The Fermi Thread - Part 3

[SKYMTL]

"reviewer's guides"It doesnt make any sense like those PR PPTs...

Chiphell's nApoleon has reviewed GTX480.He expected it would have 20% more performance than HD5870 at least,but its performance disappointed him.

Huh? I've found that reviewer's guides usually fly in the face of PR BS and are actually quite accurate.

[GoldenTiger]

http://futuremark.yougamers.com/foru...70&postcount=1

I saw that as well, but since I had no idea who the guy was other than 7500+ posts on that particular forum, I didn't repost it... that would be a nice way for things to turn out however!

[DavidGraham]

looks like he was right about the performance, though...

May be , but there is a high chance he will be wrong about that too , let's just wait and see .

welcome btw, are you from B3D? I remember seeing you on some other forum

Thanks for the welcome , Yes I am from B3D .. been waiting to post here since October 09 , the activation of my account was put on hold , I don't know why , not that I regret it anyway , you guys have done a wonderful job of messing with 3 threads ! disastrous !!

Anyhow , this picture :


It is fake , the site uses conventional names for it's drivers (like Cat 10.1) , not version names ..

Here is the confirmation that it is fake :

This is a fake. Our benchmarks are different.

http://translate.google.com/translat...tcount%3D18157


There is someone , or some group that is trying to mislead the fans with fake , undermined results for Fermi , they are exploiting the strong veils of secrecy invoked by Nvidia .

Anyhow , let's wait and see how it turns out !

[RSC]

(...)

It is fake , the site uses conventional names for it's drivers (like Cat 10.1) , not version names ..

Here is the confirmation that it is fake :
Quote:
This is a fake. Our benchmarks are different.
http://www.forum-3dcenter.org/vbulle...ostcount=18157


There is someone , or some group that is trying to mislead the fans with fake , undermined results for Fermi , they are exploiting the strong veils of secrecy invoked by Nvidia .

Anyhow , let's wait and see how it turns out !

That link is broken...

[MrDiaz]

Your link is down Graham

[NaMcO]

Thanks to a "certain" friend of mine who's under the NDA, i've had a quick look at nVidia FTP with pics, "reviewer's guides", drivers an most improtantly a whole plethora of the GTX470/480 benchmarks against the HD5850 and HD5870 and from what i've seen, the GTX480 beats the HD5870 easily by 25% on average. Sometimes it's 10% sometimes 50-60% but the GTX480 always wins.

My initial previsions were correct if this is verified as it completely shades the 5870 and licks 5970's feet. I sincerely doubt this is the case, but if true i'll sell my 5970 in a blink

[DavidGraham]

sorry back online :

http://www.forum-3dcenter.org/vbulle...ostcount=18157

[mindfury]

OK,let's wait and see how it turns out

I would like to see how those hype become jokes.

[Dezmen]

Omg they got it like 4 days ago.. And still now benchmark? What do they do with gtx480? xD

[derschlambi]

To clear things up, Raffael works for the German Hardware Site PCGH (PC Games Hardware) and stated that these results were fake. The test system they used differs from the system on the screen as well as the drivers and the selection of cards benched. The fps numbers they will show in their test will show only integer values.

Hope i could help. Greetings from Germany

[Krautmaster]

http://www.pcgameshardware.de/aid,70...fikkarte/News/

fake...

[saaya]

Huh? I've found that reviewer's guides usually fly in the face of PR BS and are actually quite accurate.

reviewers guides try to fool refviewers by hilighting good points and ignoring bad points... if you know how to handle that, then thez are accurate, yes... if you think its a good roundup of pro and con of a product, then no, they are very misleading...

It is fake , the site uses conventional names for it's drivers (like Cat 10.1) , not version names ..

that doesnt mean the numbers are fake though... if a site posts results they get into trouble, but if they post a picture of "leaked" results from another site... thats not breaking anz ndas

i think the numbers are fake though, wasnt the 470 supposed to be slightly faster than the 5870 in crysis warhead?

[mindfury]

i think the numbers are fake though, wasnt the 470 supposed to be slightly faster than the 5870 in crysis warhead?

Leaks from asuka/PHK/Chiphell all proved 470 is slower than 5870 in crysis warhead.

[GoldenTiger]

i think the numbers are fake though, wasnt the 470 supposed to be slightly faster than the 5870 in crysis warhead?

They SAID it's fake .

Leaks from asuka/PHK/Chiphell all proved 470 is slower than 5870 in crysis warhead.

Considering chiphell is posting benches that the site they're supposedly from says are fake, I sincerely doubt any of those right now...

[Sushi Warrior]

reviewers guides try to fool refviewers by hilighting good points and ignoring bad points... if you know how to handle that, then thez are accurate, yes... if you think its a good roundup of pro and con of a product, then no, they are very misleading...

And what site do you review for? I'm pretty sure Sky has a bit more insight than you on the topic.

So 2 more days, wonder if we will get any actual, final benchmark leaks tomorrow. Maybe a review posted early?

[Dimitriman]

That is the sexiest looking reference heatsink I've ever seen

[eric66]

http://futuremark.yougamers.com/foru...70&postcount=1

reviewers guide

[takamishanoku]

That is the sexiest looking reference heatsink I've ever seen

Yup. For the first time a Nvidia card looks nicer than an ATI one. The reference cooler is top notch

[RSC]

Yup. For the first time a Nvidia card looks nicer than an ATI one. The reference cooler is top notch

GTX2xx reference coolers were much betther than HD48xx reference coolers...

[SKYMTL]

reviewers guides try to fool refviewers by hilighting good points and ignoring bad points... if you know how to handle that, then thez are accurate, yes... if you think its a good roundup of pro and con of a product, then no, they are very misleading...

I am talking about the benchmark numbers that accompany ATI's and NVIDIA's reviewer's guide. You are confusing the reviewer's guide with the PR deck.

Remember, the PR slides usually try to highlight only the positive points while the reviewer's guide usually acts as a factual document listing specifications, talking about core layout and using pure benchmark numbers (no 100%, etc crap) without PR nonsense.

[eric66]

I am talking about the benchmark numbers that accompany ATI's and NVIDIA's reviewer's guide. You are confusing the reviewer's guide with the PR deck.

Remember, the PR slides usually try to highlight only the positive points while the reviewer's guide usually acts as a factual document listing specifications, talking about core layout and using pure benchmark numbers (no 100%, etc crap) without PR nonsense.

we will see 25 percent lead ? wow nvidia must have some good reasons to hide this monster

[GoldenTiger]

GTX2xx reference coolers were much betther than HD48xx reference coolers...

Yarly.

[Dimitriman]

GTX2xx reference coolers were much betther than HD48xx reference coolers...

I think Fermi Heatsink is shaping out to be n1 best designed reference HS but second place also belongs to Nvidia, me thinks this is n2 hands down:

http://www.3dnews.ru/documents/12278...-front-big.jpg
http://www.3dnews.ru/documents/12278...cooler-big.jpg

[eric66]

ati cards had always fuglier coolers since x1900 series but custom models rock especially iceq and vapor series best coolers imho

[***Deimos***]

Anybody seen this yet?
http://www.tomshardware.com/news/fer...tage,9965.html

ASUS GTX480 50% faster with overvolting!
Overvolting? Seriously? I thought we were working with that safety bear to try to reduce household electrical fires.