4 ms·
> Say your success rate for replicating CUDA+Nvidia[...] Rational hyperscalers would just stop as soon as their tooling/workloads/models are functional on AMD
by sangnoir 2y ago
> Say your success rate for replicating CUDA+Nvidia[...]
Rational hyperscalers would just stop as soon as their tooling/workloads/models are functional on AMD hardware within an acceptable perf envelope - just like they already do with their custom silicon. Replicating CUDA is just unnecessary, expensive and time-consuming completionism; if some workloads require CUDA, they will be executed on Nvidia clusters that are part of the fleet.