Radeon Vega Frontier Edition for Data Scientists and ML engineers.
Today AMD announced the first of several Vega architecture-based products, Radeon Vega Frontier Edition: http://pro.radeon.com/en-us/vega-frontier-edition/
Machine Learning application:
The Radeon™ Vega Frontier Edition, combined with our ROCm open software platform, delivers industry-leading performance in deep learning. Developers can now tap into the power of the Vega architecture to do machine learning algorithm development before deploying it to massive servers that will be equipped with Radeon Instinct accelerators. In the initial testing (early clocks and drivers) we did against the fastest competitive card we could get our hands on, we found Radeon Vega Frontier Edition graphics cardsto be more than 30% faster as measured by the DeepBench Benchmark 1 – and also delivers a disruptive performance per dollar equation that enables AMD’s leadership in compute TCO for the foreseeable future.Empower Scientists to Explore New Frontiers
Training techniques used today on neural networks in machine intelligence and deep learning applications have become very complex and require the handling of massive amounts of data when training those networks to recognize patterns within that data. This requires lots of floating point computation spread across many cores, and traditional CPUs can’t handle this type of computation as efficiently as GPUs handle it.
Radeon Vega v. NVidia Titan Xp benchmark: