I believe that there is a hidden potential in the nv30 and
- Thread starter K.I.L.E.R
- Start date
Mintmaster
Veteran
Back to the original topic of "potential" in NV30, I don't really see much. NVidia started to have great per-clock efficiency starting with the Geforce3, and it got even better (although marginally) with the GF4. Even though there were bandwidth problems, after I saw what the 9500 PRO was capable of I thought maybe bandwidth isn't as important as I had originally thought.
However, it doesn't seem like NV30 is near 85% faster than the 9500 PRO, even though you'd expect that from the clock rate difference. If the 9500 PRO ran at 500/500, it would probably outrun GFFX by some 20%, even though it would have similar bandwidth and fillrate. It seems like the NV30 architecture is flawed somewhere, much like the 8500 is not as efficient as the GF4.
I wouldn't mind seeing some 270/270 benchmarks from the GFFX to make this comparison. Maybe the Beyond3D article could include this?
However, it doesn't seem like NV30 is near 85% faster than the 9500 PRO, even though you'd expect that from the clock rate difference. If the 9500 PRO ran at 500/500, it would probably outrun GFFX by some 20%, even though it would have similar bandwidth and fillrate. It seems like the NV30 architecture is flawed somewhere, much like the 8500 is not as efficient as the GF4.
I wouldn't mind seeing some 270/270 benchmarks from the GFFX to make this comparison. Maybe the Beyond3D article could include this?
Doomtrooper
Veteran
As for shader lengths being a selling point, I disagree...we have no real DX9 titles out there anyways, and by the time there is all these current cards will be budget cards...and we will have moved on to R400 and NV35.
I used to believe in the feature set of the cards as important, but in the last 3-4 years all I see ever used on any of the current cards is FSAA and AF and or course hardware improvements...
Look at the FX reviews...they didn't even demonstrate ANY DX9 demos, concentrated more on graphs...
Yes its great DX9 cards are here, I personally don't think its that important...different if developers caught up with hardware like the 'old' days.
I used to believe in the feature set of the cards as important, but in the last 3-4 years all I see ever used on any of the current cards is FSAA and AF and or course hardware improvements...
Look at the FX reviews...they didn't even demonstrate ANY DX9 demos, concentrated more on graphs...
Yes its great DX9 cards are here, I personally don't think its that important...different if developers caught up with hardware like the 'old' days.
I keep saying that we've had the time for shaders to begin to be used, and that changing from DX 8 to DX 9 shaders is a much less onerous step. I do think shader features are very important for a card being purchased now. What I also think is that the level of DX 9 support required for the GF FX to theoretically excel compared to the R300 might be what is far off (in games).
At the same time, I think the R300 is still competitive in this regard in any case, so even realization of this theoretical advantage doesn't automatically hand the GF FX a victory.
At the same time, I think the R300 is still competitive in this regard in any case, so even realization of this theoretical advantage doesn't automatically hand the GF FX a victory.
Doomtrooper
Veteran
Yes we are saying the same thing, yet yours is much longer, elegant and better spelled..
Similar threads
