Together, the Radeon Instinct accelerators and open-source software solutions are designed to dramatically increase performance, efficiency, and ease of implementation of deep learning workloads.
Radeon Instinct accelerators are designed to address a wide-range of machine intelligence applications:
• The Radeon Instinct MI6 accelerator, based on the acclaimed Polaris GPU architecture will be a passively cooled inference accelerator optimised for jobs/second/Joule with 5.7 TFLOPS of peak FP16 performance at 150W board power and 16GB of GPU memory
• The Radeon Instinct MI8 accelerator, harnessing the high-performance, energy-efficient “Fiji” Nano GPU, will be a small form factor HPC and inference accelerator with 8.19 TFLOPS of peak FP16 performance at less than 175W board power and 4GB of High-Bandwidth Memory (HBM)
• The Radeon Instinct MI25 accelerator will use AMD’s next-generation high-performance Vega GPU architecture and is designed for deep learning training, optimised for time-to-solution
AMD also announced MIOpen, a free, open-source library for GPU accelerators to help solve high-performance machine intelligence implementations. The library is planned to be available in early next year to provide GPU-tuned implementations for standard routines like convolution, pooling, activation functions, normalisation and tensor format. This open-source software offering builds on AMD’s commitment to open source solutions and the ROCm platform revealed in November.