llvm/execution/cuda: Always use pycuda ArgumentHandler as cached buffer

PrincetonUniversity / PsyNeuLink

A block modeling system for cognitive neuroscience

https://psyneulink.org

Apache License 2.0

87 stars 31 forks source link

Closed jvesely closed 3 months ago

jvesely commented 3 months ago

Only the ArgumentHandler class provides the .array member that is used to check if a gpu_buffer is not stale.

Fixes: 36e52b8ceb50d69c04d8e9aa81cd2139f27f01ee ("llvm/execution/cuda: Use numpy argument handlers to access compiled structures")

github-actions[bot] commented 3 months ago

This PR causes the following changes to the html docs (ubuntu-latest-3.11):

No differences!

...

See CI logs for the full diff.

github-actions[bot] commented 3 months ago

This PR causes the following changes to the html docs (ubuntu-latest-3.11):

No differences!

...

See CI logs for the full diff.