Hi Joshua, As you understood, interaction of a vehicle with deformable terrain will always be more computationally expensive than with rigid terrain. Having said that, the SCM option for deformable terrain is the fastest among those offered in Chrono; a 4-wheel vehicle can run in real time on a commodity computer. You can see more details about the latest incarnation of SCM in Chrono, including performance measurements, in this paper<https://www.researchgate.net/publication/368353573_Real-time_Simulation_of_Ground_Vehicles_On_Deformable_Terrain>.
The SCM implementation in the latest Chrono code was not modified significantly from what was in release 8.0. Having said that, I strongly encourage you to use a newer Chrono version (9.0.1 or even the current development branch) as there were other changes, fixes, and new features implemented. Furthermore, it is much easier for any of us to provide support for the latest Chrono version. Finally, you indicate that you are planning on running multiple parallel, independent runs. That is certainly what you should be doing for ML training. However, your comment about concerns as you try to do more parallel simulations suggests that you are running all of them on the same node (processor). If that is correct, it is not surprising that overall run times are going down – all these processes will fight over the same hardware resources. The proper way to do scalable high throughput computing is to distribute these parallel runs over the nodes in a cluster. On a single node, you will always see continuously degraded performance as you run more an more processes, regardless of the computational efficiency of any individual simulation. --Radu From: projectchrono@googlegroups.com <projectchrono@googlegroups.com> On Behalf Of Joshua Philip Martin Sent: Thursday, November 21, 2024 3:45 PM To: ProjectChrono <projectchrono@googlegroups.com> Subject: [chrono] Tips for Speeding Up SCM Calculation? I'm looking at using Chrono for reinforcement learning, specifically because of the availability of its SCM contact model which is unique among simulator options. Understandably, the SCM model is going to be slower than the rigid contact model and will increase in complexity as the resolution of the SCM terrain increases. However, for RL applications, I'd like/need the system to run several environments in parallel, ideally at very fast speeds to increase the amount of data I can get in a given time. Long story short, does anyone have any recommendations for increasing the speed/reducing the load for SCM calculation? Since using something like gym_chrono<https://urldefense.com/v3/__https:/github.com/projectchrono/gym-chrono__;!!Mak6IKo!N_GgbHAd6ox8zXoYgK3avQSZxLoS7cAkw0skO2C8dvwkovvjpUZ8IEoCs_C-eA_HoIvvZqLIS509P6NU$> requires a single CPU thread for each sim done in parallel, I'm definitely concerned about compute times increasing as I try to load more parallel simulators or more complex SCM geometries. I'm already creating moving patches at the agents' wheels as is done in the demos, though I'm not sure what I can do beyond that. I've only seen one paper published that uses SCM at the training stage<https://urldefense.com/v3/__https:/arxiv.org/pdf/2408.09253__;!!Mak6IKo!N_GgbHAd6ox8zXoYgK3avQSZxLoS7cAkw0skO2C8dvwkovvjpUZ8IEoCs_C-eA_HoIvvZqLISzI5Wfww$>, but that means it is possible to do what I'm looking to do (though potentially slowly). In case it matters, I'm also still using Chrono 8.0.0 for this setup, as was recommended by the gym_chrono repository. I'm not sure how much their custom Chrono branch matters, though, since I'm making my own robot objects; if Chrono 9.0.0 offers dramatic speed increases in SCM calculation, maybe I can switch over? -- You received this message because you are subscribed to the Google Groups "ProjectChrono" group. To unsubscribe from this group and stop receiving emails from it, send an email to projectchrono+unsubscr...@googlegroups.com<mailto:projectchrono+unsubscr...@googlegroups.com>. To view this discussion visit https://groups.google.com/d/msgid/projectchrono/6e67eda0-e591-4ae0-9546-dea5454c2377n%40googlegroups.com<https://urldefense.com/v3/__https:/groups.google.com/d/msgid/projectchrono/6e67eda0-e591-4ae0-9546-dea5454c2377n*40googlegroups.com?utm_medium=email&utm_source=footer__;JQ!!Mak6IKo!N_GgbHAd6ox8zXoYgK3avQSZxLoS7cAkw0skO2C8dvwkovvjpUZ8IEoCs_C-eA_HoIvvZqLIS0FAr5O6$>. -- You received this message because you are subscribed to the Google Groups "ProjectChrono" group. To unsubscribe from this group and stop receiving emails from it, send an email to projectchrono+unsubscr...@googlegroups.com. To view this discussion visit https://groups.google.com/d/msgid/projectchrono/CH3PPF46CDC218522F472479DE717D57680A72A2%40CH3PPF46CDC2185.namprd06.prod.outlook.com.
