User Details
User Details
- User Since
- Dec 24 2019, 5:47 AM (169 w, 4 d)
Wed, Mar 22
Wed, Mar 22
Nice, thanks!
Fri, Mar 17
Fri, Mar 17
Wed, Mar 15
Wed, Mar 15
antiagainst committed rG59e4fbfcd064: [mlir][memref] Fold subview into GPU subgroup MMA load/store ops (authored by antiagainst).
[mlir][memref] Fold subview into GPU subgroup MMA load/store ops
antiagainst added inline comments to D146150: [mlir][memref] Fold subview into GPU subgroup MMA load/store ops.
antiagainst updated the summary of D146150: [mlir][memref] Fold subview into GPU subgroup MMA load/store ops.
antiagainst updated the diff for D146150: [mlir][memref] Fold subview into GPU subgroup MMA load/store ops.
Address comments
antiagainst added inline comments to D146150: [mlir][memref] Fold subview into GPU subgroup MMA load/store ops.
antiagainst requested review of D146150: [mlir][memref] Fold subview into GPU subgroup MMA load/store ops.
Tue, Mar 14
Tue, Mar 14
antiagainst committed rG141b7d49a3a4: [mlir][spirv] Fix UnifyAliasedResourcePass for 64-bit index (authored by antiagainst).
[mlir][spirv] Fix UnifyAliasedResourcePass for 64-bit index
antiagainst added inline comments to D145079: [mlir][spirv] Fix UnifyAliasedResourcePass for 64-bit index.
antiagainst added inline comments to D146080: [mlir][spirv] Fix scf.yield pattern conversion.
Fri, Mar 10
Fri, Mar 10
antiagainst accepted D145760: [mlir][spirv][vector] Add pattern to convert reduction to SPIR-V dot prod.
Awesome!
Mon, Mar 6
Mon, Mar 6
Thanks!
Wed, Mar 1
Wed, Mar 1
Nice, thanks for fixing this!
antiagainst requested review of D145079: [mlir][spirv] Fix UnifyAliasedResourcePass for 64-bit index.
Sun, Feb 26
Sun, Feb 26
antiagainst committed rG85365b16c8c3: [mlir][spirv] Fix Physical32/Physical64 support for OpenCL (authored by antiagainst).
[mlir][spirv] Fix Physical32/Physical64 support for OpenCL
antiagainst committed rG9a4c768a7d83: [mlir][spirv] Respect client API requirements for 64-bit index (authored by antiagainst).
[mlir][spirv] Respect client API requirements for 64-bit index
antiagainst added inline comments to D144819: [mlir][spirv] Respect client API requirements for 64-bit index.
antiagainst added a comment to D144827: [mlir][spirv] Fix Physical32/Physical64 support for OpenCL.
antiagainst added inline comments to D144819: [mlir][spirv] Respect client API requirements for 64-bit index.
antiagainst committed rG80bb9477a973: [mlir] NFC: Prepare GPUToSPIRV tests for supporting 64bit index (authored by antiagainst).
[mlir] NFC: Prepare GPUToSPIRV tests for supporting 64bit index
antiagainst requested review of D144827: [mlir][spirv] Fix Physical32/Physical64 support for OpenCL.
antiagainst updated the diff for D144819: [mlir][spirv] Respect client API requirements for 64-bit index.
Rebase
antiagainst requested review of D144826: [mlir] NFC: Prepare GPUToSPIRV tests for supporting 64bit index.
Sat, Feb 25
Sat, Feb 25
antiagainst requested review of D144819: [mlir][spirv] Respect client API requirements for 64-bit index.
Fri, Feb 24
Fri, Feb 24
antiagainst planned changes to D107846: [mlir] Graduate op availability to OpBase.td.
antiagainst planned changes to D133873: [mlir][tensor] Add pattern to hoist extract/insert slice ops.
antiagainst planned changes to D135546: [mlir][vector] Simplify vector transfer hoisting logic.
antiagainst accepted D144460: [mlir][GPUToVulkan] Port conversion passes and `mlir-vulkan-runner` to opaque pointers.
Thanks for plumbing this through!
Feb 16 2023
Feb 16 2023
antiagainst added a comment to D144113: [mlir][vectorToGPU] Fix type used when folding transpose into read op.
To close the loop: https://reviews.llvm.org/D144159 landed for improving the accessor methods.
Feb 15 2023
Feb 15 2023
antiagainst committed rGa1aad28d297a: [mlir][vector] NFC: Improve vector type accessor methods (authored by antiagainst).
[mlir][vector] NFC: Improve vector type accessor methods
antiagainst added a comment to D144159: [mlir][vector] Improve vector type accessor methods.
antiagainst requested review of D144159: [mlir][vector] Improve vector type accessor methods.
antiagainst committed rG78e172fc92e7: [mlir][spirv] Support i32 memref.atomic_rmw conversion (authored by antiagainst).
[mlir][spirv] Support i32 memref.atomic_rmw conversion
antiagainst added a comment to D144113: [mlir][vectorToGPU] Fix type used when folding transpose into read op.
IMO, the accessor name could use some improvements---it's very confusing and error-prone just being getVectorType(); making it something like getSrcVectorType() would make it far easier to avoid pitfalls. At least for me, I need to read the impl to double check what it means by just getVectorType().
Feb 14 2023
Feb 14 2023
Feb 13 2023
Feb 13 2023
antiagainst requested review of D143952: [mlir][spirv] Support i32 memref.atomic_rmw conversion.
Nice, thanks!
antiagainst accepted D143922: [mlir][gpu] Add support for unsigned integer extend in vector to gpu.subgroup_mma lowering.
LG, thanks!
Feb 11 2023
Feb 11 2023
Feb 1 2023
Feb 1 2023
antiagainst committed rG2c46051aa9d3: [mlir][spirv] Fix vector type mismatch in UnifyAliasedResourcePass (authored by antiagainst).
[mlir][spirv] Fix vector type mismatch in UnifyAliasedResourcePass
Jan 31 2023
Jan 31 2023
antiagainst updated the diff for D143042: [mlir][spirv] Fix vector type mismatch in UnifyAliasedResourcePass.
Update
antiagainst requested review of D143042: [mlir][spirv] Fix vector type mismatch in UnifyAliasedResourcePass.
Jan 30 2023
Jan 30 2023
antiagainst committed rGd3ddbe153e4c: [mlir][vector] Clarify vector.contract promotion behavior (authored by antiagainst).
[mlir][vector] Clarify vector.contract promotion behavior
antiagainst retitled D142915: [mlir][vector] Clarify vector.contract promotion behavior from [mlir][vector] Clarify integer vector.contract promotion behavior to [mlir][vector] Clarify vector.contract promotion behavior.
antiagainst updated the summary of D142915: [mlir][vector] Clarify vector.contract promotion behavior.
antiagainst updated the diff for D142915: [mlir][vector] Clarify vector.contract promotion behavior.
Address comments
antiagainst updated the diff for D142915: [mlir][vector] Clarify vector.contract promotion behavior.
Grammar
antiagainst added inline comments to D142915: [mlir][vector] Clarify vector.contract promotion behavior.
antiagainst updated the diff for D142915: [mlir][vector] Clarify vector.contract promotion behavior.
Address comments
antiagainst requested review of D142915: [mlir][vector] Clarify vector.contract promotion behavior.
Jan 28 2023
Jan 28 2023
Jan 23 2023
Jan 23 2023
antiagainst committed rGb41eb9601cf1: [mlir][spirv] Move uint asm name test to the proper place (authored by antiagainst).
[mlir][spirv] Move uint asm name test to the proper place
antiagainst added inline comments to D142295: [mlir] support unsigned int in mlir::spirv::ConstantOp::getAsmResultNames.
antiagainst added inline comments to D142295: [mlir] support unsigned int in mlir::spirv::ConstantOp::getAsmResultNames.
Jan 19 2023
Jan 19 2023
Jan 17 2023
Jan 17 2023
antiagainst added a comment to D141856: Fix crash in Spirv -lower-host-to-llvm pass.
Nice, thanks Mehdi!
Jan 13 2023
Jan 13 2023
Jan 12 2023
Jan 12 2023
Jan 11 2023
Jan 11 2023
Jan 9 2023
Jan 9 2023
antiagainst committed rG84ce59a0e4e4: [mlir][spirv] Fix MLIR code snippet marker in TD files (authored by antiagainst).
[mlir][spirv] Fix MLIR code snippet marker in TD files
antiagainst requested review of D141316: [mlir][spirv] Fix MLIR code snippet marker in TD files.
Jan 6 2023
Jan 6 2023
Thanks for the cleanup!
Jan 5 2023
Jan 5 2023
nice, thanks for fixing it!
Jan 4 2023
Jan 4 2023
Dec 30 2022
Dec 30 2022
Dec 29 2022
Dec 29 2022
antiagainst committed rG56c069887b95: [mlir][spirv] Fail vector.bitcast conversion with different bitwidth (authored by antiagainst).
[mlir][spirv] Fail vector.bitcast conversion with different bitwidth
antiagainst requested changes to D140576: [mlir][gpu][spirv] Lower gpu reduction ops to spirv.
Nice! Just a few nits.
Dec 22 2022
Dec 22 2022
antiagainst committed rG1a0453eb44cd: [mlir][vector] Fix bug in extractOp folding (authored by ThomasRaoux).
[mlir][vector] Fix bug in extractOp folding
antiagainst committed rGc03fe1ebbd8e: [mlir][spirv] Add StreamingInterfaceINTEL to SPIRVBase.td (authored by mmendell).
[mlir][spirv] Add StreamingInterfaceINTEL to SPIRVBase.td
antiagainst added a comment to D140476: [mlir][spirv] Add StreamingInterfaceINTEL to mlir SPIRVBase.td.
Should this be set somewhere in Phabricator?
antiagainst added a comment to D140476: [mlir][spirv] Add StreamingInterfaceINTEL to mlir SPIRVBase.td.