CUDA will also work with AMD GPUs
It looks like the industry has managed to break Nvidia’s software stack dominance. Previously, CUDA libraries were owned by AMD. ROCmWe’ve seen the emergence of ZLUDA, an open source migration project that makes it work with . And now Nvidia’s CUDA capabilities allow the creation of high-end hybrid models. SCALE A new tool called is entering the scene.
CUDA is used to leverage the power offered by GPUs for advanced simulation, physics calculations, scientific modeling, crypto mining, and yes, machine learning. Amid the AI boom, companies have opted for Nvidia’s GPUs due to the strong case for machine learning. Widely viewed as the industry standard, this had caused competitors like AMD to look elsewhere for software on their GPUs.
According to what has been reported, SCALE does not require changing the CUDA program or build system. Making a statement on this subject, Spectral Compute CEO Michael Sondergaard said, “We believe it should be possible to write code once and build/run it on any hardware platform. “This has been a reality for CPU code for many years, so why not for GPUs as well?” he said.
Spectral Compute’a göre SCALE has been in development for seven years and it doesn’t rely on Nvidia’s code but builds into a CUDA-compatible toolchain, making SCALE highly adaptable across multiple platforms like AMD’s RDNA GPUs. So SCALE works similarly to CUDA Toolkit and simply emulates the toolset. Thus, tools and scripts can be run directly. Spectral Compute also says that they have successfully tested SCALE with applications such as Blender, Llama-cpp, XGboost, FAISS, GOMC, and more, all running on AMD GPUs.
Source
https://www.tweaktown.com/news/99396/scale-tool-lets-nvidia-cuda-applications-run-on-amd-radeon-gpus-without-modifying-code/index.html
https://docs.scale-lang.com/
This news our mobile application Download using
You can read it whenever you want (even offline):
Source link: https://www.donanimhaber.com/amd-ye-gun-dogdu-nvidia-nin-cuda-hakimiyet-scale-ile-bitecek–179739