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

2.85k stars 620 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

[Encoding] Introduce "layouts" field to EncodingAttr.

#19215 hanhanW closed 18 hours ago
1
Include default tuning specs with the compiler

#19214 kuhar opened 2 days ago
4
[runtime][python] Fix device array deepcopy when not mappable

#19213 sogartar closed 2 days ago
1
[Codegen] Harden yielding logic in TileDispatchUsingForall

#19212 qedawkins closed 22 hours ago
1
Improving linking support for ROCM and ukernels.

#19211 benvanik closed 2 days ago
0
Use `iree-import-onnx --opset-version N` in ImportOnnxAction.

#19210 ScottTodd closed 2 days ago
0
[GPU] Add gather fusion tests for vector distribution

#19209 Groverkss opened 2 days ago
1
Make strip assertions default

#19208 IanWood1 opened 2 days ago
1
[Codegen] Clean up MaterializeUserConfigs. NFC.

#19207 kuhar closed 2 days ago
0
Use `gfx942`, not `gfx940` for MI300.

#19206 bjacob closed 2 days ago
0
Replace unmaintained `create-release` action

#19205 marbre closed 2 days ago
1
[LLVMGPU] add unit test for GPU shared memory reuse

#19204 manupak closed 19 hours ago
5
Fix compiler errors in CUDA PJRT plugin

#19203 PragmaTwice closed 2 days ago
0
PJRT plugin cannot be compiled successfully

#19202 PragmaTwice closed 2 days ago
1
[EmitC] Adapt lit tests to emitc.funcs default dialect

#19201 simon-camp closed 1 day ago
1
Add arith-expand pass to lower ceildiv, floordiv ops

#19200 harsh-nod closed 2 days ago
12
[tuner]: two new utility functions which are more friendly for c binding

#19199 bangtianliu closed 2 days ago
0
[WIP] Enable scatter fusion.

#19198 MaheshRavishankar opened 3 days ago
0
Adapt `test_ukernel.py` to an API change

#19197 bjacob closed 3 days ago
0
[Codegen][llvmgpu] Adding supoort for scf.if in prefetch shared memory pass

#19196 jerryyin opened 3 days ago
10
[Util] Fix AssumeIntOp::inferResultRanges bug

#19195 JamesMBartlett closed 2 days ago
1
Port existing ROCM ukernels from HIP to C.

#19194 bjacob closed 1 day ago
2
Consolidate pip index pages across repositories

#19193 ScottTodd opened 3 days ago
0
Release tracker - 3.1.0

#19192 ScottTodd opened 3 days ago
0
Update documentation for release promotion process.

#19191 ScottTodd closed 3 days ago
0
Bump version to 3.1.0 after releasing 3.0.0.

#19190 ScottTodd closed 3 days ago
0
[Codegen][LLVMGPU] Correctness issue with softmax due to TileDispatchUsingForall failing to fuse

#19189 qedawkins opened 3 days ago
0
[Codegen][LLVMGPU] Drop TransposeSharedMem pipeline

#19188 qedawkins opened 3 days ago
2
[Util] Erase state of modified ops

#19187 IanWood1 closed 3 days ago
3
[Codegen] Add pass to verify workgroup distribution

#19186 qedawkins closed 2 days ago
0
Fix crash due to complex types not being considered in `KernelDispatch.cpp`

#19185 giacs-epic closed 3 days ago
3
Integrate LLVM at 2f925d75dee8b4012d747d889ac4bb1d8a31d5a0

#19184 Groverkss closed 3 days ago
0
error: One or more operations with large vector sizes (16384 bytes) were found

#19183 pdhirajkumarprasad opened 4 days ago
1
fix(TensorSliceOp::fold): ignore DenseResourceElementsAttr

#19182 chrsmcgrr closed 1 day ago
0
[GPU]: error: <unknown>:0:0: in function main_graph$async_dispatch_2_softmax_Dx9xf32_dispatch_tensor_store void (ptr addrspace(1), ptr addrspace(1), ptr addrspace(1), i32, i32): unsupported dynamic alloca

#19181 pdhirajkumarprasad opened 4 days ago
0
[GPU]: stack frame size (294916) exceeds limit (131056) in function 'torch_jit$async_dispatch_1_softmax_64x4x144x144xf32_dispatch_tensor_store

#19180 pdhirajkumarprasad opened 4 days ago
1
[GPU]: 'arith.extui' op operand type 'i64' and result type 'i32' are cast incompatible

#19179 pdhirajkumarprasad opened 4 days ago
0
[DispatchCreation] Add CSE before canonicalization of `flow.dispatch.workgroups`

#19178 MaheshRavishankar closed 3 days ago
2
Modify concat decomposition to only decompose non-outer concats.

#19177 MaheshRavishankar closed 3 days ago
1
[i1] Do not emit `arith.trunci` cast from i1 to i1

#19176 lialan closed 2 days ago
8
Fuse all attention related dispatches.

#19175 MaheshRavishankar opened 5 days ago
4
[DispatchCreation] Enable bubble up extract slice for `linalg.generic` op with a single use.

#19174 MaheshRavishankar closed 3 days ago
0
Support rank-reduction slices in bubble-up extract slice

#19173 MaheshRavishankar closed 3 days ago
0
Check `isIntOrFloat` before querying bitwidth

#19172 IanWood1 closed 3 days ago
6
Turn on blocking of contractions by default

#19171 MaheshRavishankar closed 3 days ago
4
Import `iree_amdgpu_library.cmake` from `users/benvanik/amdgpu`

#19170 bjacob closed 6 days ago
4
Yet more IREEGPUAttrs cleanup: drop `get{A,B,C}SingleSubgroupLayout` methods

#19169 bjacob closed 5 days ago
0
Bump Torch-MLIR to c26ca8b

#19168 zjgarvey closed 6 days ago
0
Crash in OptimizeIntArithmetic Pass

#19167 IanWood1 closed 3 days ago
0
Change to <= dispatch count regression checks.

#19166 saienduri closed 6 days ago
1

Previous Next