iree-org iree issues - Githubissues

iree-org / iree

A retargetable MLIR-based machine learning compiler and runtime toolkit.

http://iree.dev/

Apache License 2.0

2.47k stars 551 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

[LLVMGPU] Landing VectorDistribution pipeline for attention

#17773 raikonenfnu opened 12 hours ago
0
Run sharktank llama tests on CI, qualify test names.

#17772 ScottTodd opened 12 hours ago
1
[codegen] Add max_workgroup_counts to TargetWgpAttr

#17771 krzysz00 opened 15 hours ago
0
Integrate llvm-project @0cc3fe460105c4c0c78139d7a78da557c3502298

#17770 Max191 opened 16 hours ago
0
ASAN issues on hip / rocm runtime

#17769 suryajasper opened 16 hours ago
3
Integrate llvm-project @2c1ae801e1b66a09a15028ae4ba614e0911eec00

#17768 Max191 closed 16 hours ago
3
Switch back to more stable, better performance mi210 benchmarking

#17767 saienduri opened 1 day ago
0
Start cleaning up tests/e2e/.

#17766 ScottTodd opened 1 day ago
2
Fold `arith.exti` operation into conv

#17765 IanWood1 opened 1 day ago
1
[LLVMGPU][ROCm] Plumb through i8, i8 -> i32 MFMA intrinsics

#17764 kuhar closed 21 hours ago
1
[Encoding] Add an optional bcast_map attribute to EncodingAttr.

#17763 hanhanW opened 1 day ago
5
Expose llvm dialect in python bindings

#17762 harsh-nod opened 1 day ago
1
[tilingInterface] Update the tile sizes to i64 attr type

#17761 pashu123 closed 19 hours ago
0
expand_shape fails to compile with more than one dynamic dim in a reassoc group

#17760 zjgarvey opened 1 day ago
0
dominance-related failure occurs when compiling an opt model

#17759 zjgarvey opened 1 day ago
3
[Codegen] Add interface tensor reshape foldings to TileAndDistribute

#17758 Max191 opened 1 day ago
1
[Codegen] Do not consider parallel regions in bufferization analysis

#17757 Max191 opened 1 day ago
2
`EncodingAttr` to track type of source op

#17756 lialan opened 1 day ago
1
Exclude arm-sme tests by default on Bazel CI.

#17755 ScottTodd closed 1 day ago
1
Integrate llvm-project @34e34a03ac83b51e90f8788945f9668446e468f8

#17754 Max191 closed 22 hours ago
2
Update artifact file names in validate_and_publish_release.yml.

#17753 ScottTodd closed 1 day ago
0
[Codegen] Change CompilationInfoAttr to take a lowering config interface

#17752 qedawkins closed 1 day ago
1
Add e2e test suite for FA2

#17751 erman-gurses opened 2 days ago
0
[Riscv] Implement a empty cpu initialize function for generic-riscv_64

#17750 ElEHsiang opened 2 days ago
0
Add Gemma pipeline to benchmark_large

#17749 mariecwhite opened 2 days ago
0
Update dotprod microbenchmark artifacts

#17748 mariecwhite closed 2 days ago
0
[Flow] Enable softmax-like fusion under aggressive fusion.

#17747 MaheshRavishankar closed 19 hours ago
3
Fix build_package.yml by giving uploads unique names.

#17746 ScottTodd closed 2 days ago
1
[Flow] Do not fuse operations that truncate bitwidths with operations that extend it

#17745 MaheshRavishankar opened 2 days ago
0
[LLVMGPU] Set prefetching on translation info

#17744 Groverkss closed 2 days ago
2
Add docs for Encoding, IREECodegen, and IREEGPU dialects.

#17743 ScottTodd closed 2 days ago
0
[Codegen] Improve ROCm-specific LLVM translations

#17742 krzysz00 opened 3 days ago
0
Change to mi250 runner for sdxl benchmarking.

#17741 saienduri closed 3 days ago
1
Fixes for upstream mlir changes.

#17740 hanhanW opened 3 days ago
0
[GlobalOpt] Disable pack->expand_shape propagation.

#17739 hanhanW closed 3 days ago
2
Refactor `check` dialect to support multiple devices via `hal.tensor.device`.

#17738 benvanik opened 3 days ago
0
Ensure IREE GPU dialect is registered for all GPU targets (fixes #17736)

#17737 andfau-amd closed 3 days ago
1
Build fails due to "LLVM ERROR: can't create Attribute 'mlir::iree_compiler::IREE::GPU::ComputeBitwidthsAttr' because storage uniquer isn't initialized: the dialect was likely not loaded, or the attribute wasn't added with addAttributes<...>() in the Dialect::initialize() method."

#17736 andfau-amd closed 3 days ago
2
Add loop-invariant-subset-hoisting pass to iree-opt

#17735 Groverkss opened 3 days ago
0
Assertion `input.size() == permutation.size() && "expected input rank to equal permutation rank"' failed

#17734 AmosLewis closed 12 hours ago
6
[CPU] Set tile_size=16 only when the innermost dimension is unit dim.

#17733 hanhanW opened 4 days ago
0
Add developer doc for model development debugging.

#17732 ScottTodd closed 3 days ago
0
[CPU] Tile outer parallel dims with 1 before lowering to ukernels.

#17731 hanhanW closed 2 days ago
2
[WIP] Making HAL command buffers take buffers as indirect args.

#17730 benvanik opened 4 days ago
0
Merge `round_dims_to` and `matmul_narrow_{M,N}` into a single `max_padding` array attribute

#17729 bjacob opened 4 days ago
0
[Im2col] Add decomposition pass for iree_linalg_ext.im2col

#17728 Max191 closed 2 days ago
2
[CPU] Limit the use of [8, 32, 16] gemm vector sizes to CPUs w/ avx512f feature

#17727 hanhanW closed 4 days ago
0
Add check tests for more tensor dialect ops.

#17726 ScottTodd closed 4 days ago
0
Collapse dims when producer is unpack op

#17725 IanWood1 opened 1 week ago
5
Removing nested command buffers and adding indirect execution.

#17724 benvanik opened 1 week ago
0