Vote count:
0
So if I run an openCL Kernel function on a CPU device and not a GPU device, does the kernel automatically make use of all the cores on the CPU? For instance my system says I have 4 cores on the CPU will the kernel make use of all 4 cores automatically?
If the above is true, then does it mean that running an openCL kernel on a single CPU device would be equivalent to using standard std::threads (assuming we're using C++) to carry out the same tasks?
I ask because on my current computer openCL seems to only be able to access one CPU and no GPUs. So if I were to use openCL to parallelize my code it seems like it might be overkill if it will essentially be doing the same thing as an std::thread based implementation.
asked 1 min ago
OpenCL on CPU Device - what happens under the hood?
Aucun commentaire:
Enregistrer un commentaire