search

radical-collaboration / CyberManufacturing

CDSE Multi-scale CI Project
1 stars 0 forks source link

Caption for Figure 10 #54

Open shantenujha opened 6 years ago

shantenujha commented 6 years ago

"Figure 10: Average time taken to run the PBM simulation which consisted of 25 seconds of mixing time and 75 seconds of granulation time at different core configurations. There was a steady decrease as the number of MPI processes were increased, but the improvements on increasing the OMP threads were not that significant."

i. which machine/resource? ii. why the plateau after 16 cores? How many cores does the node have ? Is the plateau due to multinode parallelism?

It appears there might be changes in the configurations (OpenMP+MPI) in addition to the number of cores employed. This should be avoided. I would prefer to see "strong scaling" and "weak scaling" plots.

csampat commented 6 years ago

The simulations was performed on stampede2. This has been mentioned in the methods sections, I shall discuss the changes in this as well as figure 6 and 7 with @yukteshwar and @FranklinBetten and modify them