Skip to content
View BBuf's full-sized avatar
🏠
Working from home
🏠
Working from home
2.7k followers · 60 following

Achievements

Achievement: Pull Sharkx4Achievement: Pair Extraordinairex4Achievement: YOLOAchievement: QuickdrawAchievement: Starstruckx3Achievement: Arctic Code Vault Contributor

Achievements

Achievement: Pull Sharkx4Achievement: Pair Extraordinairex4Achievement: YOLOAchievement: QuickdrawAchievement: Starstruckx3Achievement: Arctic Code Vault Contributor

Block or report BBuf

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
BBuf/README.md
terminal typing banner
root@bbuf-gpu-node:~# ./profile --scan[ok] identity : Xiaoyu Zhang / BBuf[ok] role : Core Developer @ SGLang | RadixArk[ok] focus : LLM inference, CUDA kernels, AI infrastructure[ok] preferred io : clean kernels, fast serving, measurable wins[ok] status : optimizing the path from model weights to tokens

./current_work

$ bbufctl top --sort impact
PID SYSTEM MODE SIGNAL
001 SGLang serving high-throughput LLM and VLM inference
002 CUDA / CUTLASS / Triton kernels memory bandwidth, latency, occupancy
003 Kernel Pilot agents AI-assisted kernel engineering loops
004 AI Infra Skills workflows reproducible infra/debug/profiling playbooks

./selected_systems

SGLanghow-to-optim-algorithm-in-cuda

kernel-pilotAI-Infra-Auto-Driven-SKILLS

./toolchain

./runtime_stats

GitHub statsTop languagesactivity graph

./contact

root@bbuf-gpu-node:~# cat /etc/contactGitHub : https://github.com/BBufBlog : https://www.giantpandacv.comWork : SGLang / RadixArk

Pinned Loading

  1. tvm_mlir_learntvm_mlir_learnPublic

    compiler learning resources collect.

    Python 2.8k 370

  2. how-to-optim-algorithm-in-cudahow-to-optim-algorithm-in-cudaPublic

    how to optimize some algorithm in cuda.

    Cuda 3.2k 290

  3. sgl-project/sglangsgl-project/sglangPublic

    SGLang is a high-performance serving framework for large language models and multimodal models.

    Python 32.1k 8k

  4. vipshop/cache-ditvipshop/cache-ditPublic

    A PyTorch-native inference engine with cache, parallelism, quantization and cpu offload for DiTs.

    Python 1.3k 77