Skip to main content

TUNA

Tuna is a distributed tuning infrastructure that provides tuning through automated Jenkins pipelines and SLURM scalable architecture. For MIOpen customers we provide pre-compiled kernels.

Prerequisites

Install python3.9 ` apt-get update && apt-get install software-properties-common add-apt-repository ppa:deadsnakes/ppa apt install python3.9 `

Install pip for python3.9 ` wget https://bootstrap.pypa.io/get-pip.py -o get-pip.py python3.9 get-pip.py rm get-pip.py `

Install MySQL server ` apt-get install mysql-server `

Enable the service ` systemctl start mysql `

Install ipmitool ` apt-get install ipmitool `

Setup passwordless ssh between all machines, for example: ` ssh-keygen -t rsa ssh-copy-id <user>@<ip-address> ssh <user>@<ip-address> ` For the tuning cycle, every machine needs to be able to access every other machine through passwordless ssh.

Installation

Clone the repo using ` git clone <repo url> ` Enter the Tuna directory ` cd MITuna ` Create a virtual envornment, and activate it (by sourcing its activate script) ` virtualenv -p python3.9 myvenv source myvenv/bin/activate ` Install the required dependencies: ` python3.9 -m pip install -r requirements.txt ` The above assumes that Tuna lives in the home directory and the virtual environment was created using the command indicated above.

Add the following environment variables to a local file and then source the file: ` export TUNA_DB_USER_NAME=root export TUNA_DB_PASSWORD=<password for root> export TUNA_DB_HOSTNAME=localhost export TUNA_DB_NAME=<database_name> export gateway_ip=<gateway_ip> export gateway_port=<gateway_port> export gateway_user=<gateway_user> `

All machines used in the tuning process must have ssh-keys enabled. MITuna needs to have all-to-all machine communication available and passwords must not be required at run-time.

Run the setup scripts: ` python3.9 setup.py develop `

The root tuna folder needs to be appended to the PYTHONAPTH: ` export PYTHONPATH=/<path_to_MITuna>/:$PYTHONPATH `

To create the database run the following script: ` ./tuna/miopen/db_tables.py `

The installation and setup are now complete. To start a tuning cycle, please follow the steps documented in [TuningCycle](https://github.com/ROCmSoftwarePlatform/MITuna/blob/develop/doc/src/TuningCycle.md)

Code formatting

MITuna used yapf for code formatting: ` yapf -i --style='{based_on_style: google, indent_width: 2}' --recursive tuna/ yapf -i --style='{based_on_style: google, indent_width: 2}' --recursive tests/ `

Static code analysis

In order for a PR to be accepted the following pylint command needs to result in 10/10 analysis: ` cd MITuna/tuna pylint -f parseable -d duplicate-code --max-args=8 --indent-string ' ' *.py ` ## Coverage

To extract coverage percentage from the underlying code base prior to opening Pull Request:

` cd MITunaX ./tests/covscripts/coverage_script.sh `

Metadata

Release files for MITuna 1.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for MITuna 1.0
File Size Uploaded Provenance Actions
MITuna-1.0.tar.gz 39.6 kB Without provenance Details for MITuna-1.0.tar.gz

Release files / MITuna-1.0.tar.gz

Download URL MITuna-1.0.tar.gz
Size 39.6 kB
Tags Source
SHA-256 checksum
How to use checksums
b86d5800d2e185c0aa424734489c62d23aee229edab3b57e15bd33f76bf0a72a
BLAKE2b-256 checksum
How to use checksums
1ef5be6fb5f5a7b0c7ea59cd16b824a7b68cdeac45985fd9a17995e0b3547484
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/4.0.2 CPython/3.9.16

Security for MITuna 1.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

1.0 This release

1 release file No provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page