Artificial Intelligence Computing Leadership from NVIDIA
P40 and TeamCenter
Hi All We are running a series of test's for customers who use Teamcenter and Catia. The hardware is a Dell R730 2 x Nvidia P40's each server The processor is Intel Xeon GOLD 6148 2.40 GHZ The storage is NetApp full flash. We are using XenDesktop 7.15.3 and XenServer 7.6. The VDI machines are running Windows 10 64 Bit 1803 with P40-3Q card and vdi machines are running 8 vcpu's. We have large size car drawings. Our engineers are designing the model with the Catia and wiewing with Siemens Teamcenter Visualization. Our issue: when we rotate the model, 2 seconds delay on first click. But after the first click, we have no problems when we want to rotate the model. Our problem is that during the first click. This delay time increases when the model file grows. What can we do for this 2-second delay?
Hi All
We are running a series of test's for customers who use Teamcenter and Catia.
The hardware is a Dell R730 2 x Nvidia P40's each server
The processor is Intel Xeon GOLD 6148 2.40 GHZ

The storage is NetApp full flash.

We are using XenDesktop 7.15.3 and XenServer 7.6.

The VDI machines are running Windows 10 64 Bit 1803 with P40-3Q card and vdi machines are running 8 vcpu's.

We have large size car drawings. Our engineers are designing the model with the Catia and wiewing with Siemens Teamcenter Visualization.

Our issue: when we rotate the model, 2 seconds delay on first click. But after the first click, we have no problems when we want to rotate the model.
Our problem is that during the first click. This delay time increases when the model file grows.

What can we do for this 2-second delay?

#1
Posted 05/30/2019 10:31 AM   
First of all I would check the power state of the GPU with nvidia-smi to figure out if this may be related. You could for example run a youtube video in the background to see/force power state in P0 and try again. If the behavior is different then there are options to keep the GPU in higher power state. Another issue is the weak CPU. 2.4Ghz is much to low for Catia and you won't keep the GPU under load as the CPU is "overloaded" already. regards Simon
First of all I would check the power state of the GPU with nvidia-smi to figure out if this may be related. You could for example run a youtube video in the background to see/force power state in P0 and try again. If the behavior is different then there are options to keep the GPU in higher power state.
Another issue is the weak CPU. 2.4Ghz is much to low for Catia and you won't keep the GPU under load as the CPU is "overloaded" already.

regards
Simon

#2
Posted 05/31/2019 07:10 AM   
Hi, I have similar issue with PTC Creo. CPUs are Intel® Xeon® Gold 6146 12C, 3.20 GHz. Is it possible to force P40 power state to P0 with nvidia-smi? Regards, Károly
Hi,

I have similar issue with PTC Creo.

CPUs are Intel® Xeon® Gold 6146 12C, 3.20 GHz.

Is it possible to force P40 power state to P0 with nvidia-smi?

Regards,

Károly

#3
Posted 09/10/2019 06:40 AM   
Scroll To Top

Add Reply