HGPU group

High performance computing on graphics processing units (GPU): AMD/ATI, nVidia, Intel Xeon Phi, CUDA, OpenCL, OpenGL, GPGPU, HPC

[HGPU group @hgpu]

ATTUNE: Agentic Throughput Tuning for LLM Training

#CUDA #LLM #Package

https://t.co/pi2iOXRWPg

[HGPU group @hgpu]

GPU Goldmines: Tunable WebGPU Kernels for Edge LLM Inference

#CUDA #OpenCL #LLM

https://t.co/YWhv0q32Av

[HGPU group @hgpu]

A New Lens on Multi-GPU Communication: API-Level Memory Models for Portable Optimization

#AMD #HPC #MemoryModel

https://t.co/8YEJAnuhLG

[HGPU group @hgpu]

Understanding NVIDIA Grace-Blackwell Unified Memory Through Placement-Controlled Microbenchmarks

#CUDA #Performance #NVIDIA #Blackwell #Memory

https://t.co/2cmieY8uMj

[HGPU group @hgpu]

HPCAgent-Bench: A Framework for Sound Benchmarking of Code-Optimizing Agents, Harnesses, and Skills

#Triton #ROCm #HIP #OpenMP #AMD #HPC #Package

https://t.co/zSDX9m068M

[HGPU group @hgpu]

AgentGarten: Code Worlds for Evolving Agents

#CUDA #Triton #LLM #Package

https://t.co/NYR2KENi98

[HGPU group @hgpu]

A Unified Intermediate Representation and Execution Architecture for Heterogeneous and Distributed Neural Simulation

#CUDA #Package

https://t.co/8XOsVasG30

[HGPU group @hgpu]

ActTune: Action-Aware Precision and GPU Operating-Point Adaptation for Energy-Efficient Vision-Language-Action Inference

#LLM #Robotics

https://t.co/73QoSazOyD

[HGPU group @hgpu]

GPU Acceleration of Awkward Arrays: Using Python cuda.compute

#CUDA #Physics #Python

https://t.co/giBzAp7c3R

[HGPU group @hgpu]

KernelLens: Deploying Triton Kernels to C++ Inference Engines

#Triton #CUDA #PTX #Package

https://t.co/EIsvqO2Ijs

[HGPU group @hgpu]

[Book] Introduction to Computer Vision

#Book #ComputerVision #OpenCV #Python

https://t.co/Y7Z7IoWK4u

[HGPU group @hgpu]

KernelZero: Co-Evolving Proposer and Coder for Continuously Improved GPU Kernel Generation

#CUDA #Triton #LLM #Package

https://t.co/hTJ9ozIP0Z

[HGPU group @hgpu]

AIM: Agentic Idea Management for Automated Research

#CUDA #LLM

https://t.co/SPNyEYJvNt

[HGPU group @hgpu]

Automatically Generating ML Compiler Backends from Tensor Accelerator ISA Descriptions

#ISA #CodeGeneration #ML #Compilers #Package

https://t.co/fLhBbWhmeF

[HGPU group @hgpu]

AI as a Compiler: Compiling Triton kernels without the Triton compiler

#Triton #PTX #LLM

https://t.co/OqNNBIApTX

[HGPU group @hgpu]

KernelOPT: Dispatch-Aware Agentic Search for GPU Kernel Optimization

#Triton #CUDA #DeepLearning #LLM

https://t.co/JlWkC1lwhh

[HGPU group @hgpu]

Xtrace: High-Fidelity GPU Intra-Kernel Tracing via Binary-Level Instruction Splicing

#LLM #Performance

https://t.co/vCsAR275q1

[HGPU group @hgpu]

Microarchitectural Memory Bandwidth Saturation, KV-Cache Paging Dynamics, and Time-to-First-Token Latency: A Comparative Benchmark of vLLM, TensorRT-LLM, and FlashAttention-3 on NVIDIA Hopper H100 versus AMD Instinct MI300X

#CUDA #ROCm #AMD #vLLM

https://t.co/FSqewEKejG

[HGPU group @hgpu]

Accelerating the Mitigation of LLM Inference Nondeterminism Across GPU Architectures

#CUDA #LLM #Package

https://t.co/qpcPa8GgBz

[HGPU group @hgpu]

pytest-gpu-proof: Enabling Cloud-CPU Continuous Integration for GPU Code with Local GPU Attestation

#Python #Cloud #CI #Package

https://t.co/ew2OTOEZh9