gpu-algorithms-labs
gpu-algorithms-labs copied to clipboard
illinois-impact
IMPACT GPU Algorithms Teaching Labs
Broadly: * each lab is in its own folder, with its own self-contained resources. Any folder should stand alone to facilitate reuse among our collaborators * create a standalone `.cu`...
https://github.com/illinois-impact/gpu-algorithms-labs/blob/b283d976996ed54bab96e36452ef0260b6d7d7a8/labs/sgemm-regtiled-coarsened/helper.hpp#L64 It's not used.
to be consistent with the class lectures
https://github.com/illinois-impact/gpu-algorithms-labs/blob/490e9f8060457634df553d8897f60bfaa49179e6/labs/tiled_conv/rai_build.yml#L4 This rai job should be launched with the following parameters: ```yaml gpu: count: 1 ```
https://github.com/illinois-impact/gpu-algorithms-labs/blob/490e9f8060457634df553d8897f60bfaa49179e6/labs/basic_conv/main.cu#L180 Students will often launch a kernel with one of these dimensions as a kernel parameter, which will cause their launch to fail. I think the pedagogical value of that...