Skip to content

[CodeGen][OpenCL] Limit OpenCL built-in vector lanes to 2, 3, 4, 8, 16.#7777

Merged
tqchen merged 1 commit intoapache:mainfrom
xndcn:opencl
Apr 1, 2021
Merged

[CodeGen][OpenCL] Limit OpenCL built-in vector lanes to 2, 3, 4, 8, 16.#7777
tqchen merged 1 commit intoapache:mainfrom
xndcn:opencl

Conversation

@xndcn
Copy link
Copy Markdown
Contributor

@xndcn xndcn commented Mar 31, 2021

OpenCL does not support any vector lanes except for 2, 3, 4, 8 and 16.

P.273 https://www.khronos.org/registry/OpenCL/specs/opencl-2.0.pdf
OpenCL

@tqchen tqchen merged commit 3825161 into apache:main Apr 1, 2021
@tqchen
Copy link
Copy Markdown
Member

tqchen commented Apr 1, 2021

Thanks @xndcn !

trevor-m pushed a commit to trevor-m/tvm that referenced this pull request May 6, 2021
trevor-m pushed a commit to trevor-m/tvm that referenced this pull request May 6, 2021
trevor-m pushed a commit to trevor-m/tvm that referenced this pull request May 6, 2021
trevor-m pushed a commit to neo-ai/tvm that referenced this pull request May 11, 2021
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.

2 participants