VeriGPU
GPU core
An open-source implementation of a GPU core for machine learning, designed to run PyTorch and other frameworks on an ASIC.
OpenSource GPU, in Verilog, loosely based on RISC-V ISA
852 stars
29 watching
94 forks
Language: SystemVerilog
last commit: 12 months ago
Linked from 1 awesome list
asic-designgpugpu-accelerationhardware-designsmachine-learningrisc-vrisc-v-assemblyverificationverilog
Related projects:
| Repository | Description | Stars |
|---|---|---|
| | A pre-configured environment for running PyTorch on OpenCL-enabled GPUs | 147 |
| | An experimental GPGPU processor focused on compute-intensive tasks with a synthesizable hardware design and software libraries. | 2,021 |
| | A Vulkan-based framework for accelerating computations on graphics processing units. | 347 |
| | A 4-stage RISC-V core for compute-intensive applications with a general-purpose extension interface | 225 |
| | A C++17 GPU-accelerated numerical computing library with Python-like syntax | 1,229 |
| | A Vulkan-based GPGPU engine that compiles and executes SPIR-V compute shaders for general-purpose computing tasks. | 133 |
| | A collection of teaching materials and tutorials focused on GPU programming in C++ | 62 |
| | A library for portable, modular, and performant GPU compute programming in Rust. | 1,597 |
| | A RISC-V processor core design implemented in SystemVerilog RTL | 830 |
| | A secure, 4-stage RISC-V core designed for high-security applications with both machine mode and user mode capabilities | 133 |
| | A 32-bit RISC-V CPU core with a two-stage pipeline designed for low-cost embedded control applications. | 29 |
| | An offline compiler and code analysis tool for various GPU architectures and programming languages. | 422 |
| | A tool for detecting and displaying information about graphics processing units (GPUs), with support for NVIDIA and Intel GPUs. | 138 |
| | Functional verification project for RISC-V cores | 458 |
| | A high-performance deep learning architecture designed to balance accuracy and efficiency on GPUs. | 471 |