All Projects → mountassir → Gmonitor

mountassir / Gmonitor

Licence: gpl-3.0
gmonitor is a GPU monitor (Nvidia only at the moment)

Projects that are alternatives of or similar to Gmonitor

Komputation
Komputation is a neural network framework for the Java Virtual Machine written in Kotlin and CUDA C.
Stars: ✭ 295 (+74.56%)
Mutual labels:  nvidia, gpu, cuda
Cudasift
A CUDA implementation of SIFT for NVidia GPUs (1.2 ms on a GTX 1060)
Stars: ✭ 555 (+228.4%)
Mutual labels:  nvidia, gpu, cuda
Thrust
The C++ parallel algorithms library.
Stars: ✭ 3,595 (+2027.22%)
Mutual labels:  nvidia, gpu, cuda
Libcudacxx
The C++ Standard Library for your entire system.
Stars: ✭ 1,861 (+1001.18%)
Mutual labels:  nvidia, gpu, cuda
Optix Path Tracer
OptiX Path Tracer
Stars: ✭ 60 (-64.5%)
Mutual labels:  nvidia, gpu, cuda
Gprmax
gprMax is open source software that simulates electromagnetic wave propagation using the Finite-Difference Time-Domain (FDTD) method for numerical modelling of Ground Penetrating Radar (GPR)
Stars: ✭ 268 (+58.58%)
Mutual labels:  nvidia, gpu, cuda
Ilgpu
ILGPU JIT Compiler for high-performance .Net GPU programs
Stars: ✭ 374 (+121.3%)
Mutual labels:  nvidia, gpu, cuda
Nvidia Modded Inf
Modified nVidia .inf files to run drivers on all video cards, research & telemetry free drivers
Stars: ✭ 227 (+34.32%)
Mutual labels:  nvidia, gpu, cuda
Cuda
Experiments with CUDA and Rust
Stars: ✭ 31 (-81.66%)
Mutual labels:  nvidia, gpu, cuda
Cub
Cooperative primitives for CUDA C++.
Stars: ✭ 883 (+422.49%)
Mutual labels:  nvidia, gpu, cuda
opencv-cuda-docker
Dockerfiles for OpenCV compiled with CUDA, opencv_contrib modules and Python 3 bindings
Stars: ✭ 55 (-67.46%)
Mutual labels:  gpu, cuda, nvidia
Deep Learning Boot Camp
A community run, 5-day PyTorch Deep Learning Bootcamp
Stars: ✭ 1,270 (+651.48%)
Mutual labels:  nvidia, gpu, cuda
peakperf
Achieve peak performance on x86 CPUs and NVIDIA GPUs
Stars: ✭ 33 (-80.47%)
Mutual labels:  gpu, cuda, nvidia
Deep Diamond
A fast Clojure Tensor & Deep Learning library
Stars: ✭ 288 (+70.41%)
Mutual labels:  nvidia, gpu, cuda
Plotoptix
Data visualisation in Python based on OptiX 7.2 ray tracing framework.
Stars: ✭ 252 (+49.11%)
Mutual labels:  nvidia, gpu, cuda
Cuda Api Wrappers
Thin C++-flavored wrappers for the CUDA Runtime API
Stars: ✭ 362 (+114.2%)
Mutual labels:  nvidia, gpu, cuda
Macos Egpu Cuda Guide
Set up CUDA for machine learning (and gaming) on macOS using a NVIDIA eGPU
Stars: ✭ 187 (+10.65%)
Mutual labels:  nvidia, gpu, cuda
Genomeworks
SDK for GPU accelerated genome assembly and analysis
Stars: ✭ 215 (+27.22%)
Mutual labels:  nvidia, gpu, cuda
Pyopencl
OpenCL integration for Python, plus shiny features
Stars: ✭ 790 (+367.46%)
Mutual labels:  nvidia, gpu, cuda
Parenchyma
An extensible HPC framework for CUDA, OpenCL and native CPU.
Stars: ✭ 71 (-57.99%)
Mutual labels:  nvidia, gpu, cuda

What is gmonitor?

This is a GPU monitoring program, it monitors the core usage, VRAM usage, PCI-E & memory bus usage and the temperature of the GPU. I write CUDA programs and always needed a way to monitor how they behave in real time, searched for an MSI afterburner like programs on GNU/Linux systems but couldn't find any, so I decided to put this together and have been using it for quite some time now. Though this is quite minimal, it's good enough for what I needed it.

Running environment:

This has been developed and used in a machine running Ubuntu, I have not tested it in any other systems.

Requirements:

At least one NVIDIA GPU, this only supports NVIDIA GPUs for now as I only needed this to monitor my CUDA programs. Support for AMD/Intel cards might come in the future if needed, feel free to add them as well.

This program probes the NVIDIA drivers for statistics through nvidia-settings, make sure the NVIDA drivers are properly installed on your system.

CUDA install:

Follow Nvidia CUDA install guide to install CUDA on your machine, If you face issues with the nouveau driver, try this Nouveau workaround.

Building from Source:

Configuring with CMake

Compile and install gmonitor in /usr/local/bin

$ cd gmonitor
$ mkdir build
$ cd build
$ cmake ..
$ make
$ sudo make install

Usage:

Download the executable from the latest release or compile your own from the source code in the folder "src", run the executable with the following arguments:

-d [displayMode]
0 (default) Monitor both current and previous GPU states.
1 Monitor most recent GPU state only.
2 Monitor previous GPU states only.
3 Same as 0, print current states for all GPU then print history.

-g [gpuNumber]
GPU numbers to monitor, primary GPU's number is 0...

-r [refreshRate]
Monitoring refresh rate (default is 2 seconds).

-o
Enable support for Optirun.

-s
Needed if you are running over ssh.

-h
Display this menu.

Examples:

Use defaults

gmonitor

Monitor the most recent state only for the first and third gpu.

gmonitor -d 1 -g 02

Monitor current and history states for 4 GPUs.

gmonitor -d 0 -g 0123

Monitor both current and previous states for all GPUs, refresh every 3 seconds.

gmonitor -d 0 -r 3

Limitations:

This only supports Nvidia cards at the moment and support for other cards might be added in the future. Multi GPU support have not been tested, it should scale and work with any number of available GPUs but I have not tested it myself as I only have one card.

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].