3D Modelling Lag with Quadro 5000

3D Modelling Lag with Quadro 5000

Anonymous
Not applicable
1,328 Views
1 Reply
Message 1 of 2

3D Modelling Lag with Quadro 5000

Anonymous
Not applicable

Morning,

 

I work for a small temporary works company and we have just upgraded out desktop towers to enable us to start modelling in 3D. We have been doing so recently using our current setups which have been less than ideal. Fine for small modelling work but severely struggles with larger or more complicated models.

 

We purchased new HP Z840 units (http://www8.hp.com/us/en/campaigns/workstations/z840.html) and then added some new SSDs along with the new NVIDIA Quadro 5000 (http://www.nvidia.com/object/product-quadro-5000-us.html). We assumed this would enable us to open and work with any and all models smoothyl, bar the exceptionally complex ones. However, we have found that this has had little to no effect on the 3D modelling lag that we were experiencing, and whilst running a model and monitoring the GPU usage we found it using only 10-20% of the GPU. 

 

My query is really just if anyone else has experienced this, and if there is anything that can be done to make AutoCad access and use the full capability of the GPU? Or is AutoCad limited in the resources it can use?

 

Thank you in advance for any help.

0 Likes
1,329 Views
1 Reply
Reply (1)
Message 2 of 2

Anonymous
Not applicable

good morning,

sorry to be a little late on this,

the graphics card is listed as quadro 5000,

but there are multiple cards with this similar name.

which one do you have?

quadro 5000

quadro 5000M

quadro K5000

quadro K5000M

quadro M5000

quadro M5000M

 

all of these are on the list for cerified drivers with autocad 2016,

and more than likely the wrong one might have been used.

 

I also found that the most current driver for my card works less than perfect.

so i went back to my older driver and it works fine now.

but i would start with current and perform a good test with 3d orbit.

 

also the most likely bottle neck is RAM on each system,

you want at least 6 gigs per core, for full functionality,

and if serious modeling 12 gigs per core and up.

 

hope this gets you going in the right direction

0 Likes