3 ms·
While this sounds attractive, OpenCL performance just doesn't seem to be there at the moment. In CNN benchmarks[1] the CL based implementations consistently fi
by nl 11y ago
While this sounds attractive, OpenCL performance just doesn't seem to be there at the moment.
In CNN benchmarks[1] the CL based implementations consistently finish at the bottom. I think this is mostly because of the weak nVidia OpenCL implementations, but that doesn't help, since most people use nVidia (eg, Amazon GPU instances are nVidia based)
[1] https://github.com/soumith/convnet-benchmarks https://github.com/soumith/convnet-benchmarks