covi
Joined 14 years ago
Karma
203
Submissions
110
Member since
Sep 5
https://github.com/skypilot-org/skypilot/
View on Hacker NewsRecent submissions
Comment · 3 days ago
Why do you target millisecond-level cycle time? Are you looking at very fine-grained tasks like elastic sandboxes, in addition to beefy pre- and post-training jobs?
Comment · 7 months ago
This feels like the chimpanzee with a power drill. An agent is honestly just brute-force search, but guided.
Comment · 8 months ago
The post says Slurm supports gang scheduling, k8s doesn't (out of the box).
Comment · 1 year ago
Take a look at SkyPilot. Good for running these batch workloads. You can use spot instances to save costs.
Comment · 1 year ago
To massively increase the reliability to get GPUs, you can use something like SkyPilot (https://github.com/skypilot-org/skypilot) to fall back across regions, clouds, or GPU choices. E.g.,
$ sky launch --gpus H100
will fall back across GCP regions, AWS, your clusters, etc. There are options to say try either H100 or H200 or A100 or <insert>.
Essentially the way you deal with it is to increase the infra search space.
Comment · 1 year ago
Related: https://skyplane.org/en/latest/ (mentioned in OP)
From what I know this idea underpins a few FAANG-level companies' data transfer systems. OP's value = a simple implementation of the idea that's OSS and applied to AI.