Skip to content

[gpu][alpaka] Add initial integration of ALPAKA#2

Open
sanjibansg wants to merge 61 commits into
mainfrom
gpu/alpaka
Open

[gpu][alpaka] Add initial integration of ALPAKA#2
sanjibansg wants to merge 61 commits into
mainfrom
gpu/alpaka

Conversation

@sanjibansg

Copy link
Copy Markdown
Member

This PR adds initial implementation to generate GPU inference code with ALPAKA from SOFIE.

sanjibansg and others added 10 commits July 13, 2026 12:12
* GTests for Conv GPU groups>1 inference

Co-authored-by: Sanjiban Sengupta <sanjiban.sg@gmail.com>
…alAveragePool) (#27)

* 2d maxpool gpu kernel and test

* empty codegen for unsupported pool variants

* 1d and 3d maxpool gpu kernels and tests

* added AvgPool and GlobalAvgPool GPU kernels + tests

* test: port Pool GPU tests into test/alpaka structure
* fix: wrong pad indices in pool for asym padding

* test: use random input for MaxPool2d_AsymPad and drop pre fix comment
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants