Clone the repository with submodules
git clone https://github.com/UnaryLab/Primus.git --recursive
cd PrimusSet up a python virtual environment
curl -LsSf https://astral.sh/uv/install.sh | sh
uv venv --python 3.12.0
.venv/bin/activateInstall chopper
uv pip install third_party/chopperRequest access to Llama 3.1 8B. Make sure to export your huggingface key.
export HF_TOKEN=<key>This script will submit all jobs needed to build the container, collect the data, and plot figures.
./submit_all.sh account_nameOnce all jobs are completed, match the generated plots with their corresponding figures in the paper:
-
Figure 16
- a) deepseek_llama_straggler_per_gpu.png
- b) deepseek_llama_lead_and_throughput.png
- c) deepseek_llama_avg_pow_freq.png
-
Figure 9
- a) red_realloc_slosh_lead_and_throughput.png
- b) red_realloc_slosh_freq_pow.png
- add raw data
- add script for generating every figure, and its corresponding match in the paper