NVIDIA Quadro FX 5800 £2,673?

Grazie wrote on 11/28/2012, 9:32 AM
NVIDIA Quadro FX 5800 by PNY Graphics card - 4 GB - GDDR3 SDRAM £2,673?

Is this over-the-top for VP12?

Grazie

Comments

megabit wrote on 11/28/2012, 9:41 AM
This is overpriced and won't even work with Vegas (old architecture - pre-Fermi - compute index below 2).

Piotr

AMD TR 2990WX CPU | MSI X399 CARBON AC | 64GB RAM@XMP2933  | 2x RTX 2080Ti GPU | 4x 3TB WD Black RAID0 media drive | 3x 1TB NVMe RAID0 cache drive | SSD SATA system drive | AX1600i PSU | Decklink 12G Extreme | Samsung UHD reference monitor (calibrated)

TorS wrote on 11/28/2012, 9:55 AM
Hi Grazie,
JohnnyRoy said in a different thread: "The Quadro 4000 seems to be the sweet spot as far as price/performance goes. Yea the 5000 is a bit faster but it cost twice as much and it is not twice as fast. It only has 352 CUDA cores compared to the 4000's 256 CUDA cores so I couldn't see spending double for it. I'm really happy with my Quadro 4000 and it has performed flawless with Vegas Pro 12.0."
megabit wrote on 11/28/2012, 9:57 AM
Please do not confuse the old Quadro FX 5x00 with the Quadro 5000 (Fermi)...

Piotr

AMD TR 2990WX CPU | MSI X399 CARBON AC | 64GB RAM@XMP2933  | 2x RTX 2080Ti GPU | 4x 3TB WD Black RAID0 media drive | 3x 1TB NVMe RAID0 cache drive | SSD SATA system drive | AX1600i PSU | Decklink 12G Extreme | Samsung UHD reference monitor (calibrated)

Grazie wrote on 11/28/2012, 10:23 AM
Eh? Me or TorS, Piotr?

I'm confused now.

G
john_dennis wrote on 11/28/2012, 5:04 PM
Grazie,

The one review of that FX5800 card was in 2009. It's probably too old to consider today.
Serena wrote on 11/28/2012, 5:42 PM
Grazie, I'll join you on the panel of the confused! This is an area of specialised knowledge and what is needed is a definitive spec. Looking around I came across this user report in relation to the Quadro 4000, which is clear (sort of) but also (in the V12 context) confusing:
"I did a lot of research on Qaudra Fermi Series and their semi twins GeForce GTX 400 & 500 Series. On the papers, they look almost identical. But the key difference is , GeForce is engineered for FPS (Frame rate in games), while Quadro is engineered for display complex model in 3D programs. Plus all 3D programs have designated driver for Quadro series, and will not help you if you have a software issue with unsupported GPU card. GeForce GTX 400 series has some display issues with 3Ds Max at the time I was doing my research(6 months ago), I'm not sure it got solved or not. Gaming cards tent to have display issue like chappy rotation, missing lines and crashes, if you have complex model or a large scene.

Majority of rendering engines are GPU base now, so you might not need this mid range Quadro if you don't deal with modeling much. You still need a quadro card for setting up the scene. Nowadays the common setup for rendering pro is, get a entry level quadro fermi card for display and modeling; and get one or two Geforce GTX card for rendering. GeForce GTX are in PCI slots but not connected to any monitors. And only install Quadro driver for display. Rendering engines will seek out all available GPU on your machine.

Bottom line, if you're a 3D pro who do a lot of modeling then this card will save you time and headache = money. If you're setting up a mini render farm, then Geforce GTX is more cost efficient. If you're like me who do a bit of everything, modeling, rendering, texturing and gaming (this card does pretty well on HD games), then this card is perfect for you. It does get pretty hot, make sure your machine have a good air flow. I have two 80mm fans (with fan speed control) to keep it cool. "
JohnnyRoy wrote on 11/28/2012, 9:09 PM
> "Please do not confuse the old Quadro FX 5x00 with the Quadro 5000 (Fermi)..."

Yea, there is the Quadro FX which are old cards and the plain ole' Quadro which are the new cards. You want this card:

PNY Quadro 4000 2GB PCI-Express x16 Professional Graphics Card for £616.66

~jr
Grazie wrote on 11/29/2012, 1:27 AM
O...K...

So a card that is over 3x the price is outdated, weak for Vegas and not having the Fermi technology

Is this SCAN UK the same as that MISCO offer?

Grazie

Serena wrote on 11/29/2012, 5:52 AM
Grazie, I believe that is the same card. Does have Fermi tech. Gets confusing when the various suppliers don't give it the same name. We have NVIDIA, PNY and Leadtek, but I presume despite the variations in nomenclature it is the same thing. And not cheap. What I don't see is whether it is worth the money compared to, say, the GeForce GTX560Ti. Be nice to have a quantitative measure of relative performance when using Vegas 12. One doesn't wish to find out after spending $900 that the changes are marginal. Yes, Fermi architecture is recommended for V12, but that on its own is a vague statement.

EDIT: cheaper from PNY Quadro 4000 2GB PCI-Express x16 Professional Graphics CardB&H[/link]
JohnnyRoy wrote on 11/29/2012, 5:53 AM
> "Is this SCAN UK the same as that MISCO offer?"

Yup, that's the one.

~jr
Serena wrote on 11/29/2012, 6:49 PM
Interesting conundrum. Is there an excellent reason for buying this card? If so, buy from B&H for US$700 or locally (Leadtek PCIe Quadro 4000) for A$950 (US$988). Customer feed back on the B&H site mentions card failure and poor PNY response. Guess I would really like an answer to "excellent reason", since the others fall into place.
TorS wrote on 11/30/2012, 1:23 AM
Sony Creative Software have always been reluctant to recommend hardware and I don’t suppose they will change that policy. The fact that NIDEA recommends fermi architecture for professional work may be strategic. They want to sell the more expensive cards. But the fact that Sony quotes this recommendation in their Vegas Pro 12 promotional material is significant. So even if they do say “GeForce 4xx and beyond” they put “Fermi” very prominently, thus unrecommending GeForce 6xx and beyond. That, to me, is an excellent reason, if in a roundabout way.
I think the value of benchmark comparisons is overrated. Quadro cards are made for multimedia, industry and science. GeForce cards are made for gamers and overclockers. That, too, is an excellent reason to put a Quadro card in a video editing machine.
Grazie wrote on 11/30/2012, 2:29 AM
Sony Creative Software have always been reluctant to recommend hardware and I don’t suppose they will change that policy.
Yup, that'll be the reason for SCS. NP from me here on them being "cautious".

The fact that NIDEA recommends fermi architecture for professional work may be strategic.
Maybe this nVidia link would have swayed me PLUS the rationale behind the vastly more muscular code being written for OFX and CUDA. There is a point to being "coy" about suggesting what GPU card to advise on. There's another issue to getting me what I want when I want it. Yes, I am sounding "sour". Why, 'cos I wasn't aware of the "facts". Would have been good to have been offered some reasonable "convergence" of technology and my budget? From where I am seated, I had neither.

But the fact that Sony quotes this recommendation in their Vegas Pro 12 promotional material is significant.
And there we have it: significant. I wasn't aware just how significant it could have been. I'm, like others here, have been punching in the dark.


So even if they do say “GeForce 4xx and beyond” they put “Fermi” very prominently, thus unrecommending GeForce 6xx and beyond. That, to me, is an excellent reason, if in a roundabout way.

Again, how "prominent" is prominent. When does "prominent" become: "If you want the bells and whistles of OFX and CUDA and fuller frame rates, do NOT purchase anything from nVidia that ISN'T Fermi." - right, like that is going to happen. But the Elephant was in the room and it would appear that the IT pros here about were Fermi-aware.

I think the value of benchmark comparisons is overrated.

Well maybe, we just needed the RIGHT or appropriate Benchmarks. Something from one of the IT Pros here saying that with Fermi you get A and with GTX you get B. What I am reading is that there is a suggestion of a pejorative purchase decision for those going with GTX nVidia, ie non-Fermi cards. And again how would we have known?

Quadro cards are made for multimedia, industry and science. GeForce cards are made for gamers and overclockers.

I ain't shooting the Postman here - ie YOU! LOL! - but rather a plea for more clarity and openness on the choice making. Again the Elephant in the room has not been, up until now, I think, been so carefully focused upon and intelligently debated.

That, too, is an excellent reason to put a Quadro card in a video editing machine.
Relating to the IT awareness of the IT Pros, sure. But to "Joe Shmoe" = ME, here it means little.

Is anybody here understanding what I am trying to unravel here?

TorS, great post, I mean it.

Cheers

Grazie

megabit wrote on 11/30/2012, 4:04 AM
I'd only add that while the Vegas GPU accel HW compatibility did have to start at a certain HW level (Fermi in this case), and is not supposed to be backward-compatible (i.e. pre-Fermi cards will not work) - one would expect the forward-compatibility should be a matter of course. So, the Kepler cards should work...

Do they?

Piotr

AMD TR 2990WX CPU | MSI X399 CARBON AC | 64GB RAM@XMP2933  | 2x RTX 2080Ti GPU | 4x 3TB WD Black RAID0 media drive | 3x 1TB NVMe RAID0 cache drive | SSD SATA system drive | AX1600i PSU | Decklink 12G Extreme | Samsung UHD reference monitor (calibrated)

farss wrote on 11/30/2012, 4:10 AM
"Is anybody here understanding what I am trying to unravel here?"

Yes, been down the same road and lane and path and round and round in circles.

We need to be careful not to put much weight on words written about applications that use OpenGL or similar. Vegas does not use OpenGL, it is not a "graphics" application in the same way as nVidea mean. It also does not use the video card in the same way as Resolve or Scratch use it.

Bob.
JohnnyRoy wrote on 11/30/2012, 3:33 PM
> "Vegas does not use OpenGL"

Be careful there... It does if you add Boris Continuum Complete or other plug-ins that use OpenGL. I specifically bought the Quadro 4000 because I use BCC8 a lot and wanted to be sure that I had superior OpenGL support for when I use it.

~jr