← All IntelClip / AI AgentsLatency: 500GB to 500MB
From Taking Reinforcement Learning Cross Datacenter — Nan Jiang, Modal · ≈17:14
The end-to-end result — cross-region weight sync drops from hours to seconds, which is what makes non-RDMA rollout capacity viable.
What’s in it
- The end-to-end result — cross-region weight sync drops from hours to seconds, which is what makes non-RDMA rollout capacity viable.
Clip transcript
support this aggregated roll out. So this is more like a system latency analyst incluster way syncing is fast because they have RDMA. Uh a full checkpoint across regions through network is pretty slow and but if we use the delta if we use exactly what we described previously we can decrease the number of like transfer transfer size from like 500 gigabytes to 5 500 megabytes. So it will be like extremely fast in seconds.
Comments
Checking sign-in…
Loading comments…