BASEMENTBasic Simulation Environment for computation of environmental flow and natural hazard simulationLaboratory of Hydraulics, Hydrology and Glaciology (VAW)ETH Zurich |
You are not logged in.
Pages: 1
Topic closed
hello colleagues
I ask because if I launch a simulation with a 36-core system, basement performs simulation with only 32 cores?
I'll explain better, I'm launching of simulations making use of virtual machines thanks to cloud computing.
To get faster results I'm using a machine with 36 cores to 2.9GHz per core,but basement only sees 32 cores
"- OpenMP>: Computation is executed in parallel using 32 threads"
because?
Parallel I launched the same simulation on an 8 threads per core 2.6GHz machine
"-> OpenMP: Computation is executed in parallel using 8 threads"
but not known a reduction of run time proportional to the computing power.
32 core: after 20min simulation ----> total run time: 4:18:00, RTS: 12
8-core: after 20min simulation ----> total run time: 6:58:00, RTS: 7.14
I wanted at this point to know what the limits of parallel computing? I report my experience as run time on the same simulation
Tr 200 run time:27 h RTS:1.85 machine:4 threads 3,5GHz
Tr 200 run time:16 h RTS:3.09 machine:32 threads 2,9GHz
Tr 200 run time:40 h RTS:1.25 machine:8 threads 2,6GHz
how to evaluate the right power to use? because the run time is not directly proportional to the computing power? there are limits beyond which the speed of simuazione can not go up?
best regards
Offline
this is a clone entry of this topic
Offline
Pages: 1
Topic closed