Some Tips to Speed up Dynamic Models Running Time
Increasing numbers of cores use in any dynamic model run might not be the most practical solution to boost running times without investigating run time chopping and convergence problems that would slow down model performance allot by all means.
When the dynamic model is running smoothly with few chopping time steps then you will be in better position to get more speed by adding more cores in the simulation run. However there is a limit with respect to the maximum numbers of cores can be used per run where if you exceed it the model running speed will drop down due to more nodes communications are taking place.
The following slides are discussing some of key items that would assist you to obtain additional running speed, to allow you to carry out more tuning runs for better history match achievement:
Sr Reservoir Engineer
3 年Totally agree on this. A good model should have a good physics-solution without much problem on chopping time and convergence related issue. which means, all the input data must be properly QC-ed and should give a ‘consistent-story line’ among all the data. Other facilities for example no of cores, GPU, etc will later on support the run time coz this is how numerical simulation works..