PlaidML Deep Learning Framework Benchmarks With OpenCL On NVIDIA & AMD GPUs
Written by Michael Larabel in Graphics Cards on 14 January 2019 at 03:30 PM EST. Page 3 of 4. 5 Comments.
For many of the inference benchmarks, the Radeon RX Vega 64 was generally running in line with the GTX 1070.
The FP16 performance was good out of the NVIDIA RTX Turing cards but not as significantly better as we've seen out of some other FP16 workloads.