New cad machine time again. Inventor 2000 part assemblies and Max design.
Well I think that the CPU is a certainty with a 4770k but now the graphics card. Does anyone have any opinions, a NVIDIA GeForce GTX 780 Ti seams to tick all the boxes and has one of the highest g3d pass marks, even faster than the k6000. I know it not all about g3d but all opinions welcomed.
Regards
Warren.
My work machine is just over 1 yr old and can't justify a new card but it would be on my wish list.
Almost any nVidia video card with a number XY0, where X is 5 or higher and Y is 6 or higher will handle Inventor 2014 very nicely. Almost any ATI video card with a number XYZ0, where X is 4 or higher, Y is 7 or higher and Z is 5 or higher will do very nicely as well.
For reference, when I upgraded my computer at home last summer, I got an i7-4770k and an SSD ... but I just went ahead and kept my ATI 4870's, because they're still pretty much killing it with Inventor, my CAM software, and most of the games that I'd want to play.
Rusty
GTX Titan or GTX 780 TI, no benefit in Revit. It will not help in hardware acceleration either. I have both types of card and when you go into options it tells you so. It's a shame that Revit does not take advantage of these super cards. Rendering a file on a 24 core, 32gb, two GTX Titans 5 hours. This is just nuts!!
@Anonymous wrote:
Rendering a file on a 24 core, 32gb, two GTX Titans 5 hours. This is just nuts!!
thats because Revit has never used the video card for rendering
its always been CPU and RAM
DarrenP
Did you find this post helpful? Feel free to Like this post.
Did your question get successfully answered? Then click on the ACCEPT SOLUTION button.
But if you have one of the cards that Revit mentions, you get hardware acceleration. If you don't use one of these cards it says that you are going to use software acceleration.
That's for live navigation and viewing inside the program, not rendering. Somewhat off-topic, but your core count vs. RAM numbers are a bit off for rendering as well. With "only" 24 cores (48 in this case, as most people have hyperthreading enabled) and 32 GB RAM you end up with around 0.5 GB of RAM per core, assuming its *all* available for use. For multi-core rendering you want to be in the range of 1 - 2 GB per core, with higher numbers for complex scenes/materials/lighting/etc. I suspect you might find rendering times decreasing if you experiment with turning off hyperthreading, but I don't recommend doing that on a working basis. There's a host of other things which can impact rendering times which are easier (and safer) to implement.
Thanks for all the information. I have two opteron 12 cores so there is no hyperthreading here. Still will be off if there is 2gb per core. Thanks anyway.
Can't find what you're looking for? Ask the community or share your knowledge.