You might not have heard about AMD’s ROCm. It (unofficially) stands for Radeon Open Compute PlatforM—a pretty terrible acronym—and it’s pronounced “rock ’em.” This is a way for programs to take advantage of the computing power in a graphics card instead of a CPU. Think of it as a software accelerator, sort of like AMD’s version of Nvidia CUDA. And it’s about to become a lot more relevant.
PCWorld contributor Will Smith got to speak with Andrej Zdravkovic, Senior Vice President of GPU Technologies at AMD. While ROCm is mostly relevant for large-scale “big iron” enterprise applications right now, the latest changes are making it more relevant for regular Windows users.
How so? Well, the new HIP (Heterogeneous Interface for Portability) SDK. It’s a bit of special sauce that lets programs designed to use CUDA and similar systems tap into ROCm’s power to leverage Radeon graphics cards. And it has the potential to be a huge game-changer. Basically, any local program that needs extra power—from AI applications to rendering to file processing and beyond—can benefit from this, with only small tweaks needed to get existing code up and running in most cases.
Check out the full video interview above for the technical ins and outs. And for more deep dives into the latest PC tech, be sure to subscribe to PCWorld and The Full Nerd Network on YouTube.
