1
'I paid for the whole GPU, I am going to use the whole GPU'
A guide to maximizing the utilization of GPUs, from cloud allocations to FLOP/s.
Modal Blog
Cloud infrastructure for teams that develop, train, and serve AI applications at scale.
- Popular posts
- 5
- Total points
- 678
- HN comments
- 272
- Rank
- #1,986 company
Visit modal.com → On Hacker News Mar 2024 – Jan 2026
Infrastructure & DevOps 4AI & Machine Learning 3Software Engineering 2Networking & Internet 1
2
Keeping 20k GPUs healthy
How we do active and passive monitoring on hyperscalers and neoclouds.
3
We reverse-engineered Flash Attention 4
Asynchrony, fast approximate exponents, and 10x more efficient softmax.
4
Lambda on hard mode: serverless HTTP in Rust
In this post, we'll talk about how Modal handles real-time HTTP requests and WebSockets in serverless functions.
5
Static IPs for Serverless Containers
Now your Modal containers around the world can have static outbound IPs. Featuring WireGuard, policy-based routing, and NAT.