G

gpu

Multidimensional arrays for C++. (Not an official Boost library) GitHub mirror: https://github.com/correaa/boost-multi

Template Numerical Library https://tnl-project.org/

k8s-config — Kubernetes Cluster Configuration

GitOps repository for the homelab k3s cluster — everything the cluster runs is defined here.

Manages all Kubernetes workloads and cluster-level configuration for a bare-metal k3s cluster with NVIDIA GPU passthrough. Deployed via Flux CD with Kustomize overlays for production and staging environments.

What's in here Flux GitOps: Kustomization CRDs in clusters/production/ and clusters/staging/ Base layer: Namespaces, ingress, NVIDIA device plugin, Weave GitOps dashboard Apps layer: Ollama (LLM serving), Open WebUI, ComfyUI (image generation), GitLab runner CI/CD: Schema validation (kubeconform), kustomize build checks, Flux CRD validation, AI code review Tech Stack Orchestration: k3s (bare-metal) GitOps: Flux CD v2 Config management: Kustomize GPU: NVIDIA passthrough with runtimeClassName: nvidia

Isard VDI (Virtual Desktop Infraestructure) Project. Open Source KVM Virtual Desktops based on KVM Linux and dockers. https://isardvdi.com/

Bandicoot: user-friendly C++ library for GPU accelerated linear algebra, integrating with CUDA and OpenCL

Event Generator, written in CUDA C++. Simulates LEP and LHC Collisions.

An EDSL that compiles to SPIR-V for graphics programming on the GPU in Haskell.

Special relativistic rendering for games

Track AI spending across API providers, MCP tools, subscriptions, and self-hosted GPUs from your terminal. One CLI to see what you're actually paying — Anthropic, OpenAI, OpenRouter, vLLM on your own hardware — with real TCO math. Compare your self-hosted $/MTok vs cloud pricing. Local-first, no SaaS required.

Kubernetes device plugin implementing the sharing of Nvidia GPUs between workloads

Zig Vulkan compute library for GPU-accelerated computing. Provides a high-level API for compute shaders, memory management, and kernel execution on Vulkan-capable GPUs.

A comprehensive CUDA compute library for Zig providing GPU memory management, kernel compilation/launching, device probing, and profiling capabilities

Bitcoin Mini Key Hunter is a tool that focuses on searching for 22-digit private keys starting with S and performing cracking drills. It uses the powerful GPU parallel computing capabilities of CUDA technology, similar to research tools such as brainflayer,keyhunt,BitCrack, ecloop,keyhunt cuda,etc.

KeyKiller is designed to achieve extreme performance on modern NVIDIA GPUs, solving the Satoshi puzzle. It leverages CUDA, warp-level parallelism, and batched EC operations to push the boundaries of cryptographic key discovery. It is commonly used in research projects such as Secp256k1-CUDA,Brainflayer,BitCrack, Keyhunt CUDA BitCrack,RCKangaroo.

After optimization, version 3080 is equivalent to version 4090. I fixed address matching errors and private key conversion errors, modified the random generator, added puzzles suitable for specific regions, and more. It is commonly used in puzzle games, Brainflayer, Keyhunt, BitCrack, VanitySearch, etc.

Volume-Language: Experimental volume description languages aimed at GPU distance field rendering.

DeepSpeed / PyTorch Distributed + TensorRT export

Genetics Algorithms and Genetic Programming library. https://genetics4j.org

This is a catalogue presenting the information displayed on Google Chrome's internal "chrome://gpu" page, collected from various hardware and operating system platforms.

Comparing the CPU and GPU implementations of LHAPDF