I’m planning to work in Visual Components designing warehouses. These are very resource-intensive simulations. Could you advise what’s more important for faster performance, CPU or GPU? Also, what computer specs would you recommend to avoid low FPS issues?
According to my understanding, the simulation of visual components is divided into two parts: data simulation mainly relies on single-core processing by the CPU, while model rendering requires GPU processing.
CPU with a high single-core performance is the main performance contributor from hardware side.
Hello Este,
I tried to do a comparison, I compared a computer with a GTX4060 graphics card, 32G RAM, and an i7-14700K processor with a computer with a GTX5090 graphics card, 32GB of RAM, and an Ultra 9 275HX processor, and I found that the results were basically the same with the same model, the same viewing angle, and the same rendering method (all of which were forced to use a discrete graphics card setting). The GTX4060 computer turned out to be even better than the GTX5090 computer, which made me wonder, and the VC’s CPU and GPU call rates were kept at a relatively low level (30-50% CPU usage, 30-45% GPU usage average, data from NVIDIA performance monitoring), which I don’t think is in line with most people’s perception that upgrading hardware can improve performance.
So, does this mean that this is the upper limit of VC applications, and is there any hope that the simulation capabilities will be improved in the future (more fully using the CPU or GPU or both)?
Thank you in advance for your reply, and apologize in advance if the issue is offensive. 
(The translation comes from the translator, hopefully it’s accurate)
GPU performance is rarely the bottleneck for VC simulations, so no wonder that using a better GPU made no change.
With every VC version there are new tools and functionalities added that increase performance. A few years ago it was not possible to create simulations at the scale that we regularly see now. Thus I am hopeful that again in a few years we can run simulations that were too demanding with current technology.
1 Like