How to utilize full computer power?

[aysadic (Ayberk Sadic)]

Hello,
I am using AMS daily for my research. I use a workstation with 4.2 GHz CPU, a 64 GB RAM and an Nvidia GeForce 1060.

My question is that I run some benchmarks on AMS and check the load on each core, gpu and total RAM usage. To my surprise, AMS only utilizes a maximum of 12930 MB of RAM while using a maximum of 30% per core. Also it doesn't use any GPU at all when the model rendering is turned off. Lastly, the temperatures are barely above 40C degrees on CPU, GPU and RAMs, so we know it is not thermal throttling either.

I use AnyPyTools too but it barely improved the performance.

How can I utilize the rest of my computer to speed up the analysis? Because I don't see any bottlenecks on the hardware side... I want to allocate more RAM to the AMS.

EDIT: I have declared the priority property of the AnyPyProcess class as 65536 (app = AnyPyProcess(timeout=65536,priority=65536) to allow it use all available RAM. However it still doesn't go beyond 12930 MB of RAM usage. Let's say that is the maximum RAM required by my simulation then why I cannot utilize even a single core 100%.

Please advise

Best,

[dschander (Divyaksh S. Chander)]

Hi Ayberk,

AnyBody simulation runs on a single thread. And the RAM used really depends on the model. AnyBody uses as much RAM as it needs. Of course, having extra RAM allows you to run multiple simulations in parallel, where each simulation runs on its individual thread.

I am not entirely sure how CPU usage is reported but I can suggest you look at the clock speed of your CPU when you are running the simulation.

Best regards,
Dave

[aysadic (Ayberk Sadic)]

Hi,

Thank you for your response.

I am using 3rd party software to check the load of my cores. (i.e. CoreTemp). Even though it is single thread process and I am aware that Windows has an architecture to divide single core processes into multiple cores to not load a single core too much, The sum of individual CPU Core loads is not equal to near 100% levels.