5 ms·
Exactly my thoughts - instead of one datacenter class gpu, you can use multuiple desktop components for better performance and more performance per buck. I've s
by altmind 8y ago
Exactly my thoughts - instead of one datacenter class gpu, you can use multuiple desktop components for better performance and more performance per buck. I've seen some sweet supermicro 2U chassis that can fit and run cool four 2080Ti.
I remember, though, that NVIDIA disallows in EULA the use of nvidia driver blob in consumer gpu in datacenter environment. Time will show how legally enforcible is this restriction.
- freeone3000 8y agoWe have one of these. Aside from the mechanical differences (it's actually a 2.5U server due to how the power cables connect to consumer cards), you're losing out on NVLink. You're going to be restricted to single-card training, which means either batching or other restrictions to ensure it fits in the memory of a single card. This setup is able to train four one-card models in parallel; it is not able to train a four-card model.
- riku_iki 8y agoNVLink supports 2080 Ti
- freeone3000 8y agoThere's a physical connector also called NVLink, yes, but the cards don't present as unified at the driver level; nvidia-smi shows "link: off" even using the bridge. It's effectively SLI, which can reduce memory bandwidth load and cross-card transfer costs, but doesn't unify the cards the way Volta (actual) NVLink does.