newling

Follow

James Newling newling

Follow

16 followers · 0 following

AMD
https://newling.github.io/

Achievements

Achievements

Pinned Loading

idiap/zentas Public

Partitional data clustering around centers

C++ 8 3
idiap/eakmeans Public

Implementation of fast exact k-means algorithms

C++ 45 10
ROCm/MIOpenGEMM Public archive

MIOpenGEMM is now deprecated

C++ 62 11
graphcore/poprithms Public

Graph algorithms for machine learning frameworks

C++ 29 2

669 contributions in the last year

Skip to contributions year list

Learn how we count contributions

Less

More

Contribution activity

July 2025

Created 11 commits in 4 repositories

Created 1 repository

newling/mlir-www SCSS
This contribution was made on Jul 18

Created a pull request in llvm/llvm-project that received 8 comments

Revert [mlir][vector] Use vector.broadcast in place of vector.splat

This reverts commit a599e60. This should only be landed after #148027, at which point we don't need to assume that vector.broadcast has been lowere…

+6 −7 lines changed • 8 comments

Opened 16 other pull requests in 3 repositories

llvm/llvm-project 5 open 6 merged

[mlir][vector] vector.splat deprecation: folding/canonicalizing parity with broadcast
This contribution was made on Jul 23
[mlir][vector] Avoid use of vector.splat in transforms
This contribution was made on Jul 23
Reland "[mlir][vector] Use vector.broadcast in place of vector.splat"
This contribution was made on Jul 22
[WIP][mlir] DenseElementsAttr::reshape(arg): make arg a shape, not a type
This contribution was made on Jul 21
[mlir][vector][NFC] Add deprecation notice to splat's .td
This contribution was made on Jul 18
[mlir][llvm dialect] Verify element type of nested types
This contribution was made on Jul 15
[mlir][vector] Use vector.broadcast in place of vector.splat
This contribution was made on Jul 10
[mlir][vector] Support direct broadcast conversion (LLVM & SPIRV)
This contribution was made on Jul 10
[mlir][armsme][vector] Replace splat with broadcast
This contribution was made on Jul 10
[WIP][MLIR][Vector] Deprecate vector.splat
This contribution was made on Jul 9
[WIP] Linearization without conversion
This contribution was made on Jul 8

iree-org/iree 1 closed 1 merged 2 open

[not for review] [CI test] Triangulating where tests run in CI
This contribution was made on Jul 22
[codegen] use vector.broadcast instead of vector.splat
This contribution was made on Jul 21
[wip][llvmgpu][codegen] Remove warp reduction pipeline
This contribution was made on Jul 21
[Codegen] linalg.generic with dynamic reduction dim: use LLVMGPUVectorDistribution.
This contribution was made on Jul 21

llvm/mlir-www 1 merged

Notification and Discourse link for vector.splat deprecation
This contribution was made on Jul 18

Reviewed 19 pull requests in 3 repositories

llvm/llvm-project 15 pull requests

[mlir][vector] Avoid use of vector.splat in transforms
This contribution was made on Jul 23
[mlir][Vector] Remove vector.extractelement and vector.insertelement ops
This contribution was made on Jul 23
[mlir][NFC] update Conversion create APIs (4/n)
This contribution was made on Jul 23
[mlir][vector] Add a check to ensure input vector rank equals target shape rank
This contribution was made on Jul 23
[mlir][vector] Add a check to ensure input vector rank equals target shape rank
This contribution was made on Jul 23
[mlir][armsme][vector] Replace splat with broadcast
This contribution was made on Jul 22
[WIP][mlir] DenseElementsAttr::reshape(arg): make arg a shape, not a type
This contribution was made on Jul 22
[mlir][NFC] update mlir/Dialect create APIs (19/n)
This contribution was made on Jul 22
[mlir][vector] Folder: shape_cast(extract) -> extract
This contribution was made on Jul 21
[mlir][llvm dialect] Verify element type of nested types
This contribution was made on Jul 16
[mlir][vector] Support direct broadcast conversion (LLVM & SPIRV)
This contribution was made on Jul 16
[MLIR][Vector]Add constraints to vector.shape_cast(constant) -> constant
This contribution was made on Jul 14
[MLIR] [Vector] Linearization patterns for vector.load and vector.store
This contribution was made on Jul 14
[mlir][Vector] Add vector.shuffle tree transformation
This contribution was made on Jul 8
[mlir][Vector] Support scalar 'vector.insert' in vector linearization
This contribution was made on Jul 7

iree-org/iree 3 pull requests

[wip][llvmgpu][codegen] Remove warp reduction pipeline
This contribution was made on Jul 23
[Codegen] linalg.generic with dynamic reduction dim: use LLVMGPUVectorDistribution.
This contribution was made on Jul 23
[LLVMGPU] Add canonicalization for select(pred, true, false) -> broadcast(pred)
This contribution was made on Jul 14

nod-ai/iree-amd-aie 1 pull request

[Reduction] Verify different sizes and types work for reduction.
This contribution was made on Jul 22

Created an issue in iree-org/iree that received 1 comment

Deprecate LLVMGPUWarpReduction pipeline

TODO(newling)

1 comment

Opened 1 other issue in 1 repository

iree-org/iree 1 open

[LLVMGPU] Numerical error: softmax with dynamic K
This contribution was made on Jul 23