Accelerating ResNet-50 Training on the IPU:  Behind our MLPerf Benchmark

Accelerating ResNet-50 Training on the IPU: Behind our MLPerf Benchmark

A technical guide on efficient hardware scaling, memory optimization strategies and performance tools

Written by: Dr. Mario Michael Krell, Zhenying Liu, Emmanuel Menage, and Bartosz Bogdanski

Graphcore engineers delivered outstanding performance at scale for the latest MLPerf v1.1 training results published in December 2021 [9], with our IPU-POD₁₆ outperforming Nvidia’s flagship DGX A100 on ResNet-50.

Here, we will explain how our team achieved these accelerations at …

artificial intelligence computer vision machine learning mlperf resnet50 training