I would have expected that going from one node (which can't hold the weights in RAM) to two nodes would have increased inference speed by more than the measured 32% (21.1t/s -> 27.8t/s).
With no constraint on RAM (4 nodes) the inference speed is less than 50% faster than with only 512GB.
Am I missing something?
I don't think that's true. At least not without heavy performance loss in which case "just be memory mapped" is doing a lot of work here.
By that logic GPUs could run models much larger than their VRAM would otherwise allow, which doesn't seem to be the case unless heavy quantization is involved.
You'd need to be in a weirdly compute-limited situation before you can replace significant amounts of RAM with SSD, unless I'm missing something big.
> MoE architecture should help quite a bit here.
In that you're actually using a smaller model and swapping between them less frequently, sure.
MoEs is great for distributed deployments, because you can maintain a distribution of experts that matches your workload, and you can try to saturate each expert and thereby saturate each node.
With a cluster of two 512GB nodes, you have to send half the weights (350GB) over a TB5 connection. But you have to do this exactly once on startup.
With a single 512GB node, you'll be loading weights from disk each time you need a different expert, potentially for each token. Depending on how many experts you're loading, you might be loading 2GB to 20GB from disk each time.
Unless you're going to shut down your computer after generating a couple of hundred tokens, the cluster wins.