Chinese Biren Technology is preparing to produce its first 7nm GPU

Chinese Biren Technology is preparing to produce its first 7nm GPU

NVIDIA and AMD continue to dominate the GPU market, but Chinese manufacturer Biren Technology hopes to change that as the company is on track to begin production of its first 7nm GPU during the third quarter of this year. . In a recent interview (as reported by CnTechPost), the company stated that its compute GPU will compete with NVIDIA's next generation 5nm competitors (“Hopper” and “Lovelace”).

This will be a GPU compute designed for AI workloads, therefore designed for both training and inference. Biren will send the design to production in the third quarter of this year, while it will debut on the market in 2022. Biren's chip will be made on the 7nm node, but it is unclear whether the developer will use the Taiwan-based TSMC or its own. rival, Samsung Foundry, based in Korea. Given the typical amount of time companies spend manufacturing their chips (assuming everything works as expected), it is reasonable to expect Biren Technology to deploy its compute GPU compute in the third or fourth quarter of 2022.

Biren Technology received 170 million in funding in 2020. However, the development and implementation of a moderately complex FinFET-based 7nm processor costs around $ 297.8 million, according to estimates by International Business Strategies (cited by SemiEngineering), while a highly complex project can cost up to $ 1. , $ 5 billion, according to IBS.

Interestingly, Biren's CTO, as well as Chief Architect, Mike Hong said in an interview with Chinese media that the performance of Biren's chip is able to rival that of the 5nm GPU compute. NVIDIA. Of course, being competitive does not mean offering superior performance, but rather coexisting on the market. Furthermore, there is no way that Hong will know what NVIDIA's Hopper or Lovelace architectures are really capable of.

Biren Technology was founded in 2019 to develop GPUs suitable for artificial intelligence, high computing performance (HPC) and graphics rendering, we read in a description on the company's website. Biren did not plan to immediately create a universal GPU, so the first product will only be designed for cloud AI workloads (that is, with a focus on INT8 and FP16 processing). However, it will take advantage of “high-end packaging technology” to tackle demanding applications. Finally, Biren will switch to HPC applications, followed by a GPU that will be used for AI, HPC and graphics.

Biren will be the second Chinese company to use a 7nm node to build a GPU for AI and HPC. The first was Tianshu Zhixin with his Big Island GPU.

Don't miss out on this incredible offer: the Bqeel kit including Raspberry Pi 4 Model B 2GB RAM, 64GB MicroSD, 5.1V 3A Type-C power supply , fan, Micro HDMI cable, USB card reader and black protective case is available on Amazon at a discounted price!

Powered by Blogger.