Twitxy Saved

Dylan Lim

Verified account Together AI

@dylan__lim

Stanford, CA dylan-lim.com Joined April 2024
519 Following 469 Followers 43 Posts

Latest posts 5

Pinned post

Dylan Lim Verified account @dylan__lim
A little over a year ago, I started my journey into kernels and inference optimization. Since then, the field has rapidly evolved, and many tasks like kernel writing have become increasingly obsolete. I’ve spent a lot of time reflecting on ML Systems in the agentic era, and how such systems must be designed to efficiently and thoroughly explore the optimization space. I’m proud to announce Together Compiler, a continuous optimization system for the age of polymorphic inference!
Simran Arora Verified account @simran_s_arora 3d
Introducing the Together Compiler: continual compilation for AI inference! I’m super proud to finally share this work from our research team! ✨

Today’s inference is “Polymorphic” or shape-shifting, it doesn’t sit still. On OpenRouter, 80% of traffic goes to models released in the last three months and the top ten most popular open models each use a different architecture. Our customers constantly change their usage behaviors. Together Compiler leverages agents to continually optimize our deployments immediately as these changes occur.

Together Compiler research highlights include:
1. 252x better throughput: TKCC lets agents try 252x more optimization experiments per hour compared to prior agent inference optimization systems.
2. State-of-the-art TPS across InferenceX workloads we tried: TKCC outperforms the best InferenceX engine across all tested settings in just 24 hours & 2 nodes per compile. Including 1111 TPS on MiniMax M3.1 B200 and 1205 TPS on DeepSeek v4.1 Flash GB300 at batch size 1. NVIDIA, AMD, Trainium2.
3. Continuous recompilation: TKCC currently serves many trillions of production tokens per day and adapts to traffic shifts within just a few hours of recompilation.
4. Bitter lessons: TKCC’s design builds on 2+ years of building frameworks to accelerate AI Inference, starting from ThunderKittens and Megakernels. Compilers have always used abstractions and IRs, but we show that agentic compilers now reach higher performance ceilings without abstractions.

From an incredible team @reyna_abhyankar @dylan__lim @realDanFu @togethercompute!!
View quoted post on X
8 replies 17 reposts 118 likes 7.2K views View on X
Dylan Lim Verified account @dylan__lim
One of the best blogs out there on writing AMD kernels!
Simran Arora Verified account @simran_s_arora Sep 21
we got early access and hipkittens is now running on amd helios mi455! here we share an educational gemm ladder to demonstrate how to use the new helios features.

overall, we found that the patterns identified in hipkittens for writing performant mi350/355 kernels translate well to helios making the forward port quite smooth!
View quoted post on X
0 replies 0 reposts 14 likes 1.5K views View on X
Dylan Lim Verified account @dylan__lim
ThunderKittens is now on Vera Rubin! It's been a fun past week playing with new kernels and hardware. Check out the blog below for more details on how we pushed our GEMMs to cuBLAS levels of performance! 🐱⚡️🔭
Together AI @togethercompute Sep 10
thunderkittens is now running on nvidia vera rubin!

our kernels team got early access to nvl72 and spent the past few days digging through the new isa and bringing nvfp4 + fp8 gemms to life

after reworking the kernels for rubin, we pushed them past 22 and 12 pflops respectively — competitive with cublas + cute dsl
View quoted post on X
0 replies 5 reposts 37 likes 4.3K views View on X
Dylan Lim Verified account @dylan__lim
Megakernels continue with our 8-GPU Llama-70B release! Please check out below!
Benjamin F Spector Verified account @spectorb Sep 28, 2025
(1/8) We’re releasing an 8-GPU Llama-70B inference engine megakernel! Our megakernel supports arbitrary batch sizes, mixed prefill+decode, a paged KV cache, instruction pipelining, dynamic scheduling, interleaved communication, and more! On ShareGPT it’s 22% faster than SGLang.
View quoted post on X
Dylan Lim Verified account @dylan__lim
Happy to announce that multi-GPU ThunderKittens is finally here! Help your GPU's meow better by checking out the following blog!
Stuart Sul Verified account @stuart_sul Sep 22, 2025
(1/6) We’re happy to share that ThunderKittens now supports writing multi-GPU kernels, with the same programming model and full compatibility with PyTorch + torchrun.

We’re also releasing collective ops and fused multi-GPU GEMM kernels, up to 2.6x faster than PyTorch + NCCL.

(Joint with @dylan__lim, @spectorb, and @HazyResearch. Generously supported by @cursor_ai)
View quoted post on X