high performance computing on graphics processing units: hgpu.org

hgpu.org » Precision

True 4-Bit Quantized Convolutional Neural Network Training on CPU: Achieving Full-Precision Parity

Shivnath Tathe

View

Download (PDF)

Source codes

Tags: CNN, Compression, Deep learning, Package, Precision

March 18, 2026 by hgpu

Diagnosing FP4 inference: a layer-wise and block-wise sensitivity analysis of NVFP4 and MXFP4

Musa Cim, Burak Topcu, Mahmut Taylan Kandemir

View

Download (PDF)

Tags: AMD, Computer science, LLM, nVidia, nVidia GeForce RTX 5090, nVidia RTX PRO 6000, Precision

March 15, 2026 by hgpu

Practical FP4 Training for Large-Scale MoE Models on Hopper GPUs

Wuyue Zhang, Chongdong Huang, Chunbo You, Cheng Gu, Fengjuan Wang, Mou Sun

View

Download (PDF)

Source codes

Tags: Computer science, CUDA, LLM, nVidia, nVidia GB200, Package, Precision

March 8, 2026 by hgpu

Mixing Low-Precision Formats in Multiply-Accumulate Units for DNN Training

Mariko Tatsumi, Silviu-Ioan Filip, Caroline White, Olivier Sentieys, Guy Lemieux

View

Download (PDF)

Tags: Computer science, CUDA, Deep learning, FPGA, nVidia, Precision

December 11, 2022 by hgpu

APNN-TC: Accelerating Arbitrary Precision Neural Networks on Ampere GPU Tensor Cores

Boyuan Feng, Yuke Wang, Tong Geng, Ang Li, Yufei Ding

View

Download (PDF)

Source codes

Tags: Algorithms, Computer science, CUDA, Deep learning, Neural networks, nVidia, nVidia GeForce RTX 3090, Package, Precision, Python, Tesla A100

June 27, 2021 by hgpu

Tools for Reduced Precision Computation: A Survey

Stefano Cherubin, Giovanni Agosta

View

Download (PDF)

Tags: Approximate computing and storage, Computer science, CUDA, HPC, Mixed precision, nVidia, OpenCL, Precision, survey

May 3, 2020 by hgpu

FPGA Implementation of Reduced Precision Convolutional Neural Networks

Muhammad Mohid Nabil

View

Download (PDF)

Tags: Computer science, CUDA, Deep learning, FPGA, HLS, Image recognition, Machine learning, Neural networks, nVidia, nVidia GeForce GTX 1080, Precision, Thesis

May 12, 2019 by hgpu

Quasi-maximum Accuracy Floating-point Computations with GPGPU for Applications in Digital Signal Processing

Adam Dabrowski, Pawel Pawlowski, Mateusz Stankiewicz, Filip Misiorek

View

Download (PDF)

Tags: CUDA, nVidia, nVidia Quadro FX 580, Precision, Signal processing

December 24, 2011 by hgpu

Precision and Performance: Floating Point and IEEE 754 Compliance for NVIDIA GPUs

Nathan Whitehead, Alex Fit-Florea

View

Download (PDF)

Tags: Computer science, CUDA, nVidia, Precision

October 26, 2011 by hgpu

Agentic Code Optimization via Compiler-LLM Cooperation

MegaTrain: Full Precision Training of 100B+ Parameter Large Language Models on a Single GPU

Device Virtual Machine (DVM)

DVM: Real-Time Kernel Generation for Dynamic AI Models

AutoKernel: Autoresearch for GPU kernels. Give it any PyTorch model, go to sleep, wake up to optimized Triton kernels

AutoKernel: Autonomous GPU Kernel Optimization via Iterative Agent-Driven Search

SOL-ExecBench: Speed-of-Light Benchmarking for Real-World GPU Kernels Against Hardware Limits

Triton-Sanitizer: A Fast and Device-Agnostic Memory Sanitizer for Triton with Rich Diagnostic Context

* * *

high performance computing on graphics processing units: hgpu.org

True 4-Bit Quantized Convolutional Neural Network Training on CPU: Achieving Full-Precision Parity

Diagnosing FP4 inference: a layer-wise and block-wise sensitivity analysis of NVFP4 and MXFP4

Practical FP4 Training for Large-Scale MoE Models on Hopper GPUs

Mixing Low-Precision Formats in Multiply-Accumulate Units for DNN Training

APNN-TC: Accelerating Arbitrary Precision Neural Networks on Ampere GPU Tensor Cores

Tools for Reduced Precision Computation: A Survey

FPGA Implementation of Reduced Precision Convolutional Neural Networks

Quasi-maximum Accuracy Floating-point Computations with GPGPU for Applications in Digital Signal Processing

Precision and Performance: Floating Point and IEEE 754 Compliance for NVIDIA GPUs

Recent source codes

Agentic Code Optimization via Compiler-LLM Cooperation

MegaTrain: Full Precision Training of 100B+ Parameter Large Language Models on a Single GPU

Device Virtual Machine (DVM)

AutoKernel: Autoresearch for GPU kernels. Give it any PyTorch model, go to sleep, wake up to optimized Triton kernels

SOL-ExecBench: Speed-of-Light Benchmarking for Real-World GPU Kernels Against Hardware Limits

Triton-Sanitizer: A Fast and Device-Agnostic Memory Sanitizer for Triton with Rich Diagnostic Context

LLM.Q: Quantized LLM training in pure CUDA/C++

True 4-Bit Quantized CNN Training on CPU

cuFuzz: A GPU-oriented coverage-guided fuzzer for userland CUDA application

KernelSkill: A Multi-Agent Framework for GPU Kernel Optimization

Most viewed papers (last 30 days)