Hello to all,
for my graduation work I am about to work with MCD in NX 10. It is a plant with many parts. I found out that when I simulate the plant in the current state I have worked out, the processor only works on one kernel. So my question is if there is any way to increase the performance of the simulation for example by outsourcing the calculations on more kernels. I work with an Intel Xeon CPU E3-1240 v5, 16 GB RAM and NVIDIA Quadro K2200.
Solved! Go to Solution.
Hello Alex and welcome to the Siemens PLM Community,
in terms of Performance we differentiate two aspects in MCD, these two are interconnected.
The first is the Visualisation (Graphics) Performance in NX, it is also the indicator how many frames per seconds you can get from the graphics window.
The Second one is the MCD Simulation Performance, it is the indicator how fast the MCD Simulation is running (e.g. realtime: time scaling factor 1.0 or with time scaling factor 0.5; half of real time).
Keep in mind these two visualisation and MCD simulation performance also influence each other (simplier assembly = faster MCD simulation...)
But coming back to your question and the Bad news: the MCD part is currently only using one Core and there is no possibility to change that other then running more instances of NX and assigning them to different cores.
Probably the good news: There are alot of Performance improvement possibilities in NX as well as MCD.
To Name some of them:
I would start with increasing the Simulation Step Time in the MCD Preferences to a level where you still have the accuracy you can live with and then looking into the other aspects.
Also If you have many Collision Bodies with many Rigid Bodies e.g. a product coming from an "Object Source" the simulation will naturally get slower over time, espacially if you dont remove the objects with an "Object Sink".
Also something to note is: if for example you want to simulate a material flow in your whole plant it probably should be noted that Plant Simulation is the better tool in doing so.