AMD GPU stats
Project description
pyamdgpuinfo
AMD GPU stats
Install
pip3 install pyamdgpuinfo
Only linux is supported, using the amdgpu driver.
Precompiled wheels for python 3.5, 3.6, 3.7 and 3.8 are the default method of install. This means that you don't need cython or any other dependencies to install it normally.
The library is written using cython, meaning that cython and and a C compiler are needed to build and install from source. Additionally, libdrm-dev is needed.
Usage
Example:
import pyamdgpuinfo
gpus = pyamdgpuinfo.setup_gpus()
# query first device
first_gpu = list(gpus.keys())[0]
vram_usage = pyamdgpuinfo.query_vram_usage(first_gpu)
print(vram_usage)
All documentation is in the docstrings of each function/class.
Available functions are (see docstrings for more info):
- setup_gpus - Sets up devices so they can be used.
- start_utilisation_polling - Starts polling GPU registers for utilisation statistics.
- stop_utilisation_polling - Stops the utilisation polling thread.
- cleanup - Cleans up allocated memory (only recommended if de-initialising the module before the main program is ended).
Query functions (again see docstrings):
- query_max_clocks - Queries max GPU clocks
- query_sclk - Queries shader (core) clock
- query_mclk - Queries memory clock
- query_vram_usage - Queries VRAM usage
- query_gtt_usage - Queries GTT usage
- query_temp - Queries temperature
- query_load - Queries GPU load
- query_power - Queries power consumption
- query_utilisation - Queries utilisation of different GPU parts (requires utilisation polling to be running)
VRAM and GTT sizes are returned by setup_devices (if they are available).
Mentions
Parts of this package were inspired by radeontop.
License
GPLV3
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distributions
Hashes for pyamdgpuinfo-1.0.5-cp38-cp38-manylinux2014_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | d3f455f227ae653461a6066f89b5bed7b6758213e68ec81cee6949e2e7a87a23 |
|
MD5 | 05573d937fd28d3a93b3fb21929fd525 |
|
BLAKE2b-256 | a91482ed9259ea072316afcb95d53259f0c831909a856c3c4b2e7f065615569a |
Hashes for pyamdgpuinfo-1.0.5-cp37-cp37m-manylinux2014_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 36ce39d3cbe204c4864ce2a86000244b04fd518f965b12412d2cb2481188c77f |
|
MD5 | 09a5277ad04059776f6806363d386150 |
|
BLAKE2b-256 | 385139278b8b85aa96c554e9a0e15976cfbd3f5032185b06f803ba9788680753 |
Hashes for pyamdgpuinfo-1.0.5-cp36-cp36m-manylinux2014_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | f06a1f1706aef4090492c072025dc8fb86e0496396657e81427b12405c77d946 |
|
MD5 | aa2d9eed4ace5e8912cf57399a0d37b3 |
|
BLAKE2b-256 | 2d9f92eed6e9e3e5c9831420ba1222e5857aae36390611e06acda35d3695e87a |
Hashes for pyamdgpuinfo-1.0.5-cp35-cp35m-manylinux2014_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | ac677f1aab8d1a0dd6eaeae4c72389634097f8bef9be22e376e461e6ee7fdb54 |
|
MD5 | b99a1b0ab913bc9b437eb79f74aac9d5 |
|
BLAKE2b-256 | e5c4dc5b5cd3ef7985b9f5045a39709cb2c790f6586e1ecc38c40fe08569583c |