← All creators
S
Stuart Sul
0 recommended products
Everything Stuart personally uses, recommends, or has created — plus things they don't recommend — sourced from their own show and appearances on other podcasts.
Recent episodes
All →Showing 5 of 6 products
Clear filtersMentioned in 1 episode
"Together AI is also using Parallel Kittens to optimize its inference workloads."
Mentioned in 1 episode
"Parallel Kittens builds on all the trade-offs and principles that we discussed and it's a highly opinionated set of CUDA programming primitives that extends Thunder Kittens, which is one of our previous works for single GPU kernels."
Mentioned in 1 episode
"Over the past few years the ML systems community has devoted a lot of effort into single GPU efficiency. For instance we have IO-aware algorithms like Flash Attention, linear attention models, mega kernels and many many programming libraries and frameworks for writing efficient kernels."
Mentioned in 1 episode
"We have NVLink which allows direct communication of data between the GPUs. And that's essentially what Nvidia DGX machines are all about."
Mentioned in 1 episode
"You also have scale-up architectures like NVL72 which packs 72 GPUs inside a single NVLink domain and soon this is going to extend to hundreds of GPUs."
0 recommended · 0 sponsored · 0 warned against · 0 episodes