Application Benchmarks: ParaView & VisIt
SpecViewPerf is a good start, but I felt the need to try some more ‘application’ benchmarks. To do this, I worked with the developers of two popular open-source visualization packages: ParaView & VisIt. I obtained a copy of the Supernova Dataset used in the recent Ultrascale Visualization paper, and extracted one default isosurface (in the median of the data) and constructed a simple flyaround. I disabled all attempts at level-of-detail, and forced displaylist generation. This way, the geometry is loaded to the video card and the problem becomes a pure geometry-pushing problem: How fast can it push triangles through the pipeline to the screen?
I used FRAPS to track the framerate over the runs, and wound up with results like this:
You can see easily how the Quadro5000 doubles the performance of the previous generation 4800 and 5600. One interesting thing to note here is how the 5600 and 4800 benchmark almost identically. During the early discussions about the 5000 and the Fermi architecture in general, NVidia gave us some background about the last few generations of hardware. With the sudden success of GPGPU and CUDA, and the advent of larger screens, much of their energy has been devoted to increasing the ability of the hardware to simply push pixels (Fill Rate). While this was great for many applications, raw geometry performance stagnated.
The Fermi architecture, combined with Quadro, brings a 5x increase in raw geometry performance. This is even easier to see in some Maya Benchmarks.
Dear Randall,
My company is involved in marine simulation (both for training and research purposes). We will upgrade to Presagis Vega Prime. We are currently debating the choice for a GTX 480 or Quadro 5000 for our image generator PCs in our simulators. Our Visual Database development workstations run on the Quadro GTX 4800.
I am curious to hear your advise!
Kind regards,
Martijn
@ Griffin Doh, you’re right.. I just checked the BIOS, I’ve got 2 Xeon 5550’s (8 cores) with HyperThreading enabled. I missed that. I’ll update the article.
Randall,
Thanks a lot for the review, please keep up the great work.
I was wondering if you could clarify the test setup, It was my understanding that Xeon 5550’s could only run 2-way, not 4-way like the 7500’s? Could you please shed some light on your system and its configuration.
Thanks a lot for your time
@ Matt It depends heavily on the application, and faster/newer is always better. But, if you’re in the CAD or 3d Graphics market then I’m pretty sure you would see some improvement.
Randall,
thanks for posting your assessment. I habe a pretty basic question as I am in no way an expert in graphics hardware. Do I benefit of the new GPU with all kinds of workstations or do I need to have a minimal configuration in terms of CPU, working memory etc. to really feel the difference between, let’s say, a 5000 and its predecessor, the 4800?
Just wondering if the Fermi based cards are Mac compatible…
@ Nick D Well, for reasons similar to why you would choose a Quadro4800 over a GTX285. The GTX285 has 240 cores [link], while the Quadro4800 only has 192 [link], but you can look at the benchmarks I posted and see how the Quadro4800 smokes the GTX285 in several benchmarks.
While more cores is nice, it’s not the sole indicator of performance. The Quadro cards are optimized to use the advanced graphics features not typically used by games, but frequently used by CAD & Visualization products. Because of these optimizations, a Quadro will always beat a GeForce when those features come into play.
Randall,
Other than drivers and memory, can you tell me why the 480 core nVidia GTX 480 would not be better than the Quadro 5000 with only 352 cores?
After waiting so long, I am really disappointed the Quadros don’t have the same AND more cores than the GTX 480…
I wonder if NVIDIA could give you a 6000.
@ chris True, but I don’t know of many professional types that are using Stereo on 3 monitors.
“Add in the support for the new OptiX 2, 3d Vision Pro technology” but you would need 2 cards to do stereo on 3 monitors? makes it expensive