Really grateful to @GPU_MODE for the opportunity to talk about my recent Tiny TPU project: 🧵 https://t.co/9hqftJY82z.

btw today at 3pm PST (in ~4 hours) we're having Vicki Wang from NVIDIA giving a @GPU_MODE talk on CuTe DSL, its features, and how to use the most of it if you're currently competing in the NVFP4 Blackwell competition this will be very helpful, but it's open to anyone!

Kernel Challenge #2 is LIVE In partnership with @GPU_MODE, we've dropped the next high-performance task: NVFP4 GEMM. Prove your kernel optimization skills and push the limits of low-precision computing. 📝 Problem: NVFP4 GEMM 🗓️ Deadline: Dec 19 🔗 Join Now:

Training LLMs with NVFP4 is hard because FP4 has so few values that I can fit them all in this post: ±{0, 0.5, 1, 1.5, 2, 3, 4, 6}. But what if I told you that reducing this range even further could actually unlock better training + quantization performance? Introducing Four

Had a good time competing in @GPU_MODE first hackathon for NVFP4 GEMV. This was especially fun since I'm taking @kayvonf 's CS149 this quarter, and it was nice applying some of the concepts we learned in class. I made a fork of the popcorn-cli tool to turn off the terminal user

Sign up! https://t.co/J6umUfPeia we're at over 1,300 registrations and still going strong!

▓▓▓░░░░░░░░░ 25% We just concluded the GEMV problem for the Blackwell NVFP4 competition. And we've started on a new GEMM problem. You can still sign up and be eligible for prizes per problem and the grand prize. glhf!

After 3 weeks, we have concluded our first problem of the @GPU_MODE x @nvidia competition, NVFP4 GEMV. Thanks to everyone who has participated, we have collected over 40k submissions from >200 users. Congrats to the winners and good luck with the next problem, NVFP4 GEMM 🔥

CuteDSL 4.3.1 is here 🚀 Major host overhead optimization (10-40µs down to a 2µs in hot loops_, streamlined PyTorch interop (pass torch.Tensors directly, no more conversions needed) and export and use in more languages and envs. All powered by apache tvm-ffi ABI

Today we are releasing our first public beta of Nsight Python! The goal is to simplify the life of a Python developer by proving a pythonic way to analyze your kernel code! Check it out, provide feedback! Nsight Python — nsight-python

My Triton version for the NVFP4 gemv kernel competition @GPU_MODE 🧵 https://t.co/4u3hAFIlpS

Saturday, November 15, at 12:00 PM PST we have a special talk lined up. Paulius Micikevicius arguably the man most responsible behind the efficiency revolution in GPUs with low bit dtypes and sparsity will be speaking. Cohosting with @cHHillee

AI has been built on one vendor’s stack for too long. AMD’s GPUs now offer state-of-the-art peak compute and memory bandwidth — but the lack of mature software / the “CUDA moat” keeps that power locked away. Time to break it and ride into our multi-silicon future. 🌊 It's been a

1,000 registrations so far!

updates to https://t.co/asKuaIFf5P. working on the runtime and eager kernels now. picograd is taking longer than other "hobby" autograds i've seen. but our plan is to be the *definitive* resource on building your own pytorch. we agree with @karpathy that course building is a

The most intuitive explanation of floats I've ever come across, courtesy of @fabynou https://t.co/XNiZNZTNlf

Congrats again to the winners of the GPU MODE IRL #2 hackathon held in the beautiful @Accel office A giant thank you to the only neocloud that could pull off a Blackwell hackathon for us @nebiusai 1st Place – Symmetric Minds: Enabled multi-GPU expert-parallel MoE inference in

Surprising properties of low-precision floating point numbers are in the news again! These numerical formats are ubiquitous in large NNs but new to most programmers. So I worked with @klyap_ last week to put together this little visualizer: https://t.co/ntHlazyDmK.

🔥 New Blog: “Disaggregated Inference: 18 Months Later” 18 months in LLM inference feels like a new Moore’s Law cycle – but this time not just 2x per year: 💸 Serving cost ↓10–100x 🚀 Throughput ↑10x ⚡ Latency ↓5x A big reason? Disaggregated Inference. From DistServe, our

If you are a student and you want a career follow @GPU_MODE, attend events, and do the completions. You WILL be employable.