Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/andylamp/gpurelperf
A handy utility to find relative GPU performance quickly in multi-gpu boxes
https://github.com/andylamp/gpurelperf
gpu load-balancing mxnet
Last synced: 2 months ago
JSON representation
A handy utility to find relative GPU performance quickly in multi-gpu boxes
- Host: GitHub
- URL: https://github.com/andylamp/gpurelperf
- Owner: andylamp
- License: apache-2.0
- Created: 2018-06-26T20:51:40.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2018-06-26T21:01:59.000Z (over 6 years ago)
- Last Synced: 2024-08-01T22:42:23.283Z (5 months ago)
- Topics: gpu, load-balancing, mxnet
- Language: Python
- Size: 12.7 KB
- Stars: 3
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- Awesome-MXNet - gpurelperf
README
# GPU relative performance
This is a solution to a problem that if you probably have not
even encountered if you are not running a multiple different
GPU's in one way of another... of course, this package tries
to solve that!
# First world problems.
Be it for Neural Network training or Mining you need to have
a way to distribute the load equally to each GPU. For example
although `mxnet` *supports* multiple GPU's the scheduler currently
does *not* know the relative performance of each GPU available
and either you have to do it by hand or a uniform load
distribution is applied to each GPU (more [here][1]).
# The (quick) solution.
Now, I've thought of how to tackle this problem with
micro-benchmarks, load measurements and so on... but I like
the KISS principle and hence I am using something simple with
minimal overhead which works surprisingly well in practice.
Basically what I want to accomplish is to get a relative
performance of each GPU against the lowest performing one
currently installed; this can be achieved by using
[Geekbench][2] aggregated CUDA benchmarks scores and construct
a relative performance index for the currently installed GPU's.
The main gist of this solution is fetch the raw `JSON` CUDA
benchmark data from Geekbench, parse it, find the GPU's
installed in the system while matching and normalizing their
performance using the CUDA benchmark scores. These results
can then be immediately used in the `mxnet` scheduler as
percentages.
# Requirements
Currently this exists as a source distribution and requires `nvidia-smi`
to be installed -- which if you are using an NVIDIA GPU with either
Windows or Linux it should already be installed. Roughly, the requirements
are as follows:
* Python > 3
* Nvidia Drivers (in scope)
Unfortunately, MacOS does not have `nvidia-smi` yet, but a workaround exists
which I will probably include in a future update.
# Usage
Using this package is easy, first of all do:
```
$ pip install gpurelperf
```
Then, after install completes you can use it as a normal package:
```python
from gpurelperf import get_sys_cards
print(get_sys_cards())
```
## A quick example with mxnet
```python
import mxnet
from gpurelperf import get_sys_cards
# returns a tuple with the ratios
(wl_list, gfx_list) = get_sys_cards()
# then you would set the work load list as such
work_load_list=wl_list
```
# License
This project is licensed under the terms and conditions of the Apache 2.0 license.
[1]: https://mxnet.incubator.apache.org/faq/multi_devices.html
[2]: https://browser.geekbench.com/cuda-benchmarks