Bodun Hu (@BodunHu) header
Bodun Hu (@BodunHu) profile picture

CS Ph.D. student @UTAustin. Student researcher @Meta. Working on systems for ML

Austin, TX
Joined January 2021
Photo by Bodun Hu (@BodunHu)Photo by Bodun Hu (@BodunHu)Photo by Bodun Hu (@BodunHu)Photo by Bodun Hu (@BodunHu)
GPUs still optimize compute and communication separately. That leaves performance on the table. CUCo is our agentic, training-free framework that rewrites CUDA+NCCL programs to jointly co-design compute + communication on device, unlocking overlap, fusion, and pipelining. Up to 1.57× end-to-end latency reduction on MoE, KV-cache transfers, FlashAttention+context parallelism, and GEMM+AllGather. 📄 arxiv.org/abs/2603.02376 🌐 ut-infraai.github.io/cuco/ 💻 github.com/UT-InfraAI/cuco #CUDA #MLSystems #GPU #HPCL @BodunHu @utnslab
2
15
2
157
11,527
For decades, OSes have relied on hand-tuned heuristics. What if they could write their own? 🔥 Meet PolicySmith, our LLM-driven framework that generates high-performing, interpretable system code. 📄arxiv.org/abs/2510.08803 at #HotNets25 #LDOS #MLforSystems #CodeGeneration
4
2
8
2,362
Check out our hot take on policies in computer systems, especially machine learned ones. As we dig deeper into LDOS implementation, we're seeing a foundational role for learned heuristic code driving resource management, allocation, and planning policies. In the limit, code learned on the fly can tailor an entire OS to optimally serve currently running workloads while meeting complex system objectives. This work will appear at HotNets 2025. Led by the brilliant Rohit Dwivedula and @divytweet with collaborators @swarat and @daehyeok_kim .
For decades, OSes have relied on hand-tuned heuristics. What if they could write their own? 🔥 Meet PolicySmith, our LLM-driven framework that generates high-performing, interpretable system code. 📄arxiv.org/abs/2510.08803 at #HotNets25 #LDOS #MLforSystems #CodeGeneration
1
11
1,478
Thrilled to hear USENIX ATC will now officially continue under ACM SIGOPS sponsorship! Same name, agenda, and abbreviation. This is a testament to the community's strength. 👏🏽
1
2
7
8,143
KV caching in LLMs, clearly explained (with visuals):
14
157
6
1,880
436,453
Introducing ChartMuseum🖼️, testing visual reasoning with diverse real-world charts! ✍🏻Entirely human-written questions by 13 CS researchers 👀Emphasis on visual reasoning – hard to be verbalized via text CoTs 📉Humans reach 93% but 63% from Gemini-2.5-Pro & 38% from Qwen2.5-72B
2
36
5
77
18,658
Bodun Hu retweeted
News: the latest ChatGPT-4o (2025-03-26) jumps to #2 on Arena, surpassing GPT-4.5! Highlights - Significant improvement over the January version (+30 pts, #5->#2) - Tied #1 in Coding, Hard Prompts. Top-2 across ALL categories - Matching or surpassing GPT-4.5 with 10x cheaper price. Congrats @OpenAI for the new milestone! View below for more insights ⬇️
GPT-4o got an another update in ChatGPT! What's different? - Better at following detailed instructions, especially prompts containing multiple requests - Improved capability to tackle complex technical and coding problems - Improved intuition and creativity - Fewer emojis 🙃
47
111
76
1,185
308,258
Bodun Hu retweeted
BREAKING: Gemini 2.5 Pro is now #1 on the Arena leaderboard - the largest score jump ever (+40 pts vs Grok-3/GPT-4.5)! 🏆 Tested under codename "nebula"🌌, Gemini 2.5 Pro ranked #1🥇 across ALL categories and UNIQUELY #1 in Math, Creative Writing, Instruction Following, Longer Query, and Multi-Turn! Massive congrats to @GoogleDeepMind for this incredible Arena milestone! 🙌 More highlights in thread👇
Think you know Gemini? 🤔 Think again. Meet Gemini 2.5: our most intelligent model 💡 The first release is Pro Experimental, which is state-of-the-art across many benchmarks - meaning it can handle complex problems and give more accurate responses. Try it now → goo.gle/4c2HKjf
71
386
108
2,291
470,734
Applications opened today for the LDOS PhD Research School: ldos.utexas.edu/events/resea…. Please apply or share with your friends and/or students!
1
8
816
DeepSeek R1 is 100% Opensource and 96.4% cheaper than OpenAI o1 while delivering similar performance. OpenAI o1: $60.00 per 1M output tokens DeepSeek R1: $2.19 per 1M output tokens People with $200 ChatGPT subscription, let that sink in.
124
507
148
3,312
1,260,697
Nobel Prize in Physics is awarded to AI this year. Incredible!
BREAKING NEWS The Royal Swedish Academy of Sciences has decided to award the 2024 #NobelPrize in Physics to John J. Hopfield and Geoffrey E. Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks.”
1
2
384
Replying to @utnslab
@utnslab at our recent end-of-semester social.
4
29
1,829
Super happy about this!! ❤️ But to be honest, with exceptional faculty colleagues and students, @UTCompSci feels like -- and deserves to be -- a top 5 department 🙂
Kudos to @UTCompSci’s graduate program on climbing up a spot in the @USNews rankings! Details here: cns.utexas.edu/news/announce… #USNewsRankings #UTCompSci #UTAustin
1
3
26
5,798
Replying to @utnslab
@utnslab hike at Barton Creek with @VenkatArun95 @daehyeok_kim and our amazing students - @jitaotweet @Jane_JiayiChen @BodunHu @divytweet @brian0296 @RohitDwivedula Jiaxin, Geon-Woo, Jeongyoon, Yeonju, Haoyu, Darius, and Donghyun.
3
9
1
46
3,817
Generative LLMs are slow and expensive to serve. Their much smaller, distilled versions are faster and cheaper but achieve suboptimal generative performance. We show it is possible to achieve the best of both worlds. Code: github.com/flexflow/FlexFlow… Paper: cs.cmu.edu/~zhihaoj2/papers/…
2
70
7
355
82,292
Bodun Hu retweeted
A series of AI announcements by Meta: - MTIA v1: an AI chip for fast inference: ai.facebook.com/blog/meta-tr… - RSC: 5 exaflops, 16,000 GPU Research Super Cluster for AI research: ai.facebook.com/blog/superco… - AI-focused data centers: ai.facebook.com/blog/meta-ai…
72
270
53
1,448
390,707
Google is in trouble. I got early 'Alpha' access to GPT-4 with browsing and ran some tests. Here are 8 crazy things I found:
789
3,667
548
19,205
8,232,250
Bodun Hu retweeted
Alpa is a framework that uses just one line of code to easily automate the complex model parallelism process for large #DeepLearning models. Learn more and check out the code. goo.gle/3y7xZ1f
5
97
10
352