Open sumo43 opened 5 months ago
I think they'll need some more information here. How big is your dataset? For small samples, the speeds could be comparable because of the overhead spent in swapping memory from RAM to VRAM and viceversa. For big matrices, this shouldn't be the case. That's just the minimal info I would require for tackling something like this, if you could provide a full working code I guess we could understand better what's happening in your case.
Hello,
I tried running hellokan.ipynb on cpu and gpu, and the speed is similar. What are some places where parallelization can be introduced to improve performance?