2013-01-03 |
Elena Demikhovsky | Simplified TRUNCATE operation that comes after SETCC... |
blob | commitdiff | raw | diff to current |
2013-01-03 |
Hal Finkel | Add a subtype parameter to VTTI::getShuffleCost |
blob | commitdiff | raw | diff to current |
2013-01-02 |
Nadav Rotem | AVX: Fix a bug in WidenMaskArithmetic. |
blob | commitdiff | raw | diff to current |
2013-01-02 |
Chandler Carruth | Move all of the header files which are involved in... |
blob | commitdiff | raw | diff to current |
2012-12-30 |
Bill Wendling | Use the predicate methods off of AttributeSet instead... |
blob | commitdiff | raw | diff to current |
2012-12-30 |
Bill Wendling | Remove the Function::getFnAttributes method in favor... |
blob | commitdiff | raw | diff to current |
2012-12-29 |
Craig Topper | Remove intrinsic specific instructions for (V)SQRTPS... |
blob | commitdiff | raw | diff to current |
2012-12-29 |
Craig Topper | Merge similar functionality using a nested switch. |
blob | commitdiff | raw | diff to current |
2012-12-29 |
Craig Topper | Remove intrinsic specific instructions for SSE/SSE2... |
blob | commitdiff | raw | diff to current |
2012-12-29 |
Jakub Staszak | Simplify code, no functionality change. |
blob | commitdiff | raw | diff to current |
2012-12-28 |
Nadav Rotem | CostModel: initial checkin for code that estimates... |
blob | commitdiff | raw | diff to current |
2012-12-28 |
Nadav Rotem | wrap 80-col lines. |
blob | commitdiff | raw | diff to current |
2012-12-28 |
Nadav Rotem | AVX: Move the ZEXT/ANYEXT DAGCo optimizations to the... |
blob | commitdiff | raw | diff to current |
2012-12-27 |
Nadav Rotem | Reverse the 'if' condition and reduce the indentation. |
blob | commitdiff | raw | diff to current |
2012-12-27 |
Nadav Rotem | AVX/AVX2: Move the SEXT lowering code from a target... |
blob | commitdiff | raw | diff to current |
2012-12-27 |
Nadav Rotem | On AVX/AVX2 the type v8i1 is legalized to v8i16, which... |
blob | commitdiff | raw | diff to current |
2012-12-27 |
Nadav Rotem | AVX/AVX2: Move the code that lowers vector-trunc from... |
blob | commitdiff | raw | diff to current |
2012-12-26 |
Nadav Rotem | Reformat the docs. |
blob | commitdiff | raw | diff to current |
2012-12-25 |
Benjamin Kramer | X86: Shave off one shuffle from the pcmpeqq sequence... |
blob | commitdiff | raw | diff to current |
2012-12-25 |
Benjamin Kramer | X86: Custom lower <2 x i64> eq and ne when SSE41 is... |
blob | commitdiff | raw | diff to current |
2012-12-24 |
Nick Lewycky | Quiet gcc's -Wparenthesis warning. No functionality... |
blob | commitdiff | raw | diff to current |
2012-12-23 |
Nadav Rotem | whitespace |
blob | commitdiff | raw | diff to current |
2012-12-23 |
Nadav Rotem | Loop Vectorizer: Update the cost model of scatter/gathe... |
blob | commitdiff | raw | diff to current |
2012-12-22 |
Benjamin Kramer | X86: Turn mul of <4 x i32> into pmuludq when no SSE4... |
blob | commitdiff | raw | diff to current |
2012-12-22 |
Benjamin Kramer | X86: Emit vector sext as shuffle + sra if vpmovsx is... |
blob | commitdiff | raw | diff to current |
2012-12-21 |
Benjamin Kramer | X86: Match pmin/pmax as a target specific dag combine... |
blob | commitdiff | raw | diff to current |
2012-12-21 |
Benjamin Kramer | X86: Match the SSE/AVX min/max vector ops using a custo... |
blob | commitdiff | raw | diff to current |
2012-12-21 |
Nadav Rotem | Improve the X86 cost model for loads and stores. |
blob | commitdiff | raw | diff to current |
2012-12-19 |
Patrik Hagglund | Change TargetLowering::getTypeForExtArgOrReturn to... |
blob | commitdiff | raw | diff to current |
2012-12-19 |
Patrik Hagglund | Change TargetLowering::findRepresentativeClass to take... |
blob | commitdiff | raw | diff to current |
2012-12-19 |
NAKAMURA Takumi | X86ISelLowering.cpp: Fix warnings. [-Wlogical-op-parent... |
blob | commitdiff | raw | diff to current |
2012-12-19 |
Elena Demikhovsky | Optimized load + SIGN_EXTEND patterns in the X86 backend. |
blob | commitdiff | raw | diff to current |
2012-12-19 |
Bill Wendling | Rename the 'Attributes' class to 'Attribute'. It's... |
blob | commitdiff | raw | diff to current |
2012-12-18 |
Jakub Staszak | Reverse order of checking SSE level when calculating... |
blob | commitdiff | raw | diff to current |
2012-12-17 |
Craig Topper | Simplify BMI ANDN matching to use patterns instead... |
blob | commitdiff | raw | diff to current |
2012-12-15 |
Benjamin Kramer | X86: Add a couple of target-specific dag combines that... |
blob | commitdiff | raw | diff to current |
2012-12-14 |
Nadav Rotem | TypeLegalizer: Do not generate target specific nodes... |
blob | commitdiff | raw | diff to current |
2012-12-12 |
Evan Cheng | Sorry about the churn. One more change to getOptimalMem... |
blob | commitdiff | raw | diff to current |
2012-12-12 |
Evan Cheng | - Rename isLegalMemOpType to isSafeMemOpType. "Legal... |
blob | commitdiff | raw | diff to current |
2012-12-12 |
Evan Cheng | Avoid using lossy load / stores for memcpy / memset... |
blob | commitdiff | raw | diff to current |
2012-12-11 |
Patrik Hagglund | Revert EVT->MVT changes, r169836-169851, due to buildbo... |
blob | commitdiff | raw | diff to current |
2012-12-11 |
Patrik Hagglund | Change TargetLowering::getTypeForExtArgOrReturn to... |
blob | commitdiff | raw | diff to current |
2012-12-11 |
Patrik Hagglund | Change TargetLowering::findRepresentativeClass to take... |
blob | commitdiff | raw | diff to current |
2012-12-10 |
Evan Cheng | Some enhancements for memcpy / memset inline expansion. |
blob | commitdiff | raw | diff to current |
2012-12-09 |
Shuxin Yang | - Re-enable population count loop idiom recognization |
blob | commitdiff | raw | diff to current |
2012-12-08 |
Chandler Carruth | Revert the patches adding a popcount loop idiom recogni... |
blob | commitdiff | raw | diff to current |
2012-12-07 |
Bill Wendling | s/AttrListPtr/AttributeSet/g to better label what this... |
blob | commitdiff | raw | diff to current |
2012-12-07 |
Nadav Rotem | When we use the BLEND instruction that uses the MSB... |
blob | commitdiff | raw | diff to current |
2012-12-07 |
Nadav Rotem | X86: Prefer using VPSHUFD over VPERMIL because it has... |
blob | commitdiff | raw | diff to current |
2012-12-06 |
Evan Cheng | Replace r169459 with something safer. Rather than havin... |
blob | commitdiff | raw | diff to current |
2012-12-06 |
Jakub Staszak | Remove unneeded function, since PR8156 was fixed over... |
blob | commitdiff | raw | diff to current |
2012-12-06 |
Jakub Staszak | Simplify code. |
blob | commitdiff | raw | diff to current |
2012-12-06 |
Evan Cheng | Let targets provide hooks that compute known zero and... |
blob | commitdiff | raw | diff to current |
2012-12-05 |
Elena Demikhovsky | Simplified BLEND pattern matching for shuffles. |
blob | commitdiff | raw | diff to current |
2012-12-05 |
Evan Cheng | Add x86 isel lowering logic to form bit test with inver... |
blob | commitdiff | raw | diff to current |
2012-12-03 |
Chandler Carruth | Use the new script to sort the includes of every file... |
blob | commitdiff | raw | diff to current |
2012-11-29 |
Shuxin Yang | rdar://12100355 (part 1) |
blob | commitdiff | raw | diff to current |
2012-11-29 |
Elena Demikhovsky | I changed hasAVX() to hasFp256() and hasAVX2() to hasIn... |
blob | commitdiff | raw | diff to current |
2012-11-26 |
Jakub Staszak | Normalize splat 256bit vectors with 8 elements. |
blob | commitdiff | raw | diff to current |
2012-11-21 |
Craig Topper | Mark ISD::FMA as Legal instead of custom for x86 with... |
blob | commitdiff | raw | diff to current |
2012-11-16 |
Duncan Sands | Add the Erlang/HiPE calling convention, patch by Yianni... |
blob | commitdiff | raw | diff to current |
2012-11-16 |
Craig Topper | Use roundps/pd for llvm.ceil, llvm.trunc, llvm.rint... |
blob | commitdiff | raw | diff to current |
2012-11-15 |
Craig Topper | Add llvm.ceil, llvm.trunc, llvm.rint, llvm.nearbyint... |
blob | commitdiff | raw | diff to current |
2012-11-14 |
Benjamin Kramer | X86: Enable SSE memory intrinsics even when stack align... |
blob | commitdiff | raw | diff to current |
2012-11-14 |
Craig Topper | Factor out an overly replicated typecast. No functional... |
blob | commitdiff | raw | diff to current |
next |