quadro4000 vs gtx580

mmlxx · Lurker Joined: 10 Sep 2008 Posts: 24

Which better quadro4000 and gtx580? performance、stability and so on......

Erik_JE · Posted: Thu Oct 20, 2011 7:55 am GMT Post subject:

Quadro 4000. More stable and pricier.

GTX 580. Don't think there will be much viewport performance difference but should have better raw performance for gpu computations.

I'd go with a GTX 580 with 3gb memory.

fgillis · Posted: Wed Feb 01, 2012 10:21 pm GMT Post subject:

Hi all,

Am building a linux workstation for using Houdini and am still on the fence regarding gtx vs quadro.

SESI does not recommend using gtx cards, and many sites discussing the issue devolve into the "professional" Quadro camp vs the "gaming" GTX camp.

I've read the pros and cons... and based on manufacturer product descriptions, SESI recommendations, and because I will be using the card for "professional" use and not gaming, I should go with the Quadro card series.

However, it's hard to dismiss a card like the 3GB GTX580 which appears to be faster and have more memory than a much higher priced Quadro card.

Does anyone have extended experience using a GTX580 "professionally" in a real production environment running Houdini on a linux system? Can you share your experience? Modelling? Sims? High poly/point count viewport interaction? Stability? Driver support? Etc?

Thanks,

dennis.weil · Posted: Thu Feb 02, 2012 4:12 am GMT Post subject:

Hi Floyd,
I'm using a 3Gb GTX580 with Ubuntu Linux 10.04 in a professional production scenario for 4 months now and I've had no problems whatsoever. I also have some extended experience with the Quadro 4000 and I've been pretty dissapointed by its performance (especially considering its price). The GTX580 beats the Quadro 4000 in every aspect (also viewport performance). That being said I WOULD buy the best Quadro Card available to get the best performance and stability if I had the money. If GTX580 and Quadro 4000 are your only available choices I would recommend the GTX580 however.

Hope this is helpful.

-dennis

fgillis · Posted: Thu Feb 02, 2012 7:33 am GMT Post subject:

Hi Dennis,

Thank you for the information. Very interesting and helpful, (and what I was hoping to hear).

Anyone else who'd like to share their experiences, positive or negative, with using the GTX580 in a professional production environment running Houdini on Linux, (or Windows)?

Many thanks,

fsimerey · Posted: Thu Feb 02, 2012 7:36 am GMT Post subject:

Like dennis.weil, I'm using a GTX 580 - 3 GB on Linux Mint 11 in production.
All works fine and fast.
If you check the specs of Quadro 4000, it's like a GTX 285 except the ram (768 Mo on gtx vs 2 GB on Quadro).

My vote would be for the GTX.

fgillis · Posted: Thu Apr 12, 2012 1:44 am GMT Post subject:

Just a quick follow-up. I ended up buying the GTX580 3gb board for a new system build. Have been using it with Houdini on production work for a couple of months now and, so far, I'm very happy. It's fast and stable.

I am having some issues with scrubbing audio... but I doubt that's related to the graphics board.

This was the first system I've built, (since the Sinclair ZX-81), and it went smoother than expected.

some system info:
Asus P8 Z68-V/Gen3 mobo
Intel i7 2600K
Evga Geforce GTX 580
Debian Squeeze... currently running nvidia 295.33 driver

gronk33 · Lurker Joined: 08 Apr 2012 Posts: 13

This is probably a super noob question but when it comes to Mantra PBR IPR and Mantra PBR renders what makes the most difference when it comes to speed? GPU or CPU? Or both?

Is there a good explanation somewhere about how offline rendering and IPR work behind the scenes when it comes to one's hardware? I've read the "understanding mantra" section in the help docs. But it doesn't help me with practical choices like upgrading hardware.

For what it's worth, as a newcomer to Houdini I'm finding both IPR and offline rendering slower in Mantra than in Vray for Maya (hence my interest in hardware upgrades). But it could also be my general lack of experience with the software.

Thanks for any advice!