high performance computing on graphics processing units: hgpu.org

hgpu.org » CUDA

Profiling Concurrent Vision Inference Workloads on NVIDIA Jetson – Extended

Abhinaba Chakraborty, Wouter Tavernier, Akis Kourtis, Mario Pickavet, Andreas Oikonomakis, Didier Colle

View

Download (PDF)

Tags: Computer science, CUDA, nVidia, nVidia Jetson Nano, nVidia Jetson Orin Nano, Performance

August 24, 2025 by hgpu

Towards Efficient and Practical GPU Multitasking in the Era of LLM

Jiarong Xing, Yifan Qiao, Simon Mo, Xingqi Cui, Gur-Eyal Sela, Yang Zhou, Joseph Gonzalez, Ion Stoica

View

Download (PDF)

Source codes

Tags: Computer science, CUDA, LLM, nVidia, nVidia A100, Package, Performance, PyTorch

August 24, 2025 by hgpu

Bandicoot: A Templated C++ Library for GPU Linear Algebra

Ryan R. Curtin, Marcus Edel, Conrad Sanderson

View

Download (PDF)

Source codes

Tags: Computer science, CUDA, Linear Algebra, nVidia, nVidia GeForce RTX 2080 Ti, OpenCL, Package, Programming techniques

August 24, 2025 by hgpu

Fuzz4cuda: Fuzzing Your Nvidia Gpu Libraries Through Debug Interface

Yuhao Zhou, Peng Jia, Jiayong Liu, Ximing Fan

View

Download (PDF)

Tags: Computer science, CUDA, nVidia, nVidia RTX A6000, PTX, Security

August 24, 2025 by hgpu

The Fused Kernel Library: A C++ API to Develop Highly-Efficient GPU Libraries

Oscar Amoros, Albert Andaluz, Johnny Nunez, Antonio J. Pena

View

Download (PDF)

Source codes

Tags: Computer science, CUDA, nVidia, nVidia GeForce RTX 4090, Package, Performance

August 17, 2025 by hgpu

GPUHammer: Rowhammer Attacks on GPU Memories are Practical

Chris S. Lin, Joyce Qu, Gururaj Saileshwar

View

Download (PDF)

Source codes

Tags: Computer science, CUDA, Machine learning, nVidia, nVidia RTX A6000, Package, Security

August 17, 2025 by hgpu

Block: Balancing Load in LLM Serving with Context, Knowledge and Predictive Scheduling

Wei Da, Evangelia Kalyvianaki

View

Download (PDF)

Source codes

Tags: Artificial intelligence, Computer science, CUDA, LLM, nVidia, nVidia A30, Package, Task scheduling

August 17, 2025 by hgpu

SIGMo: High-Throughput Batched Subgraph Isomorphism on GPUs for Molecular Matching

Antonio De Caro, Gennaro Cordasco, Federico Ficarelli, Biagio Cosenza

View

Download (PDF)

Source codes

Tags: Biology, Chemistry, Computer science, CUDA, MPI, nVidia, nVidia V100, oneAPI, Package

August 10, 2025 by hgpu

Performance Portable Gradient Computations Using Source Transformation

Kim Liegeois, Brian Kelley, Eric Phipps, Sivasankaran Rajamanickam, Vassil Vassilev

View

Download (PDF)

Tags: AMD Radeon Instinct MI300X, ATI, Computer science, CUDA, HIP, Kokkos, Machine learning, Mathematical Software, nVidia, nVidia H100

August 3, 2025 by hgpu

Kevin: Multi-Turn RL for Generating CUDA Kernels

Carlo Baronio, Pietro Marsella, Ben Pan, Simon Guo, Silas Alberti

View

Download (PDF)

Tags: Artificial intelligence, Computer science, CUDA, LLM, Machine learning, nVidia, nVidia H100, nVidia H200, Performance

July 20, 2025 by hgpu

Specx: a C++ task-based runtime system for heterogeneous distributed architectures

Paul Cardosi, Bérenger Bramas

View

Download (PDF)

Source codes

Tags: Computer science, CUDA, Heterogeneous systems, HIP, nVidia, nVidia A100, nVidia P100, nVidia Quadro RTX 8000, nVidia V100, Package, Task scheduling

July 20, 2025 by hgpu

Pre-Training LLMs on a budget: A comparison of three optimizers

Joel Schlotthauer, Christian Kroos, Chris Hinze, Viktor Hangya, Luzian Hahn, Fabian Küch

View

Download (PDF)

Tags: Artificial intelligence, Computer science, CUDA, LLM, Machine learning, nVidia, nVidia A100

July 20, 2025 by hgpu

kvcached: Elastic KV cache for dynamic GPU sharing and efficient multi-LLM inference

Towards Efficient and Practical GPU Multitasking in the Era of LLM

Bandicoot: C++ library for GPU accelerated linear algebra

Bandicoot: A Templated C++ Library for GPU Linear Algebra

Luthier: Bridging Auto-Tuning and Vendor Libraries for Efficient Deep Learning Inference

Fused Kernel Library (FKL)

The Fused Kernel Library: A C++ API to Develop Highly-Efficient GPU Libraries

Block: Balance Loader of LLM Serving with Context, Knowledge and Predictive Scheduling

Block: Balancing Load in LLM Serving with Context, Knowledge and Predictive Scheduling

GPUHammer: Rowhammer Attacks on GPU Memories are Practical

SIGMo: Scalable Isomorphism Graph Matching on GPUs

SIGMo: High-Throughput Batched Subgraph Isomorphism on GPUs for Molecular Matching

DGEMM without FP64 Arithmetic - using FP64 Emulation and FP8 Tensor Cores with Ozaki Scheme

DGEMM without FP64 Arithmetic – using FP64 Emulation and FP8 Tensor Cores with Ozaki Scheme

GEAK-agent: LLM-based AI agent, which can write correct and efficient GPU kernels automatically

Geak: Introducing Triton Kernel AI Agent & Evaluation Benchmarks

OpenDwarfs 2025: re-engineered version of the OpenDwarfs benchmark suite, for compatibility with modern platforms

OpenDwarfs 2025: Modernizing the OpenDwarfs Benchmark Suite for Heterogeneous Computing

See all packages

* * *

high performance computing on graphics processing units: hgpu.org

Profiling Concurrent Vision Inference Workloads on NVIDIA Jetson – Extended

Towards Efficient and Practical GPU Multitasking in the Era of LLM

Bandicoot: A Templated C++ Library for GPU Linear Algebra

Fuzz4cuda: Fuzzing Your Nvidia Gpu Libraries Through Debug Interface

The Fused Kernel Library: A C++ API to Develop Highly-Efficient GPU Libraries

GPUHammer: Rowhammer Attacks on GPU Memories are Practical

Block: Balancing Load in LLM Serving with Context, Knowledge and Predictive Scheduling

SIGMo: High-Throughput Batched Subgraph Isomorphism on GPUs for Molecular Matching

Performance Portable Gradient Computations Using Source Transformation

Kevin: Multi-Turn RL for Generating CUDA Kernels

Specx: a C++ task-based runtime system for heterogeneous distributed architectures

Pre-Training LLMs on a budget: A comparison of three optimizers

Recent source codes

kvcached: Elastic KV cache for dynamic GPU sharing and efficient multi-LLM inference

Bandicoot: C++ library for GPU accelerated linear algebra

Luthier: Bridging Auto-Tuning and Vendor Libraries for Efficient Deep Learning Inference

Fused Kernel Library (FKL)

Block: Balance Loader of LLM Serving with Context, Knowledge and Predictive Scheduling

GPUHammer: Rowhammer Attacks on GPU Memories are Practical

SIGMo: Scalable Isomorphism Graph Matching on GPUs

DGEMM without FP64 Arithmetic - using FP64 Emulation and FP8 Tensor Cores with Ozaki Scheme

GEAK-agent: LLM-based AI agent, which can write correct and efficient GPU kernels automatically

OpenDwarfs 2025: re-engineered version of the OpenDwarfs benchmark suite, for compatibility with modern platforms

Most viewed papers (last 30 days)