Last month, Huawei unveiled a new AI server cluster in China's Anhui province powered by its in-house Ascend chips, not the dominant GPUs from NVIDIA. Power distribution architecture supports 2N, DR, and BR. Power distribution. Diving a bit into the specifications reported by Huawei, it is claimed that the Atlas 950 SuperPoD will feature 8,192 of the Ascend 950 AI chips, and they will bring in a cumulative performance of eight EFLOPS FP8 and 16 EFLOPS FP16 with a total interconnect bandwidth of a whopping 16. The system delivers 8 EFLOPS in FP8 precision and 16 EFLOPS in FP4 precision, with 1,152 TB of total memory. Although it costs three times more, and uses 3. So China can resource internally all the computing power it needs to pursue AI development. This development, alongside reports of performance gains and a growing domestic ecosystem, raises questions about whether US curbs are effectively. Find local businesses, view maps and get driving directions in Google Maps.
[PDF Version]