AMD readies Zen’s server software

ROCm writes and compile parallel programs

AMD has released the latest version of its ROCm software tools which make it easier to write and compile parallel programs for its new Zen GPUs and CPUs.

The software is designed to help put Zen under the bonnet of high-performing servers to turn GPUs and CPU combos into servers. If it all pays off AMD could be back in the server market after losing it totally to Chipzilla.

ROCm provides a base for the company to build GPUs for large-scale servers. It is a low-level programming framework like Nvidia’s CUDA. But it’s open source and can work with a wide range of CPU architectures like ARM, Power, and x86.

According to PC World, the ROCm platform is targeted at large-scale server installations and for multiple GPUs in a cluster of racks.

It’ll work with AMD’s latest Radeon Pro GPUs and current consumer GPUs based on the Polaris architecture. It can be used to run neural networking clusters or for scientific computing.

AMD has not revealed any of its supercomputing GPU plans but said ROCm will play a big role as the company goes after the HPC space.

ROCm is based around the Heterogeneous System Architecture (HSA) spec which are supposed to link the computing power of CPU, GPU, and other processors in a system. AMD thinks HSA specifications could replace OpenCL, which is widely used today for parallel programming.

But what is more interesting about it is that AMD is chasing open-source standards, contrary to Intel which still wants people to use its proprietary standards. This open saucy approach might be the novelty which helps AMD succeed. Open Source does well in the HPC area where stuff is a little more colaborative.  It might be that AMD has hit on a system that works and can get its foot in the door.