Benchmarking dataloading

[itamar_kanter]

I’m trying to benchmark different datasets object with torch.utils.benchmark.
I have noticed that I can control the size of threadpool through the num_threads param.
In torch.utils.data.DataLoader there exists is the num_workers param.
What is the diiffernce between thus two params; num_workers and num_threads?
I have noticed that changing num_threads didn’t affect the benchmarking results

[calwritescode (_cal)]

Maybe this thread helps!

DataLoader num_workers vs torch.set_num_threads

Is there a difference between the parallelization that takes place between these two options? I’m assuming num_workers is solely concerned with the parallelizing the data loading. But is setting torch.set_num_threads for training in general? Trying to understand the difference between these options. Thanks!