I do see the prefix sum is defined and used in the Cuda functions I may overlook something, but I did not find the parallel prefix sum in the python implementation selective_scan_ref. Is it only for ...
This project originated from 3b1b.co/talent (Grant Sanderson's talent search), which led to mlpuzzles.com - a GPU kernel optimization challenge created by METR (Model Evaluation & Threat Research).
Some results have been hidden because they may be inaccessible to you
Show inaccessible results