PlaidML Deep Learning Framework Benchmarks With OpenCL On NVIDIA & AMD GPUs

Pointed out by a Phoronix reader a few days ago and added to the Phoronix Test Suite is the PlaidML deep learning framework that can run on CPUs using BLAS or also on GPUs and other accelerators via OpenCL. Here are our initial benchmarks of this OpenCL-based deep learning framework that is now being developed as part of Intel's AI Group and tested across a variety of AMD Radeon and NVIDIA GeForce graphics cards.

Over the weekend I carried out a wide variety of benchmarks with PlaidML and its OpenCL back-end for both NVIDIA and AMD graphics cards. The Radeon tests were done with ROCm 2.0 OpenCL and it was working out fine there without any troubles while also working fine with NVIDIA's OpenCL driver stack. Benchmarks were done with a variety of neural networks, both training and inference, etc.

The graphics cards available for testing included the 16 following GPUs:

- RX 580
- RX 590
- RX Vega 56
- RX Vega 64
- GTX 980
- GTX 980 Ti
- GTX 1060
- GTX 1070
- GTX 1070 Ti
- GTX 1080
- GTX 1080 Ti
- RTX 2060
- RTX 2070
- RTX 2080
- RTX 2080 Ti
- TITAN RTX

All of the tests were run from an AMD Ryzen Threadripper 2990WX workstation with ASUS ROG ZENITH EXTREME motherboard, 4 x 8GB DDR4-3200 memory, and Samsung 970 EVO 500GB NVMe SSD. Ubuntu 18.10 was running on the system with the Linux 4.20.0 kernel and GCC 8.2 compiler.

These PlaidML benchmarks were carried out using the Phoronix Test Suite. Coming up later this week will be PlaidML CPU benchmarks across a variety of operating systems.