FX3000 on Amazon International

[vvolkman]

Did I miss something or is the Quadro FX 3000 only about 2% faster than than the Quadro FX 2000?

We're a UG shop so I could easily recommend Quadro FX 2000 over Quadro4 980XGL, since the increase is 2x rather than 2%.

I guess you really are getting an FX 2000 with Genlock and other video transmitter gear. Seems disingenuous to call it FX 3000. More like FX 2000 All-in-wonder

:wtf:

http://www.amazoninternational.com/html/news/280703-1.gif

[wilson]

I wonder if the increased memory bandwidth helps with high-resolution displays (like 3840x2400). Send me one and I'll test it for you.

Try testing at >=2048x1536 and see if there's a bigger diff.

[MUYA]

Amazon international seem to be happy with it however the full picture will only be known after a week or so, because thats how long Amazon International have said it will take to finish. We can get a better idea then.

Cheers
MUYA

[ricercar]

I'm stunned the Quadro NV30-35 performance is so close when the GeForce NV30-35 difference is so wide! That performance graph lends credence to the rumors that the NV3x family is simply the same core tweaked in different directions. Perhaps the first iteration NV30 was close to the max OpenGL performance that is possible with 3x architecture.

[Uttar]

Originally posted by ricercar
I'm stunned the Quadro NV30-35 performance is so close when the GeForce NV30-35 difference is so wide! That performance graph lends credence to the rumors that the NV3x family is simply the same core tweaked in different directions. Perhaps the first iteration NV30 was close to the max OpenGL performance that is possible with 3x architecture.

I can show you a billion of graphs showing the FX5900 losing to the FX5800, if you wanted :)
Main areas it loses in are where sheer clock power truly matters, and no updates were done for the NV35 core:
1. Vertex Shaders
2. Texturing
3. Triangle Setup ( okay, it's NEVER the bottleneck but eh! :) )
4. Number of pixel outputs

And guess what... That's exactly what workstation benchmarks are after! Oh, and they're also sometimes after sheer CPU power, which is reduced due to immature drivers. So it's already a miracle IMO the NV35 is faster...

A good example of small differences between NV30 and NV35 is Doom 3 with no AA/AF. The difference is minimal, but the NV35 still wins thanks to its superior Pixel Shading abilities. If PS wasn't that much better, they'd easily lose to the NV30. AA heavily changes the picture, of course!


Uttar