Nvidia’s CUDA Language Can Now Run On AMD, Intel GPUs
Los Angeles-based cloud rendering company OTOY has manged to reverse engineer Nvidia’s general purpose graphics processing unit (GPGPU) software, known as CUDA, to run across non-Nvidia hardware platforms. In other words, programs written in the CUDA language can now run on GPUs manufactured by Intel, AMD, and ARM.
The technology that is used in porting the CUDA language to other than Nvidia’s own GPUs, basically creates a single CUDA code base that can run across multiple platforms.
“We have been able to do this without changing a line of CUDA code, and it runs on AMD chips,” OTOY chief executive Jules Urbach told GamesBeat. “You can now program once and take CUDA everywhere. AMD has never really been able to provide an alternative.”
The tool will work on a multitude of devices including AMD-based PlayStaton 4 and Xbox One consoles, as well as Apple’s iPad and iPhone.
Although there is an independent GPGPU standard on the market, dubbed OpenCL, it isn’t necessarily as good as NVIDIA’s CUDA technology, believes Urbach. He said CUDA is “superior” and enables much richer graphics software.
OTOY research and development team was able to build the CUDA “cross compiler” in 9 weeks, which “is a big breakthrough from our point of view,” said Urbach.
OTOY’s tool for the CUDA compiler will be made available within the company’s Octane 3.1 rendering engine sometime this summer.