GPU Goldmines offers a framework for tuning WebGPU kernels across devices, targeting matrix multiplication and flash attention for edge LLM inference. The approach uses performance characterization to speed up autotuning…

#WebGPU #LLMInference #EdgeAI #GPUComputing
https://arxiv.org/abs/2610.17321

youtu.be/S-OMnc5Djug NVIDIA H100 80GB GPU for enterprise AI and HPC, highlighting Hopper architecture, fourth-generation Tensor Cores, 80GB HBM2e memory, PCIe 5.0, NVLink, MIG @nvidiabot.bsky.social NVIDIA #NVIDIAH100 #ArtificialIntelligence #DeepLearning #GPUComputing #DataCenter #soniccomponents

Abhisek Goswami: CUDA and PostgreSQL: Twins from two Worlds
tech_blogs_planetscale

#GPUComputing

Kamino: GPU-Based Simulation of Rigid Multi-Body Physics with Kinematic Loops [video]
Discussion | hackernews | Author: ImJasonH

#GPUComputing

GPGPU-SIM – Cycle-Level Simulator for Nvidia GPUs with CUDA or OpenCL Workloads
Discussion | hackernews | Author: peter_d_sherman

#GPUComputing

Show HN: Rgpu – a PyTorch device whose tensors live on a remote GPU
Discussion | hackernews | Author: boxstream

#GPUComputing

Unroll a loop, make shader 10x faster
Discussion | lobsters | Author: calvin

#GPUComputing

Efficient FlashAttention on Blackwell via Fixed-Shift Softmax and Persistent Scheduling
tech_blogs_arxiv | Author: Oleksandr Stashuk, Hongtao Yu, Jay Shah

#GPUComputing

Show HN: Enki – Write GPU compute kernels in pure stable Rust
Discussion | hackernews | Author: enki_runtime

#GPUComputing

UMAP embeddings can vary between runs because stochastic negative sampling makes the repulsive forces depend on sampling order. ibUMAP addresses this with a coherent, field-based formulation evaluated…

#UMAP #DimensionalityReduction #MachineLearning #GPUComputing
https://arxiv.org/abs/2610.01445

KernelBench: Can LLMs Write GPU Kernels? – Benchmark and Toolkit, Torch –> CUDA
Discussion | hackernews | Author: peter_d_sherman

#GPUComputing

SoRoMoX is a Python/JAX framework for reduced-order soft-robot simulation, built on Cosserat-rod theory and accelerated with Warp GPU kernels for parallel continuum-model execution. It supports differentiable,…

#SoftRobotics #JAX #Robotics #GPUComputing
https://arxiv.org/abs/2608.06650

New arXiv paper proposes map-conditioned autoregressive generation of human mobility, using a road raster to condition a decoder emitting 31.25 m mesh-cell tokens, tested on trajectories from Ishikawa Prefecture.

#Semiconductors #AIInfrastructure #GPUComputing
https://arxiv.org/abs/2609.32360

🚀 Build practical skills in HPC performance engineering at @it4innovations.bsky.social in Czechia.

Profile, benchmark and optimise parallel applications across modern CPU and GPU systems using MPI, OpenMP, CUDA, Slurm and more.

🔗 Apply Now: hpctrain.eu/traineeship/...

#HPCTRAIN #HPC #GPUComputing

🚀 Interested in AI and HPC?

Join @hpctrain.bsky.social at Research Institute in Sweden and gain hands-on experience optimising and benchmarking large-scale AI applications across CPU and GPU-based HPC systems.

🔗 hpctrain.eu/traineeship/

#HPCTRAIN #AI #HPC #GPUComputing #EuroHPC

Show HN: Agentic CUDA Kernel Optimizer
Discussion | hackernews | Author: bertaye

#GPUComputing

fatbin tools for object files
tech_blogs_redplait_blogspot_com | Author: redp

#GPUComputing

CUDA.jl 6.4 is here 🚀

🟢 Improved NVIDIA Jetson support
🟢 Compiled GPU code caching for faster TTFX
🟢 CUDA 13.4 support
🟢 LLVM 23 upgrade

See what's new → juliagpu.org/post/2026-09...

#JuliaLang #CUDA #GPUComputing

Two GPU tools for data too big to look at

cuGraph does graph analytics on the GPU; Omniverse does real-time 3D simulation.

#NCAAIIO #AI #DataScience #GPUComputing

Full NCA-AIIO explanation, free: https://navyduck.com/nvidia/ai-infrastructure/nca-aiio/q117-a-research-team-needs-to

CUDA Toolkit 13.4技術解説：Windows on Arm対応とGPU共有リソース管理の進化

CUDA 13.4が登場。Windows on Arm対応と共有GPUの制御強化を詳説。

#CUDA #GPUComputing #WindowsonArm #HPC #SystemProgramming