Installation
Pre-Installation Instructions¶
1. Set Up Ubuntu VM:¶
- Start with an Ubuntu VM equipped with Nvidia GPUs.
2. Install Dependencies:¶
-
For Docker Benchmarks: Install Docker, Nvidia Docker (for GPU support), git, Python with Virtual Environment, and Nvidia drivers.
-
For Direct Machine Benchmarks: Install Tmux, git, Python with Virtual Environment, and Nvidia drivers.
3. Setup VM With Script (requires git to clone repo):¶
-
Docker Benchmarks: Execute
./setup_vm_docker.shfrom the cloned Git repository. -
Direct Machine Benchmarks: Execute
./setup_vm_tmux.shfrom the cloned Git repository.
Installation Instructions¶
Follow these steps to set up iris-gpubench:
1.Clone the Repository¶
Start by cloning the project repository:
git clone https://github.com/bryceshirley/iris-gpubench.git
cd iris-gpubench
2.Set Up a Virtual Environment¶
Next, create and activate a virtual environment:
python3 -m venv env
source env/bin/activate
3.Install Dependencies and iris-gpubench Package¶
a. Finally, install the package along with necessary dependencies:¶
pip install wheel
pip install .
b. (For Developers)¶
pip install wheel
pip install -e .
-e for editable mode, lets you install Python packages in a way that
allows immediate reflection of any changes you make to the source code
without needing to reinstall the package.