This blog post spends a lot of time telling why gang scheduling is needed here (a fact obvious to anyone using the nvl72s), but doesn't actually go into detail about how their scheduler works.
At work, we have built something very similar, a custom topology aware scheduler for workloads, and the most interesting things about such a project are entirely absent from this article, making it more of a fluff/marketing piece than an actually interesting technical blog in this area.
Our next iteration of this scheduler aims to scale to O(millions) of GPUs, with a scheduling cycle time measured in the milliseconds.
This blog's goal is to introduce the problem around the new NVL72 system with GPU failures in production, which not all people outside of xAI are aware of. We have more blogs coming up for what we built to handle topology-aware scheduling, defragmentation and hot swapping during failures. The short version of the scheduling piece: jobs are decomposed into 16-node slices placed atomically on a single rack, with two spares per rack reserved for hot swaps.
Seems very relevant to what you are building. Quoting what Ali said: "Quick PoC?"
Feels AI-generated, so much text real estate is spent explaining the basics in sufficient detail then just two paragraphs on their scheduler with a lot of hand-waving.
At work, we have built something very similar, a custom topology aware scheduler for workloads, and the most interesting things about such a project are entirely absent from this article, making it more of a fluff/marketing piece than an actually interesting technical blog in this area.
Our next iteration of this scheduler aims to scale to O(millions) of GPUs, with a scheduling cycle time measured in the milliseconds.
Seems very relevant to what you are building. Quoting what Ali said: "Quick PoC?"
/genuine do you recommend any other blog?