Compute System Simulator: Simulation data and script for plotting Figure 6 - reliability scan
收藏官方服务:
资源简介:
The python script has an option to either generate the data or just plot it. Becuase the software is still closed source, it should only be used for plotting. Therefore, the data are also included to generate the plot and the script has run_simulations set to false. This is a scan of mean time to failure, mean time to repair, and requested job size (# of GPUs) for a GB200 system. For these, the values are: allocation_sizes = [72, 68, 64, 60, 56, 52, 48, 44, 40] # gpus mttf = [100, 200, 400, 800, 1600, 3200] # days mttr = [1, 2, 4, 8, 14] # days resulting in 270 simulations. Requirements to run the plotting script: python3 matplotlib numpy
提供机构:
Zenodo创建时间:
2025-04-24



