Ahead of the launch of the next AMD Ryzen 3000 series, the company has also revealed its new HIP SDK for Windows.
This helps developers run CUDA applications on AMD hardware, and I’ve been working with the software for a little while now.
This is a big deal for AMD, as it helps the company compete with Nvidia in the GPU computing space.
The HIP SDK bridges the divide between Nvidia’s CUDA API and AMD’s open-source HIP API, meaning developers can convert CUDA applications into C++ code compatible with both Nvidia and AMD graphics cards.
The HIP SDK is part of AMD’s ROCm platform, which is designed to help developers tackle high-performance and artificial intelligence (AI) workloads.
What’s neat about the software is that it helps developers take advantage of AMD hardware that would otherwise go unused.
I’ve been playing around with the HIP SDK for the past few weeks, and it’s extremely easy to get started with.
Porting a CUDA application to HIP SDK is not challenging since both are based on C++.
However, you will still have to ensure your code is compatible with the AMD HIP API.
There are a few tools that come with the HIP SDK that can help with this, namely the HIPIFY toolset.
This allows you to convert CUDA code into portable HIP C++, which you can then compile into an executable file with a click of a button.
While this is great for developers, I’m not sure I’d recommend the HIP SDK for people looking to automate the graphics card mining process.
It doesn’t automatically optimize code, so you’ll still need to manually optimize your code before executing it.
The HIP SDK is compatible with 32-bit and 64-bit Windows operating systems, including Windows 10 and 11.
It also supports a range of graphics cards, including workstations, mobile gaming, and even the newly-launched AMD Ryzen Embedded 3200 series.
AMD is also updating the compatibility list, and there are ten Radeon graphics cards listed at the time of writing.
Blender HIP has actually embraced AMD Radeon graphics cards since the Vega days, so it’s nice to see the company finally deliver a compatible GPU computing SDK for Windows.
You can download the AMD HIP SDK here.
The milestone release is a big win for the company, and I’m looking forward to seeing what other features and updates it plans to add.
