News for hackers. Stuff that matters.

A comment on Intel Announces Aurora GenAI, ChatGPT Competitor with 1T Parameters. Parent comment

Re:Intel Announces Aurora GenAI, ChatGPT Competitor with 1T ... by mirekrusin on Monday May 22, 2023 @02:30PM (#36034962)
...which is equivalent to just 62x H100 GPUs from Nvidia?
[ Parent | on HN ]
Re:Intel Announces Aurora GenAI, ChatGPT Competitor with 1T ... by AnthonyMouse on Monday May 22, 2023 @06:28PM (#36037490)
One H100 has the same memory bandwidth as the two CPUs in each of the compute nodes. But the supercomputer has >9000 compute nodes, so 62x H100 GPUs might fall short there.

They also might not be measuring the same thing. H100 GPUs have "3026 TFLOPS":

https://www.pny.com/nvidia-h100

...at FP8. 51 TFLOPS at FP32. Supercomputer lists typically use FP64.

[ Parent | on HN ]