The Fermi Thread - Part 3

**Olivon** · 03-23-2010, 02:39 PM

And what about HD 4850 with 120° on Furmak ? Too hot ?

**Sam_oslo** · 03-23-2010, 02:41 PM

Originally Posted by Sn0wm@n

DRIVER BUG??? and all this time i though nvidia didnt have driver bug .....

so nvidia did cheat in the past and they made it look subtle ???? so why not now .....

omg another driver bug idea ..... so lets blame the thing on a bug ..... instead of pr stunt

I don't know much about these drivers, but is is natural with some limitations so early. Specially for these cards that are based on a totally new architecture.

Are you dismissing the driver problems in such a early stage?

**Sn0wm@n** · 03-23-2010, 02:47 PM

LOL @ nvidia fanboy blaming driver bug

**blob** · 03-23-2010, 02:48 PM

Originally Posted by Sn0wm@n

DRIVER BUG??? and all this time i though nvidia didnt have driver bug .....

so nvidia did cheat in the past and they made it look subtle ???? so why not now .....

omg another driver bug idea ..... so lets blame the thing on a bug ..... instead of pr stunt

It's not a driver bug. It's the Dirt2 Demo lacking the correct profiles for the GTX 4X0 cards (The demo was released 6 months ago)
The full game does seem to have the correct profiles.
Not mentioning that the benchmark is GTX470@DX9 vs HD5870@DX11 is a differend topic

**Iconyu** · 03-23-2010, 02:50 PM

Originally Posted by Sn0wm@n

DRIVER BUG??? and all this time i though nvidia didnt have driver bug .....

so nvidia did cheat in the past and they made it look subtle ???? so why not now .....

Don't be so patronising, everyone has bugs.

As for cheating the rules have changed, in DX10 they added rasterisation rules. If a card is rendering something it has to be a certain way. So now if it isn't done the right way, it's normally because something is wrong not to get a speed boost.

Since DX10 you've not seen as many attacks on image quality except for the bs concerning the DX 10.1 of a certain game, nvidia smashed that like the hammer of Thor. But yeah as far as Microsoft are concerned they don't want to see any crap from ATI, nvidia or whoever.

So nope it's unlikely for nvidia to cheat, at least directly.

**SimBy** · 03-23-2010, 02:50 PM

3 more days and all this Fermi nonsense hopefully ends.

**Sam_oslo** · 03-23-2010, 02:51 PM

Originally Posted by Sn0wm@n

LOL @ nvidia fanboy blaming driver bug

Is it all you got to say?

It is a good idea to keep the person focus out of a technical discussion. Does this means that you dismiss driver problem in such a early stage?

**GoldenTiger** · 03-23-2010, 02:51 PM

Originally Posted by Chickenfeed

Its all in the name of PR. Anyone who is planning on buying a GTX4xx card without waiting for reviews won't have their minds changed by something such as this. As for the rest of the sane world, they will wait to take in the more level headed and honest reviews which will be out in force in a matter of days. Marketers embellishing the truth? Oh really?!

I'll be waiting for reviews, but it's definitely a 470 or 480 for me: it just depends on performance and 100% certain pricing on newegg that will decide it.

**trinibwoy** · 03-23-2010, 02:54 PM

Originally Posted by GoldenTiger

Unless something super-reliable surfaces I don't really care about speculation with only 2.5 days to go, frankly.

You know what they say. Where there's smoke....

**Sn0wm@n** · 03-23-2010, 03:01 PM

Originally Posted by Sam_oslo

Is it all you got to say?

It is a good idea to keep the person focus out of a technical discussion. Does this means that you dismiss driver problem in such a early stage?

so i have to explain it in a technical manner as to why i think its not a driver problem but more of a cheat that nvidia pulled to make fermi look like its a good piece of trash ????

come on ... all the sudo news + all the leaked benchy pointed to it months ago ... and now this dx9 stunt .....

**GoldenTiger** · 03-23-2010, 03:03 PM

Originally Posted by trinibwoy

You know what they say. Where there's smoke....

Sure, but there's been smoke spread out a lot all over, so it's hard to pinpoint where the actual fire is

.

Some claim a 480 is barely on-par with a 5870. Some claim a 470 is 10% faster than the 5870.

**Sam_oslo** · 03-23-2010, 03:05 PM

Originally Posted by Sn0wm@n

so i have to explain it in a technical manner as to why i think its not a driver problem but more of a cheat that nvidia pulled to make fermi look like its a good piece of trash ????

come on ... all the sudo news + all the leaked benchy pointed to it months ago ... and now this dx9 stunt .....

I asked you a simple question, you don't need to talk about all other stuff. Unless you want to avoid to answer, of course.

Let me ask again. Do you dismiss the possibility that driver problems may be affecting the performance in such a early stage?

**trinibwoy** · 03-23-2010, 03:06 PM

Originally Posted by GoldenTiger

Sure, but there's been smoke spread out a lot all over, so it's hard to pinpoint where the actual fire is

.

Some claim a 480 is barely on-par with a 5870. Some claim a 470 is 10% faster than the 5870.

True. It was a fun ride but it's almost over. Wonder when that Hecatoncheires speculation thread will get started....

**GoldenTiger** · 03-23-2010, 03:08 PM

Originally Posted by trinibwoy

True. It was a fun ride but it's almost over. Wonder when that Hecatoncheires speculation thread will get started....

(Goes off to make thread)

**bill_d** · 03-23-2010, 03:10 PM

Originally Posted by GoldenTiger

(Goes off to make thread)

remember to put... part 1

**ElSel10** · 03-23-2010, 03:13 PM

Originally Posted by Olivon

And what about HD 4850 with 120° on Furmak ? Too hot ?

It's OK because it's ATI, not Nvidia. It's also OK that eventually, the VRMs on that card will pop when running Furmark. Again, because it isn't nvidia.

**thatdude90210** · 03-23-2010, 03:15 PM

Originally Posted by trinibwoy

You know what they say. Where there's smoke....

"Where there's smoke, there's a..." Fermi at idle?

**Sn0wm@n** · 03-23-2010, 03:18 PM

Originally Posted by Sam_oslo

I asked you a simple question, you don't need to talk about all other stuff. Unless you want to avoid to answer, of course.

Let me ask again. Do you dismiss the possibility that driver problems may be affecting the performance in such a early stage?

again..... driver problem are dismised... wanna know why???? they had the fricking card for so long in their hand tweaked it that they had to tweak the driver package .... but hey your right its all speculation until release day ...

and why run a benchmark in dx9 and say we are having driver problem instead or do some pr spin on it ???? the reason is nvidia didnt want it to be known.... now the question is why ???

heres why i made the assumption.....

**Musho** · 03-23-2010, 03:22 PM

Originally Posted by ElSel10

It's OK because it's ATI, not Nvidia. It's also OK that eventually, the VRMs on that card will pop when running Furmark. Again, because it isn't nvidia.

The VRMs don't pop. The card simply crashes and goes into a protection mode, because furmark (or was it OCCT?) was causing the OCP to kick in. Still a design flaw, but it only affected benchers and people running stresstests. Wasn't causing any problems for real games. Still a serious and bad problem though

I haven't heard of any cards hitting 120C though

**DavidGraham** · 03-23-2010, 03:28 PM

You guys are ignoring the fact that Charlie has been wrong about clocks , Shader count , Power Consumption and nearly every aspect of GTX 4xx series except for the launch dates !

**ElSel10** · 03-23-2010, 03:30 PM

Originally Posted by DavidGraham

You guys are ignoring the fact that Charlie has been wrong about clocks , Shader count , Power Consumption and nearly every aspect of GTX 4xx series except for the launch dates !

Shhh!

**annihilat0r** · 03-23-2010, 03:30 PM

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

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

**Sam_oslo** · 03-23-2010, 03:31 PM

Originally Posted by Sn0wm@n

again..... driver problem are dismised... wanna know why???? they had the fricking card for so long in their hand tweaked it that they had to tweak the driver package .... but hey your right its all speculation until release day ...

and why run a benchmark in dx9 and say we are having driver problem instead or do some pr spin on it ???? the reason is nvidia didnt want it to be known.... now the question is why ???

heres why i made the assumption.....

It's much better now. As long as you don't forget you are amusing, guessing and speculating, then it is OK .

Let me ask you a new question.
When you know your own ideas are based on speculations and assumption, how can you be so sure about your own impressions that you dismiss others ideas so firmly (even an obvious idea as possible driver problems in early stage), by calling them fanboys and such ?

**Olivon** · 03-23-2010, 03:35 PM

Originally Posted by Musho

The VRMs don't pop. The card simply crashes and goes into a protection mode, because furmark (or was it OCCT?) was causing the OCP to kick in. Still a design flaw, but it only affected benchers and people running stresstests. Wasn't causing any problems for real games. Still a serious and bad problem though

I haven't heard of any cards hitting 120C though

I have seen the case with a HD 4850 Sapphire classic 2 slots cooler on a bad ventilated mid-ATX case with Furmark. Really massive heat

the card is designed to withstand 120°C operating temperatures - W1zzard

http://forums.techpowerup.com/showpo...0&postcount=11