Sign in
opensecura
/
3p
/
openxla
/
iree
/
HEAD
e1ce3fa
[tuner]: add c/python binding for querying mma intrinsic (#19218)
by Bangtian Liu
· 10 hours ago
main
1654ce6
Bump iree-test-suites commit used for testing onnx ops. (#19229)
by Scott Todd
· 12 hours ago
b5b8059
[LLVMGPUVectorDistribute] Fix vector step distribute (#19227)
by Manupa Karunaratne
· 14 hours ago
1aada43
fix(TensorSliceOp::fold): ignore DenseResourceElementsAttr (#19182)
by Christopher McGirr
· 17 hours ago
901db6e
[i1] Do not emit `arith.trunci` cast from i1 to i1 (#19176)
by lialan
· 27 hours ago
8fd3e0d
Add arith-expand pass to lower ceildiv, floordiv ops (#19200)
by harsh-nod
· 28 hours ago
26ef79a
[Util] Fix AssumeIntOp::inferResultRanges bug (#19195)
by James Bartlett
· 34 hours ago
latest-snapshot
47432c6
[Codegen] Add pass to verify workgroup distribution (#19186)
by Quinn Dawkins
· 35 hours ago
35b495b
Use `iree-import-onnx --opset-version N` in ImportOnnxAction. (#19210)
by Scott Todd
· 35 hours ago
cd4aa72
[runtime][python] Fix device array deepcopy when not mappable (#19213)
by Boian Petkantchin
· 36 hours ago
82a89e3
Improving linking support for ROCM and ukernels. (#19211)
by Ben Vanik
· 2 days ago
f510664
Adding `HoistExecutableObjectsPass`.
by Ben Vanik
· 2 days ago
89feab6
Allowing #hal.executable.object data to be any serializable attr.
by Ben Vanik
· 2 days ago
4396bf1
[Codegen] Clean up MaterializeUserConfigs. NFC. (#19207)
by Jakub Kuderski
· 2 days ago
23c32c6
Use `gfx942`, not `gfx940` for MI300. (#19206)
by Benoit Jacob
· 2 days ago
c0619db
Replace unmaintained `create-release` action (#19205)
by Marius Brehler
· 2 days ago
bc9cbaa
Fix compiler errors in CUDA PJRT plugin (#19203)
by Twice
· 2 days ago
495adc6
[tuner] two new utility functions which are more friendly for c binding (#19199)
by Bangtian Liu
· 2 days ago
45df99d
Add linkBitcodeObjects to ROCMTarget.
by Ben Vanik
· 2 days ago
5b0740c
Adapt `test_ukernel.py` to an API change (#19197)
by Benoit Jacob
· 2 days ago
b68c535
Check `isIntOrFloat` before querying bitwidth (#19172)
by Ian Wood
· 2 days ago
540cebf
Reland "[DispatchCreation] Run preprocessing before..." (#18939)
by Ian Wood
· 2 days ago
1ab3b49
[Util] Erase state of modified ops (#19187)
by Ian Wood
· 2 days ago
df83f8e
[DispatchCreation] Enable bubble up extract slice for `linalg.generic` op with a single use. (#19174)
by MaheshRavishankar
· 3 days ago
c581951
Update documentation for release promotion process. (#19191)
by Scott Todd
· 3 days ago
50a001f
[DispatchCreation] Add CSE before canonicalization of `flow.dispatch.workgroups` (#19178)
by MaheshRavishankar
· 3 days ago
957ae60
Modify concat decomposition to only decompose non-outer concats. (#19177)
by MaheshRavishankar
· 3 days ago
d144f06
Bump version to 3.1.0 after releasing 3.0.0. (#19190)
by Scott Todd
· 3 days ago
d01432b
Turn on blocking of contractions by default (#19171)
by MaheshRavishankar
· 3 days ago
3ed07da
Document the versioning scheme (#19114)
by Marius Brehler
· 3 days ago
480c4f4
Integrate LLVM at 2f925d75dee8b4012d747d889ac4bb1d8a31d5a0 (#19184)
by Kunwar Grover
· 3 days ago
29c451b
Yet more IREEGPUAttrs cleanup: drop `get{A,B,C}SingleSubgroupLayout` methods (#19169)
by Benoit Jacob
· 5 days ago
e10342d
e2e matmul test improvements (#19016)
by Benoit Jacob
· 5 days ago
e3b6cc3
Bump Torch-MLIR to c26ca8b (#19168)
by zjgarvey
· 5 days ago
60cf4ab
More cleanup in `IREEGPUAttrs`. (#19161)
by Benoit Jacob
· 6 days ago
e3826d5
Change to <= dispatch count regression checks. (#19166)
by saienduri
· 6 days ago
43d3d1a
Guard the blocking of dynamic dimensions of contractions. (#19164)
by MaheshRavishankar
· 6 days ago
f750087
Update references to SHARK-Platform using new shark-ai name. (#19165)
by Scott Todd
· 6 days ago
bd51591
Integrate LLVM at 24c2c74bd29d4d550974f8249cbf8fdf1d033bfd (#19120)
by Andrea Faulds
· 6 days ago
c80fa3b
Support i1 datatype with an experimental flag. (#18713)
by lialan
· 6 days ago
8cb8743
Revert LLVM changes from #19007 (#19153)
by Andrzej Warzyński
· 6 days ago
5318fce
Strip .egg-info folders from iree-base-compiler whl files. (#19156)
by Scott Todd
· 6 days ago
81dd4e6
[Util][NFC] OptimizeIntArithmetic: reduce calls to `eraseState` (#19130)
by Ian Wood
· 6 days ago
d497571
[onnx][importer] Add support for externalized params (#18880)
by Vinayak Dev
· 7 days ago
34d9d5f
[Encoding] Implement OpAsmDialectInterface for IREE Encoding dialect. (#19146)
by Han-Chung Wang
· 7 days ago
9eaa4ef
[tuner]: Add a utility function to query supported MMA intrinsics (#19124)
by Bangtian Liu
· 7 days ago
4c0fd90
[Encoding][DT][NFC] Simplify the codes for EncodingAttr creation. (#19147)
by Han-Chung Wang
· 7 days ago
2a2bd06
More IREEGPUAttrs.cpp cleanups (#19142)
by Benoit Jacob
· 7 days ago
f828914
[GPU] Move tile and distribute pass before packing to intrinsic for TileAndfuse pipeline (#19053)
by Nirvedh Meshram
· 7 days ago
8391943
Bump torch-mlir to 30c519369ed7eabad0282d0f874500a9b41fcbbd (#19152)
by Vivek Khandelwal
· 7 days ago
c27c90c
Add `--pre` flag to pip installs for using nightly versions. (#19145)
by Scott Todd
· 7 days ago
eef2c3a
[GPU] Do not generate insert_strided_slice for 0-d vectors (#19149)
by Kunwar Grover
· 7 days ago
bf711a1
Add strip assertions flag (#19014)
by Ian Wood
· 7 days ago
dd9d616
[Flow] Add patterns to convert from `tensor.concat` to `flow.tensor.update`. (#19126)
by MaheshRavishankar
· 7 days ago
ef241f9
[LLVMGPU] Cleanup VirtualMMA functions to match refactoring on base MMAAttr (#19144)
by Stanley Winata
· 7 days ago
a70ea83
Moved windows workflow to Kubernetes hosted runner (#18967)
by Eliasj42
· 7 days ago
43b22de
Add punet benchmarking to the regression suite (#19088)
by saienduri
· 7 days ago
cb5d1ab
Rename `unroll_n_to_subgroups` to `subgroups_n` (#19102)
by Benoit Jacob
· 7 days ago
e10231c
[hip] Set the current device before calls into Hip. (#19103)
by Andrew Woloszyn
· 7 days ago
1a28f8d
Fixes for switch statements on enums (#19140)
by Benoit Jacob
· 8 days ago
68c35d7
Pass some `decompose-complex-ops` options in `torch-to-iree` (#19076)
by zjgarvey
· 8 days ago
4b15edd
Extend hoist collapse out of scf.forall pattern to use same offsets for all users (#19139)
by Nirvedh Meshram
· 8 days ago
ab35e1b
Simplify GPUTileSwizzleUtils and avoid creating unit dims. (#19105)
by Benoit Jacob
· 8 days ago
b08ea12
[LLVMGPU] Add 32x32x16 F8 MFMA intrinsic (#19106)
by Stanley Winata
· 8 days ago
11fe5cd
[Codegen] Add canonicalization pass to track lowering configs (#19138)
by Quinn Dawkins
· 8 days ago
1c43bcd
Fix requirements checking in pypi_deploy.sh and document scripts. (#19137)
by Scott Todd
· 8 days ago
2311e04
Populate more metadata in compiler and runtime Python packages. (#19118)
by Scott Todd
· 8 days ago
4477091
[python][tuner] Add bindings for `iree_codegen.compilation_info` (#19129)
by Jakub Kuderski
· 8 days ago
bc23e59
[python][tuner] Add bindings for `iree_codegen.translation_info` (#19128)
by Jakub Kuderski
· 8 days ago
ea03080
Re-enable aarch64 package builds (#19135)
by Andrzej Warzyński
· 8 days ago
d32ce2f
Block dynamic dimensions of contraction-like operations. (#19056)
by MaheshRavishankar
· 8 days ago
f3c1467
Hoist collapse shape out of scf.forall when possible and expand its destination (#19044)
by Nirvedh Meshram
· 8 days ago
11f0099
Disable tests depending on iree-model-artifacts GCS bucket. (#19127)
by Scott Todd
· 8 days ago
e0cd5b8
Disable aarch64 package builds until runners become available. (#19116)
by Scott Todd
· 8 days ago
4aa08f2
CDNA1/2 data tiling (#19100)
by Benoit Jacob
· 9 days ago
31e7343
A few more MMAIntrinsics (#19099)
by Benoit Jacob
· 9 days ago
5b9c4d9
[python][tuner] Set up bindings for iree_codegen (#19108)
by Jakub Kuderski
· 9 days ago
bbb87aa
Refactor existing MMA intrinsics (#19098)
by Benoit Jacob
· 9 days ago
2bfc639
[DispatchCreation] Collapse `iree_linalg_ext.attention` (#19012)
by Ian Wood
· 9 days ago
e8f755d
e2e matmul tests: support f64 (#19093)
by Benoit Jacob
· 9 days ago
c29ea90
[Codegen][NFC] Add dynamic tests for DecomposeBoundaryPackUnPackOpsPass (#19079)
by Max191
· 9 days ago
87e6e09
[LLVMCPU] Update tests. (#19109)
by Prashant Kumar
· 9 days ago
da286ea
[LLVMCPU] Add an additional level of tiling (#19027)
by Prashant Kumar
· 9 days ago
d1a991c
[Codegen] Update translation_info attribute assembly format. NFC. (#19107)
by Jakub Kuderski
· 9 days ago
fa6aa1c
[python] Simplify iree_gpu dialect bindings tests. NFC. (#19104)
by Jakub Kuderski
· 9 days ago
7f7cfb0
[VectorDistribution] Fix 0-rank vector.broadcast distribution (#19007)
by Kunwar Grover
· 9 days ago
55f2fce
[python][tuner] Add bindings for lowering config (#19096)
by Jakub Kuderski
· 9 days ago
55b998a
[tuner] add an iree-opt pass to strip configuration from executable sources (#19069)
by Bangtian Liu
· 10 days ago
300e0c3
[NFC][LLVMGPU] Cleanup layout configuration (#19059)
by Kunwar Grover
· 10 days ago
48f6dee
[python][tuner] Add bindings for MMAIntrinsic (#19095)
by Jakub Kuderski
· 10 days ago
b133218
Update docs, experimental and samples (#19065)
by Marius Brehler
· 10 days ago
915b06b
[hip] Re-land #19082 and #19074 (#19101)
by Andrew Woloszyn
· 10 days ago
c0dff68
[LLVMGPU] Remove Swizzle from workgroup reordering enums (#19094)
by Jakub Kuderski
· 10 days ago
3e98592
Increment package versions to 3.0.0 (#19097)
by Marius Brehler
· 10 days ago
5c45591
Bump torch-mlir to 8eb34dae78940efe529fedef5bbe96c905f3ee3b (#19089)
by jinchen
· 12 days ago
60726cd
Increment package versions to 2.9.1 following today's release. (#19090)
by Scott Todd
· 12 days ago
fd6e157
Revert #19074 and #19082 to fix broken AMDGPU tests. (#19086)
by Scott Todd
· 13 days ago
4d61d85
[Vulkan] Update sample transform spec (#19084)
by Jakub Kuderski
· 13 days ago
5d40ebe
Updated the plot types in deferred_work_queue. (#19082)
by Andrew Woloszyn
· 13 days ago
42f7822
[hip] Implement async allocation/deallocation. (#19074)
by Andrew Woloszyn
· 13 days ago
Next »