NVIDIA GF100 & Friends speculation

Ah correct. Had a long time to see how Furmark fairs in framerate terms.

My 8800GT at the office runs at around 40fps at 1920X1080 noAA, so yes, there's probably something fishy here.

If Furmark was a win, it would be in the leaked benchmark graphs :p
 
I'm still surprised Furmark remains in use as anything other than a way to see if either vendor has working precautions against run away power viruses.

Regards,
SB
 
That possibility I finde highly unlikely, but still…

It's exactly what they did for 480 when people complained about the high noise at idle. 580 is the same process node, same high power requirements, same big heatsinks, etc.

What I find unlikely is that Nvidia have suddenly developed some magic that deals with the same problems in a different way. You're not going to see a fixed Fermi with more cores and higher frequency being able to cut it's idle temp in half whilst maintaining noise levels.

Heck, it's the same temp a ATI 4870 idles at because ATI did the same trick of trading off idle temp for low noise.. but that was a couple of generations back when idle power usage and perf/watt was a lot worse than it is now...
 
It's already bigger in some of the pre-selected benches linked here, no 3DMark "11" included in those

Yeah, I saw them (LP2, etc). However, since he said synthetic I was thinking of Futuremark's new benchmark program as the Vantage score wasn't that impressive.
 
Last edited by a moderator:
For what it's worth I'm not seeing anything to support the 128 TMU rumour, reworked chip or not. At 772Mhz that would be a 135% increase in texture throughput. The current 15% perf increase in games is more in line with the increased clock (10%), flops (17%) and bandwidth (8%) so an increase in texturing of that magnitude is very unlikely.

@neliz: I think my $100 is safe!
 
Last news :
MSI test tool : 10 min stress : 86 c @ 67% fan speed (default)
Idle temp :30c @51 MHz core clock.
Easily overclocked to 900MHz .
TDP : 244W
16 SM and 512 Cuda Cores
 
For what it's worth I'm not seeing anything to support the 128 TMU rumour, reworked chip or not.

There's a possible explanation, which would also be sufficient for neliz's previous claims of GF100 having half its 128 diasbled (IIRC).
 
Isnt it weird that nVIDIA is not doing their usual stunts like "Are you ready?" on their site? After all we are at 3 days till launch, be it soft or hard...
 
Isnt it weird that nVIDIA is not doing their usual stunts like "Are you ready?" on their site? After all we are at 3 days till launch, be it soft or hard...
Maybe, just maybe they've burnt themselves one time to many over-hyping a soft launch?

I know, crazy talk..ignore me. :oops:
 
Maybe, just maybe they've burnt themselves one time to many over-hyping a soft launch?

I know, crazy talk..ignore me. :oops:

nVIDIA is really a marketing driven company, so I dont quite believe that they would stop doing it because of that. AND if they did in fact push the date sooner because of the stockholders meeting or what ever it is, wouldnt it make sense for them to make as noise as possible?
 
Back
Top