Hello.
Is there any option to speed up the software on large models (graphics display and calculations) or an option to allows ROBOT to use more hardware capacity from my PC?? (more processors simultaneously, more graphic memory, etc...).
What are the limitations in memory and numbre of processors? Since at the moment Robot seems to take the same time in the analysis regardless of the number of processors or the physical available memory of the computer used.
In large models, the graphic display seems extremely slow even when the computer has a powerful graphic card.
Thanks!
Regards
Hi
As concerns GUI acceleration check \ set it here:
For large model calculations please use multithreaded solver :
Rafal,
I don't agree that using multi threaded solver is a valid solution. It cannot be used for footfall or other such dynamic analysis, which is when you really need massive processing capability.
Besides, I think most analysis should be confirmed as valid wih another solver prior to using the multi threaded one, as it only has limited instability reporting. This means you still need to use the slower solvers until your model is completely sorted. However, once the model is completely sorted there is little need to re-analyse, making the multi treaded solver kind of redundant.
Just my 2 cents.
Cheers,
Tom,
I don't agree that using multi threaded solver is a valid solution. It cannot be used for footfall or other such dynamic analysis, which is when you really need massive processing capability.
I know you know about footfall and multihtreded but other dynamics? WHich? Any example? If you have any of such examples please register issue thru subscription center.
Besides the question was about using more processors \ cores. There is only one solver in Robot using more then one of them -> multithreaded.
Besides, I think most analysis should be confirmed as valid wih another solver prior to using the multi threaded one, as it only has limited instability reporting. This means you still need to use the slower solvers until your model is completely sorted. However, once the model is completely sorted there is little need to re-analyse, making the multi treaded solver kind of redundant.
You are right about fixing model but re-running model with changed ie sections will be faster with multithread.
Another example: running model with seismic analysis : ie Sparse M solver - 24h <-> multhitread ~ 5h ; checking model can be done with sparse M deactivating (setting as Auxiliary) all time consuming cases for the time of model check.