NVIDIA GF100 & Friends speculation

Edit: Deleted, I copied the wrong link.

 

Exclusive - Sapphire Radeon HD 5870 TOXIC 2,048MB graphics card review

 

Or you could just get high, and then even the official reviews will look semiaccurate.

 

I don't have any at the moment, so no high, just wasted from alcohol tonight

100MHz OC on core, hits 101C in Furmark with fan @ 80%:
http://img.hexus.net/v2/graphics_cards/nvidia/Fermi/N174/OCb.jpg

 

did Hexus not mention fan volume? they ran it at 80% for their overclock... you'ld think they would have noticed

I guess I could take that as implying the noise level wasn't bad, but that just seems unlikely. Tolerable or horrible, that is the question.

Guess we'll find out soon enough

 

<SCNR>
SCHUWEEEEEEEEEEE-
</SCNR>

None of the reviews mentioned fan noise yet. It should be okay in *normal* use considering nApoleon would have bitched about if it wasn't.

 

Isn't 5870/4890 around that mark to even if fan is around 80-100%. That aslong as the ATI driver gimping of Furmark is disabled (throttling system). I sure know my 4890 got upwards 100c for the non-gimped test.

 

Charlie's still suggesting his scores of 480 are accurate on Twitter.

http://twitter.com/CDemerjian

Whatdyya think?

 

When OC'd perhaps, for 4890, but 5870 shouldn't get near those numbers even OC'd for what I know

 

More accurate than semi.

 

If you used the name-changed exe for Furmark to evade the ATI driver throttling then upwards 100c or even more is what happens (with some variation between different GPU sections/VRAM). Search the net for lots of these tests and I can also back that up with 2 different stock 4890 and 4870. Use Everest Ultimate Edition to monitor all graphic card temps.

 

we'll he's been on top of ATI and NVIDIA these past months and he's been right over the majority of the nvidia calls

 

His claims of clocks seem too low to be vindicated, given recent "leaks". And performance is correlated with clocks.

 

Yeah, on HD4 series which is why I said perhaps on HD4890, HD5 has "heftier hardware protection" on the PWM parts
(though I think there's no indication that it would actually hit temps requiring throttling even in Furmark, regardless of what name you use for it)

 

Well his clocks were for 512SP version, were they not? And now we have 480SP version, giving possibility to throw higher clocks

 

remember he said that the clocks were at a certain level and it was shortly after that leaks were showing that they were being purposely lowered.

 

so now we are hearing a full 512sp which was originally very low yield being "prepped' as a full card later on

yeah its crippled

 

That wasn't what he stated, anyways, this isn't a charlie thread there is one dedicated just for him.

 

It is easy to add, just change the number which is a tesselation factor (doable in real time) and you're getting more triangles. Whatever else is happening means nothing here because I was talking about tesselation and nothing else.
So, yes, he should stop posting crap, not me. I can say that again because I'm tired of his replies to my postings. When he says that you need to recompile an application to change the tesselation factor he's posting crap. As simple as that.

 

Please relocate the Charlie analysis to the Charlie thread...