aboutsummaryrefslogtreecommitdiff
path: root/llvm/test/Transforms/InstCombine
AgeCommit message (Expand)AuthorFilesLines
4 hoursValueTracking: Handle constant structs in computeKnownFPClass (#184192)Matt Arsenault1-3/+2
5 hoursValueTracking: Handle ConstantDataSequential in computeKnownFPClass (#184191)Matt Arsenault1-17/+16
6 hoursInstCombine: Support extractvalue in SimplifyDemandedFPClass (#184171)Matt Arsenault1-0/+67
14 hours[InstCombine] make `foldBinOpIntoSelectOrPhi` fold on all operands (#183692)Kiva6-20/+126
15 hours[InstCombine] Don't strip leading zero index for overaligned vector GEP (#184...Nikita Popov1-1/+10
26 hours[AMDGPU] Add dereferenceable retAttr to a call to llvm.amdgcn.implicitarg.ptr...Yoonseo Choi1-0/+78
34 hoursRevert "Avoid maxnum(sNaN, x) optimizations / folds (#170181)" (#184125)Lewis Crawford1-4/+2
2 days[InstCombine] Always fold nonnull assumptions into operand bundles (#169923)Nikolas Klauser3-68/+27
3 days[LangRef] Clarify nsz semantics (#180906)Nikita Popov1-1/+1
3 daysValueTracking/AMDGPU: handle mbcnt in computeKnownBitsFromOperator (#183229)Ruiling, Song3-18/+97
5 daysInstCombine: Stop applying nofpclass from use nofpclass attribute (#183835)Matt Arsenault1-2/+21
5 days[InstCombine] Replace alloca with undef size with poison instead of null (#18...Manuel Carrasco1-0/+29
6 days[InstCombine] Combine extract from get_active_lane_mask where all lanes inact...Kerry McLaughlin1-0/+40
6 days[InstCombine] Fold shift of boolean zext to logic sequence (#180596)Gergo Stomfai1-0/+143
7 days[InstCombine] Fold min/max of two subtracts with common RHS (#183240)Nathiyaa Sengodan1-0/+169
8 days[SimplifyLibCalls] Avoid simplifying pow(x, 2.0) -> x * x with math-errno. (#...Ricardo Jesus1-38/+8
8 days[LLVM][ConstantFolding] Add vector ConstantInt/FP support to ReadDataFromGlob...Paul Walker1-1/+27
9 days[InstCombine] Restrict foldICmpOfVectorReduce to one-use (#182833)Ramkumar Ramachandra1-0/+30
10 days[InstCombine] Fold icmp (vreduce_(or|and) %x), (0|-1) (#182684)Ramkumar Ramachandra1-0/+101
10 days[InstCombine] Only ignore first zero index during GEP canonicalization (#180764)Nikita Popov1-1/+1
10 daysRevert "[InstCombine] Allow freezing multiple operands (#154336)" (#182769)Eli Friedman9-58/+53
11 daysInstCombine: Fold absorbing fmul of compared 0 into select (#172381)Matt Arsenault1-15/+7
11 days[InstCombine] Update testAiden Grossman1-1/+1
11 days[InstCombine] Transform splat before n x i1 for vec.reduce.add (#182213)Hongxu Xu1-6/+1
11 daysInstCombine: Add baseline test for fcmp-0-select combine (#172380)Matt Arsenault1-0/+601
12 days[AMDGPU] Use a general form of intrinsic for tensor load/store (#182334)Changpeng Fang1-185/+0
13 days[InstCombine] Do not perform fcmp -> icmp transformation if denormal inputs m...Justin Holewinski1-0/+52
13 daysInstCombine: Fold bitcast of vector with constant to scalar (#179042)Matt Arsenault3-7/+135
13 daysAMDGPU: Strip sign bit operations on llvm.amdgcn.trig.preop uses (#179712)Matt Arsenault1-0/+64
13 days[InstCombine] Allow simplifying FP selects of cmpxchg instructions. (#181977)Ricardo Jesus1-0/+67
14 daysInstCombine: Implement SimplifyDemandedFPClass for ldexp (#180703)Matt Arsenault1-48/+31
14 daysInstCombine: Add baseline SimplifyDemandedFPClass ldexp tests (#180702)Matt Arsenault1-0/+506
14 days[IRBuilder] Use ptrtoaddr in CreatePtrDiff() (#181855)Nikita Popov2-8/+8
2026-02-17[NFC][AMDGPU] Use `zeroinitializer` instead of `null` for `ptr addrspace(2/3/...Shilei Tian2-2/+2
2026-02-16[instcombine][x86]: simplifyx86fpmaxmin - allow negzero for single operand (#...Madhur Kumar1-0/+103
2026-02-15[InstCombine] Avoid invalid bitcast across address spaces in foldIntegerTyped...Michal Paszkowski1-0/+40
2026-02-16[InstCombine] Optimise the expression `(C && A) || (!C && B)` with `FoldOrOfL...Rajveer Singh Bharadwaj1-1/+115
2026-02-15[InstCombine][NFC] Add test for existing fold (#181555)Tirthankar Mazumder1-0/+23
2026-02-13[InstCombine] Fold ((X + AddC) & Mask) ^ Mask to (~AddC - X) & Mask (#174278)hanbeom1-0/+142
2026-02-11InstCombine: Fix wrong insert point for sqrt -> copysign simplify (#180838)Matt Arsenault1-0/+12
2026-02-11InstCombine: Fix wrong insert point for various fmul->copysign simplifies (#1...Matt Arsenault1-0/+76
2026-02-11InstCombine: Fix wrong insert point for fdiv->copysign simplify (#180839)Matt Arsenault1-0/+24
2026-02-11InstCombine: Fix insert point for rounding intrinsic -> copysign (#180837)Matt Arsenault1-0/+12
2026-02-10InstCombine: Fix broken insert point for fdiv replacement (#180830)Matt Arsenault1-0/+19
2026-02-10InstCombine: Use SimplifyDemandedFPClass on fmul (#177490)Matt Arsenault14-53/+53
2026-02-09[InstCombine] Support minimumnum/maximumnum (#180529)Nikita Popov2-0/+826
2026-02-09[InstCombine] fold `gepi _, (srem x, y)` to `gepi _, (urem x, y)` if `y` is p...Kiva1-0/+128
2026-02-09[InstCombine] Regenerate test checks (NFC)Nikita Popov1-16/+42
2026-02-09[InstCombine] Avoid overflow in `foldVecExtTruncToExtElt` (#180414)Yingwei Zheng1-0/+34
2026-02-09[InstCombine] Drop nonnull assumes if the pointer is already known to be nonn...Nikolas Klauser1-0/+37
generated by cgit v1.1 at 2026-03-04 23:02:35 +0000