According to artificial intelligence research group OpenAI, Nvidia’s graphics processing units are too difficult to program, even with Nvidia’s own programming tool, CUDA.
The San Francisco-based AI startup, which is sponsored by Khosla Ventures and Microsoft, just released the 1.0 version of Triton, a new programming language created to alleviate that burden, and inserted a link to the source code on GitHub.
According to OpenAI, Triton can provide significant ease-of-use improvements over coding in CUDA for some neural network activities at the heart of machine learning types of AI, such as matrix multiplies.
To Read More: ZDNet
Check Out The New Enterprisetalk Podcast. For more such updates follow us on Google News Enterprisetalk News.
The platform covers e entire enterprise technology space- including emerging technologies like RPA, AI, cloud, automation, and the entire gamut of digital transformation tools, strategies and management decisions.
