aboutsummaryrefslogtreecommitdiff
path: root/llvm/lib/CodeGen/SelectionDAG
AgeCommit message (Expand)AuthorFilesLines
30 hours[SelectionDAGISel] Avoid unnecessary MatchScope copy. NFC (#178957)Craig Topper1-2/+1
33 hours[SelectionDAG] Handle undef at any position in isConstantSequence (#176671)Philip Ginsbach-Chen1-14/+47
34 hours[SelectionDAGISel] Use size_t for MatcherIndex. NFC (#178828)Craig Topper1-24/+24
35 hours[DAG] Reland: Enable bitcast STLF for Constant/Undef (#178890)陈子昂1-3/+26
40 hoursRevert "[DAG] Enable bitcast STLF for Constant/Undef" (#178872)Alex Bradbury1-26/+3
43 hours[DAG] Enable bitcast STLF for Constant/Undef (#172523)陈子昂1-3/+26
2 daysSelectionDAG: Add -print-sdnode-addrs flag.Peter Collingbourne1-0/+6
2 days[RISCV] Support ISD::CLMUL/CLMULH for i64 scalable vectors with Zvbc. (#178340)Craig Topper1-2/+6
2 days[AArch64] Optimize memset to use NEON DUP instruction for more sizes (#166030)Osama Abdelkader2-10/+30
2 days[CodeGen] Fix MachineMemOperand Size of MaskedLoad (#156398)Wei Xiao1-1/+1
3 days[SelectionDAG] Add CTLS to FoldConstantArithmetic and optimize i1 CTLS to 0. ...Craig Topper1-0/+9
3 days[LLVM][DAGCombiner] Look through freeze when combining extensions of loads (#...David Sherwood4-20/+49
3 days[CodeGen] Only use actual alloca alignment (#178361)Jameson Nash1-1/+1
3 days[NFC][CodeGen] Use getAllocationSize instead of manual size computation (#178...Jameson Nash1-4/+4
3 days[DAG] SDPatternMatch - allow m_BinOp / m_c_BinOp to take an optional SDNodeFl...Simon Pilgrim1-12/+8
3 days[DAG] SimplifyDemandedBits - ICMP_SLT(X,0) - only sign mask of X is required ...Anikesh Parashar1-16/+28
4 days[LegalizeIntegerTypes] Add `PromoteIntOp_ANY_EXTEND_VECTOR_INREG` (#178144)Abhishek Kaushik2-0/+14
4 days[perf] Replace copy-assign by move-assign in llvm/lib/CodeGen/* (#178172)serge-sans-paille1-1/+1
5 daysSupport EXTRACT_SUBVECTOR in computeKnownBits for scalable vectors (#177163)Sander de Smalen1-6/+11
5 daysDAG: Remove TypePromoteFloat (#177427)Matt Arsenault6-734/+0
6 days[SDAG] Remove non-canonical fabs libcall handling (#177967)Nikita Popov1-7/+0
6 daysTargetLowering: Allow FMINNUM/FMAXNUM to lower to FMINIMUM/FMAXIMUM even with...valadaptive1-8/+4
6 daysDAG: Use poison instead of undef in some vector combines (#177612)Matt Arsenault1-48/+48
7 days[DAG] Extend MinMax matchers to detect flippable sign (#177504)Stefan Weigl-Bosker1-8/+8
8 daysDAG: Use correct shift type for big endian store forwarding case (#177752)Matt Arsenault1-2/+3
8 days[KnownBits][SelectionDAG] Add KnownBits::clmul. Support trailing bits. NFC (...Craig Topper1-5/+1
8 daysDAG: Use poison for unused shuffle operands in legalizer (#177578)Matt Arsenault3-9/+10
9 daysDAG: Use poison in more vector legalization contexts (#177580)Matt Arsenault1-30/+31
9 days[DAG] Add basic folds for CLMUL nodes (#176961)Vishruth Thimmaiah1-0/+27
9 days[NFC][MI] Tidy Up RegState enum use (2/2) (#177090)Sam Elliott2-2/+3
9 days[SelectionDAG] Add very basic computeKnownBits support for ISD::CLMUL. (#177445)Craig Topper1-0/+10
9 days[DAG] expandIntMINMAX - use getOppositeSignednessMinMaxOpcode helper to flip ...Simon Pilgrim1-17/+1
10 days[SDAG] (abd? (?ext x), c) -> (zext (abd? x, c)) (#176366)DaKnig1-0/+28
10 days[AArch64] Optimize memcpy for non-power of two sizes (#168890)Cheng Lingfei1-10/+10
10 daysFix EXTEND_VECTOR_INREG widening when input > result size (#177095)nataliakokoromyti1-2/+25
10 days[LLVM][CodeGen] Remove pass initialization calls from pass constructors (#173...Rahul Joshi1-2/+0
11 days[AArch64] Fold sext-in-reg for predicate -> fixed-length conversions. (#176883)Sander de Smalen1-5/+11
11 days[IR] Allow non-constant offsets in @llvm.vector.splice.{left,right} (#174693)Luke Lau2-33/+27
11 daysIR: Remove llvm.convert.to.fp16 and llvm.convert.from.fp16 intrinsics (#174484)Matt Arsenault1-13/+0
12 days[DAG] foldAddToAvg - add patterns to form avgceil(A, B) from ((A >> 1) + (B >...Simon Pilgrim1-0/+16
12 daysDAG: Get libcall info from LibcallLowering in more places (#176836)Matt Arsenault2-26/+28
12 daysCodeGen: Use LibcallLoweringInfo for stack protector insertion (#176829)Matt Arsenault2-6/+9
12 days[SelectionDAG] Widen <2 x T> vector types for atomic load (#148897)jofrn2-23/+74
12 days[SelectionDAG] NFC: Remove redundant assert in ComputeNumSignBits.Sander de Smalen1-1/+0
12 days[DAGCombiner] Fold min/max vscale, C -> C (#174708)Luke Lau1-10/+29
12 daysDAG: Take LibcallLoweringInfo from analysis (#176800)Matt Arsenault1-16/+29
12 daysFastISel: Thread LibcallLoweringInfo through (#176799)Matt Arsenault2-1/+6
13 daysUse AllocaInst::getAllocationSize instead of manual size calculations (#176486)Jameson Nash2-8/+6
13 days[SelectionDAG] Add TRUNCATE_SSAT_S/U and TRUNCATE_USAT_U to canCreateUndefOrP...Jerry Dang1-0/+18
13 days[X86][WinEH] Insert nop after unwinding inline assembly (#176393)Nikita Popov2-0/+4
generated by cgit v1.1 at 2026-02-01 04:34:54 +0000