NVIDIA announces that it will make CUDA compatible with RISC-V

The RISC-V platform has just signed one of the most important partnerships it could to become even more popular. At the Symposium of Architecture to China, Nvidia has sent a representative to confirm that she intends to support her Cuda ecosystem in Open-Source architecture processors.

It is an even unexpected ad, as Nvidia is somewhat closed with its Cudda platform and RISC-V processors are not yet found in the IA processing data centers that have made Nvidia the richest company in the world. Still, the port will allow a type CPU to be the main processor of CUDA -based AI systems.

GeForce’s owner seems to see a future in which we will have immense systems also using Risc-V, and certainly the port of her ecosystem is the kind of thing that will encourage us to get there one day. However, in the introduction of support, projects such as Nvidia Jetson can already benefit, opting for CPUS RISC-V instead of X86 or ARM.

We should not ignore that Nvidia’s announcement was made at a RISC-V symposium in China. According to Tom’s hardware, architecture has been rapidly growing in the country, so there are many potential users who will celebrate the ad.

At the same time, Nvidia has needed to work “anchored” while trying to uphold US government sanctions that prevent the company from exporting some of its most profitable products to China. We recently had the revelation that at least H20 accelerators could be sold there again.

Read more:

Thus, it is very likely that the largest company in the world wants to expand its presence in less advanced systems as well, with hardware that it is allowed to sell in the Asian country. Supporting open source processors such as RISC-V in these cases should help Nvidia bring more customers to the Cuda ecosystem.