aboutsummaryrefslogtreecommitdiff
path: root/llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
AgeCommit message (Expand)AuthorFilesLines
2026-01-08[AMDGPU] Fix a potential use-after-erase in `AMDGPUPromoteAlloca` pass (#174529)Shilei Tian1-3/+12
2025-12-19[AMDGPU] In promote-alloca, if index is dynamic, sandwich load with bitcasts ...Kevin Choi1-0/+31
2025-12-18AMDGPU/PromoteAlloca: Fix handling of users of multiple allocas (#172771)macurtis-amd1-9/+19
2025-12-12AMDGPU/PromoteAlloca: Refactor into analysis / commit phases (#170512)Nicolai Hähnle1-304/+347
2025-12-08Revert "[AMDGPU] Enable i8 GEP promotion for vector allocas" (#171087)Jan Patrick Lehr1-18/+3
2025-12-08[AMDGPU] Enable i8 GEP promotion for vector allocas (#166132)Harrison Hao1-3/+18
2025-12-06Reland "AMDGPU/PromoteAlloca: Always use i32 for indexing (#170511)" (#170956)Nicolai Hähnle1-6/+7
2025-12-06Reland "AMDGPU/PromoteAlloca: Simplify how deferred loads work (#170510)" (#1...Nicolai Hähnle1-46/+36
2025-12-05Revert "AMDGPU/PromoteAlloca: Simplify how deferred loads work (#170510)"Nicolai Hähnle1-34/+46
2025-12-05Revert "AMDGPU/PromoteAlloca: Always use i32 for indexing (#170511)"Nicolai Hähnle1-7/+6
2025-12-05AMDGPU/PromoteAlloca: Always use i32 for indexing (#170511)Nicolai Hähnle1-6/+7
2025-12-05AMDGPU/PromoteAlloca: Simplify how deferred loads work (#170510)Nicolai Hähnle1-46/+34
2025-12-03AMDGPU/PromoteAlloca: Extract getVectorTypeForAlloca helper (#170509)Nicolai Hähnle1-17/+28
2025-11-14[AMDGPU] Make use of getFunction and getMF. NFC. (#167872)Jay Foad1-3/+3
2025-09-12[AMDGPUPromoteAlloca][NFC] Avoid unnecessary APInt/int64_t conversions (#157864)Fabian Ritter1-12/+10
2025-09-10[AMDGPU] Generate canonical additions in AMDGPUPromoteAlloca (#157810)Fabian Ritter1-1/+1
2025-09-10[AMDGPU] Treat GEP offsets as signed in AMDGPUPromoteAlloca (#157682)Fabian Ritter1-7/+10
2025-08-26[AMDGPU] AMDGPUPromoteAlloca: increase default max-regs to 32 (#155076)Carl Ritson1-2/+6
2025-06-24[AMDGPU] Replace dynamic VGPR feature with attribute (#133444)Diana Picus1-1/+9
2025-06-20AMDGPU: Remove legacy PM version of AMDGPUPromoteAllocaToVector (#144986)Matt Arsenault1-40/+0
2025-06-16Revert "[AMDGPU] Extended vector promotion to aggregate types." (#144366)zGoldthorpe1-65/+41
2025-06-13[AMDGPU] Extended vector promotion to aggregate types. (#143784)zGoldthorpe1-41/+65
2025-06-02[AMDGPU] Promote nestedGEP allocas to vectors (#141199)Harrison Hao1-3/+28
2025-05-21[AMDGPU] PromoteAlloca: handle out-of-bounds GEP for shufflevector (#139700)Robert Imschweiler1-1/+3
2025-05-01[AMDGPU] Max. WG size-induced occupancy limits max. waves/EU (#137807)Lucas Ramirez1-23/+9
2025-04-14[AMDGPU] Avoid crashes for non-byte-sized types in PromoteAlloca (#134042)Fabian Ritter1-10/+24
2025-03-31[IRBuilder] Add new overload for CreateIntrinsic (#131942)Rahul Joshi1-3/+3
2025-03-20[Target] Use *Set::insert_range (NFC) (#132140)Kazu Hirata1-1/+1
2025-03-19[AMDGPU] Fix typing error in multi dimensional promote alloca (#131763)Carl Ritson1-12/+15
2025-03-18AMDGPU: Use freeze poison instead of undef in alloca promotion (#131285)Matt Arsenault1-1/+19
2025-03-14[NFC][AMDGPU] Replace direct arch comparison with `isAMDGCN()` (#131357)Shilei Tian1-1/+1
2025-03-12[AMDGPU] Fix typing error introduce in promote alloca changeCarl Ritson1-1/+4
2025-03-12[AMDGPU] Extend promotion of alloca to vectors (#127973)Carl Ritson1-25/+94
2025-03-03[AMDGPU] Simplify conditional expressions. NFC. (#129228)Jay Foad1-1/+1
2025-02-24[AMDGPU] Update PromoteAlloca to handle GEPs with variable offset. (#122342)Sumanth Gundapaneni1-6/+8
2025-02-09[AMDGPU] Avoid repeated hash lookups (NFC) (#126430)Kazu Hirata1-2/+1
2025-01-27[NFC][AMDGPU] Improve code introduced in #124607 (#124672)Shilei Tian1-15/+5
2025-01-27[AMDGPU] Handle invariant marks in `AMDGPUPromoteAllocaPass` (#124607)Shilei Tian1-4/+26
2025-01-23[AMDGPU] Occupancy w.r.t. workgroup size range is also a range (#123748)Lucas Ramirez1-1/+1
2024-11-06AMDGPU: Improve vector of pointer handling in amdgpu-promote-alloca (#114144)Matt Arsenault1-39/+34
2024-10-29AMDGPU: Fix producing invalid IR on vector typed getelementptr (#114113)Matt Arsenault1-4/+14
2024-10-17[LLVM] Make more use of IRBuilder::CreateIntrinsic. NFC. (#112706)Jay Foad1-17/+9
2024-10-11[NFC] Rename `Intrinsic::getDeclaration` to `getOrInsertDeclaration` (#111752)Rahul Joshi1-7/+7
2024-09-30[NFC] Use initial-stack-allocations for more data structures (#110544)Jeremy Morse1-1/+1
2024-09-19[LLVM] Use {} instead of std::nullopt to initialize empty ArrayRef (#109133)Jay Foad1-1/+1
2024-09-06AMDGPU: Remove unnecessary pointer bitcastMatt Arsenault1-4/+2
2024-08-14AMDGPU: Stop promoting allocas with addrspacecast users (#104051)Matt Arsenault1-8/+4
2024-05-28[AMDGPU][PromoteAlloca] Don't stop when an alloca is too big to promote (#93466)Pierre van Houtryve1-14/+18
2024-04-12[AMDGPU] Fix a potential wrong return value indicating whether a pass modifie...Shilei Tian1-1/+1
2024-03-19[AMDGPU][PromoteAlloca] Whole-function alloca promotion to vector (#84735)Pierre van Houtryve1-26/+108
generated by cgit v1.2.3 (git 2.39.1) at 2026-05-19 21:15:06 +0000