Experiences with the Nvidia A30?

Hi everyone. Just curious if anybody who does ML/AI or HPC work has had experiences with the Nvidia A30? The power consumption, street price, and memory bandwidth are rather appealing, but I’m worried that it might be in an awkward worst-of-both-worlds category where it’s better to either save the money and go with like a 3090/4090, or cough up the extra money for an A100/H100?

Let’s disregard FP64 at the moment as well. For me, it’s a nice to have feature (particularly for precision sanity checks), but it’s hardly critical.

This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.