What Should I Build?

A directory of what people actually want. Classified, clustered, ranked and updated daily

Dynamic autoscaler for heterogeneous ML compute (GPU/CPU/RAM)

Other · 1 mentions

#1992399734783533519

Not sure why folks are deploying machines with 4 H100 and then backing it with a TiB of RAM and a bazillion core of CPU. Except for in very short bursts, the workload is going to swing one way or the other. There must be a better way to optimise for this ?

For any inquiries, contact info@quantumedge.sk