Fermi GeForce = GTX 4XX - Hardware & Software

[QUOTE=TheStaticAge;21031810]480 beats out the dual GPU 5970s with a single GPU at about $200 less[/QUOTE] no it doesn't [editline]01:19AM[/editline] [QUOTE=acds;21051754][media]http://www.youtube.com/watch?v=g9HBKIH1e-Y[/media] PhysX, 3D Vision and so on. The only really interesting part (for me at least) is the PhysX in Mafia 2, not worth a new case, a new powersupply and 50$ more for the GPU, but I guess that if you already have the right case and right powersupply, it could be worth spending 50$ for the PhysX in certain games (like Mafia 2, most probably Mirror's Edge 2 and so on).[/QUOTE] why exactly would you need a different case? [editline]01:21AM[/editline] [QUOTE=ShaRose_;21049383]Except that with fermi, the lower end you go, the weaker the tessellation is. With ati, it's the same amount all throughout. As well, ati can easily add more performance there if they want. They have both the die area and tdp to spare for it. Just add another engine, or add a bigger / more powerful one. Nvidia doesn't have that ability, since fermi is already maxed out on tdp, die area, etc, and the limit is 32 polymorph engines for the architecture. They could overclock, but then, tdp limits. They can't go past 300 watts and still be PCI-E. [editline]01:44AM[/editline] Oh and of course, you can crossfire low end cards and double performance with tessellation afaik, because it adds another tessellation engine to the mix. Don't know if it really works though, so.[/QUOTE] so, let me make sure i'm getting this correctly, you're saying "ati has room to grow" and "nvidia can't" and your reasoning is simply just "nvidia can't"? that's an interesting argument

[QUOTE=acds;21051754][media]http://www.youtube.com/watch?v=g9HBKIH1e-Y[/media] PhysX, 3D Vision and so on. The only really interesting part (for me at least) is the PhysX in Mafia 2, not worth a new case, a new powersupply and 50$ more for the GPU, but I guess that if you already have the right case and right powersupply, it could be worth spending 50$ for the PhysX in certain games (like Mafia 2, most probably Mirror's Edge 2 and so on).[/QUOTE] A 9800GT would be fine for a graphic card which can do only PhysX (with a Radeon 5970) ?

[QUOTE=M_B;21051776]so, let me make sure i'm getting this correctly, you're saying "ati has room to grow" and "nvidia can't" and your reasoning is simply just "nvidia can't"? that's an interesting argument[/QUOTE] Well more or less "nvidia can't" is accurate. Given the current 40nm manufacturing process, and how much Nvidia has been struggling with their massive die and low yields, Nvidia just can't improve without a serious architectural re-design and a completely new approach towards GPU design. They can't make it bigger without sending the yields even further down, they can't make it draw more power without staying with in official PCI-E spec, and the result is that they simply cannot physically add more CUDA cores or whatever/raise the clock speeds. We'll see performance improve as Nvidia gets a better grip on the 40nm process, but by the time that happens 32nm/28nm (Depending on which one comes out faster from GlobalFoundies/TSMC) is probably going to be heading towards the mainstream, and AMD will be preparing their newer generation of GPUs. And at that point, if Nvidia wants to stay competitive, they'll be getting their newer generation ready too; even if the GTX 480 has god-like tessellation and compute capabilities, and it's been future proofed (to an extent), we frankly can't expect it to remain competitive against an entirely new, redesigned generation of graphic cards being manufactured on a smaller process.

[QUOTE=ADT;21052423]A 9800GT would be fine for a graphic card which can do only PhysX (with a Radeon 5970) ?[/QUOTE] You can't use a Nvidia GPU as a PhysX card anymore with an ATI card.

[QUOTE=Shogoll;21052573]Well more or less "nvidia can't" is accurate. Given the current 40nm manufacturing process, and how much Nvidia has been struggling with their massive die and low yields, Nvidia just can't improve without a serious architectural re-design and a completely new approach towards GPU design. They can't make it bigger without sending the yields even further down, they can't make it draw more power without staying with in official PCI-E spec, and the result is that they simply cannot physically add more CUDA cores or whatever/raise the clock speeds. We'll see performance improve as Nvidia gets a better grip on the 40nm process, but by the time that happens 32nm/28nm (Depending on which one comes out faster from GlobalFoundies/TSMC) is probably going to be heading towards the mainstream, and AMD will be preparing their newer generation of GPUs. And at that point, if Nvidia wants to stay competitive, they'll be getting their newer generation ready too; even if the GTX 480 has god-like tessellation and compute capabilities, and it's been future proofed (to an extent), we frankly can't expect it to remain competitive against an entirely new, redesigned generation of graphic cards being manufactured on a smaller process.[/QUOTE] Well, there's that, but I was hinting towards that while nvidia tried to make fermi scalable, it's limit is at 16 clusters. So they'd need to redesign it anyways. Also, TSMC killed 32nm. Oh, and charlie is self congratulating himself on fermi. Enjoy. [url]http://www.semiaccurate.com/2010/03/29/why-nvidia-hacked-gtx480/[/url] My favorite part is the end: As things stand, the 9,000 risk wafers seem to have produced less than 10,000 GTX480s and about twice that many GTX470s if the rumored release numbers are to be believed. That would put yields of the new lower spec GTX480 at well under the two percent Nvidia saw last fall. [editline]11:29AM[/editline] [QUOTE=Robber;21053682]You can't use a Nvidia GPU as a PhysX card anymore with an ATI card.[/QUOTE] Well, you can if you hack the drivers. Or did nvidia fix that?

[QUOTE=Robber;21053682]You can't use a Nvidia GPU as a PhysX card anymore with an ATI card.[/QUOTE] Yes, you can. But with something else, a mod: [media]http://www.youtube.com/watch?v=Fgp1mYRYLS0[/media]

[QUOTE=ADT;21055603]Yes, you can. But with something else, a mod: [media]http://www.youtube.com/watch?v=Fgp1mYRYLS0[/media][/QUOTE] Wow, cheers for that. Just put my 8600gts in and now I have physx :smug:

[QUOTE=ShaRose_;21054733]Well, there's that, but I was hinting towards that while nvidia tried to make fermi scalable, it's limit is at 16 clusters. So they'd need to redesign it anyways. Also, TSMC killed 32nm. Oh, and charlie is self congratulating himself on fermi. Enjoy. [url]http://www.semiaccurate.com/2010/03/29/why-nvidia-hacked-gtx480/[/url] My favorite part is the end: As things stand, the 9,000 risk wafers seem to have produced less than 10,000 GTX480s and about twice that many GTX470s if the rumored release numbers are to be believed. That would put yields of the new lower spec GTX480 at well under the two percent Nvidia saw last fall. [editline]11:29AM[/editline] Well, you can if you hack the drivers. Or did nvidia fix that?[/QUOTE] It's sickening how ridiculously ATI biased you are.

[QUOTE=M_B;21051776] why exactly would you need a different case? [/QUOTE] Well considering the temps, it's a good idea to get a decent case (not a no-name brand case with one puny fan in the front). It's not required, but a good airflow is a good idea with these cards (though I doubt anyone buying a 500€ card doesn't have a good case already).

[QUOTE=acds;21060909]Well considering the temps, it's a good idea to get a decent case (not a no-name brand case with one puny fan in the front). It's not required, but a good airflow is a good idea with these cards (though I doubt anyone buying a 500€ card doesn't have a good case already).[/QUOTE] The cards are made to run at that temperature. You don't need a new case.

[QUOTE=ADT;21055603]Yes, you can. But with something else, a mod: [media]http://www.youtube.com/watch?v=Fgp1mYRYLS0[/media][/QUOTE] I thought Windows 7 just enabled this without any modding. But I guess I am wrong.

[QUOTE=Odellus;21060674]It's sickening how ridiculously ATI biased you are.[/QUOTE] How was that ati biased? It's the truth. Point out where it was. Fermi really does have a hard limit of 16 shader clusters, evergreen really does use one (small) dedicated tessellation engine. Nvidia really did send out 9000 risk wafers, and if rumored allotments for fermi's launch is correct, that really is 1.something yield. TSMC really did kill of all plans for 32nm, and in fact that fucked over ati more than anything since it was designed to be a 32nm chip. Oh, and of course charlie is indeed congratulating himself for being right in the face of people like you. Perhaps if you want to scream bias, you should point some out? I could see it if I said "Fermi really does have a hard limit of 16 shader clusters BECUSE IT IS THE SUXXORS and evergreen really does use one (small) dedicated tessellation engine BECAUSE IT OOWNS FERMI ANYWATYS BECAUSE IT OWNS SO HARD", but I didn't. I suppose you could try and say it's bias if nvidia's been lying and fucking around for the past 6 months and I believed someone who actually did some research who also happens to be as biased as hell, instead of listening blindly to PR which of course is biased as hell. But hey, he was right. [editline]07:29PM[/editline] [QUOTE=Odellus;21060975]The cards are made to run at that temperature. You don't need a new case.[/QUOTE] They run at that temperature with good airflow, and it doesn't say how long lifespan for the cards are.

[QUOTE=Robbazking;21061593]A fuck i should had kept a 8800gt just for this.[/QUOTE] lol i kept mine.

[QUOTE=Odellus;21060674]It's sickening how ridiculously ATI biased you are.[/QUOTE] aaah now I remember who that is

[QUOTE=reapaninja;21064352]aaah now I remember who that is[/QUOTE] hi

[QUOTE=Odellus;21065747]hi[/QUOTE] not you, the ATi tard

[QUOTE=reapaninja;21066340]not you, the ATi tard[/QUOTE] Nice to see you two can read. I've yet to see anything from either of you pointing out where that post (or it's followup) were biased. [editline]11:50PM[/editline] Or better yet you could try and actually TALK about fermi instead of sniping me.

[QUOTE=ShaRose_;21068977]Nice to see you two can read. I've yet to see anything from either of you pointing out where that post (or it's followup) were biased. [editline]11:50PM[/editline] Or better yet you could try and actually TALK about fermi instead of sniping me.[/QUOTE] Well you're kinda jerking off to 1.7% like an idiot and a whole lot of other things that are relatively minor. [editline]10:50PM[/editline] You know [editline]10:50PM[/editline] Like a fanboy

[QUOTE=TheStaticAge;21031810]nvidia is supposed to run hotter, [b]they have their cards a lot quieter than AMD[/b], but they run a little warmer[/QUOTE] Yeah I agree totally. I can barely the fans of the 470 and 480 over the sound of the 5 Yamaha engine generators needed to power it.

[QUOTE=Dark-Energy;21069716]Yeah I agree totally. I can barely the fans of the 470 and 480 over the sound of the 5 Yamaha engine generators needed to power it.[/QUOTE] The above point applies to you too. Wait weren't you the guy who in this exact thread backpedaled out of the most idiotic and biased statement ever with "lol I trolld u"

[QUOTE=FunnyGamer;21069623]Well you're kinda jerking off to 1.7% like an idiot and a whole lot of other things that are relatively minor. [editline]10:50PM[/editline] You know [editline]10:50PM[/editline] Like a fanboy[/QUOTE] 1.7 percent? Where. Nividia send off 9000 risk wafers, and had to cut down losses from 512 shaders to 480 shaders because of yield, and even then they pull under 2 percent, because as far as we know it's only 10,000 chips for the launch. [editline]12:32AM[/editline] you know [editline]12:32AM[/editline] like math instead of charlie bashing

[QUOTE=ShaRose_;21069846]1.7 percent? Where. Nividia send off 9000 risk wafers, and had to cut down losses from 512 shaders to 480 shaders because of yield, and even then they pull under 2 percent, because as far as we know it's only 10,000 chips for the launch. [editline]12:32AM[/editline] you know [editline]12:32AM[/editline] like math instead of charlie bashing[/QUOTE] Gee I wonder what's under 2%. Certainly can't be 1.7% that thing's way up there with 50%. [editline]11:05PM[/editline] I think you're the one who needs to do the math here. Charlie pointed out numerous times the 1.7% yield, and you just did. I think you just need to get out.

[QUOTE=FunnyGamer;21069868]Gee I wonder what's under 2%. Certainly can't be 1.7% that thing's way up there with 50%. [editline]11:05PM[/editline] I think you're the one who needs to do the math here. Charlie pointed out numerous times the 1.7% yield, and you just did. I think you just need to get out.[/QUOTE] I think you should stop being so pigheaded. The 1.7 you are referring to was in september, I'm referring to nvidia's A3 risk wafers, of which they ordered 9000. The ones being used NOW. Which, if current launch quantities are correct, is under 2 percent again! It's almost as if you refuse to read and instead are too [B]BIASED[/B] to understand what I'm saying.

[QUOTE=FunnyGamer;21069840]The above point applies to you too.[/quote] lol. It's funny how you respond to everything negative that is said about fermi, you scream "BIASED" and "FANBOY" at just about everything you don't like to hear. [quote]Wait weren't you the guy who in this exact thread backpedaled out of the most idiotic and biased statement ever with "lol I trolld u"[/QUOTE] "most idiotic and biased statement ever" is your entitled opinion. If you read through what I said you'd realize I am probably right. And I didn't backpedal out my statement, did I? No. I just simply started trolling afterwards cause I thought it was funny and had nothing else to do. It's funny because you will probably find one little statement that appears, at first, that I backpedaled. Infact, I don't feel that I backpedalled at all. I had absolutely no intentions of doing that cause I was right. I bet you're going to find something that you misinterpreted as back pedalling, and then going off on a huge tangent about it. Just watch.

[QUOTE=ShaRose_;21063357]Oh, and of course charlie is indeed congratulating himself for being right in the face of people like you. -snip- But hey, he was right.[/QUOTE] Contrary to popular belief, it is indeed possible to be both right and an idiot. [editline]10:15PM[/editline] [QUOTE=Dark-Energy;21070079]Infact, I don't feel that I backpedalled at all. I had absolutely no intentions of doing that cause I was right.[/QUOTE] See above.

[QUOTE=Dark-Energy;21070079]lol. It's funny how you respond to everything negative that is said about fermi, you scream "BIASED" and "FANBOY" at just about everything you don't like to hear.[/QUOTE] That's standard modus operandi for fanboys.

[QUOTE=Roast Beast;21070090]Contrary to popular belief, it is indeed possible to be both right and an idiot. [editline]10:15PM[/editline] See above.[/QUOTE] Whatever. Call it what you want then.

[QUOTE=Roast Beast;21070090]Contrary to popular belief, it is indeed possible to be both right and an idiot.[/QUOTE] Right for such a long time says 'source'. Either that or nvidia had the EXTREME bad luck to keep fucking up until charlie was right. [editline]12:50AM[/editline] Oh, and TSMC boned ati again. Now they are coming out with southern islands with 40nm while waiting for TSMC to get it's shit together. [url]http://www.semiaccurate.com/2010/03/30/atis-next-generation-outed/[/url]

[QUOTE=ShaRose_;21070137]Right for such a long time says 'source'. Either that or nvidia had the EXTREME bad luck to keep fucking up until charlie was right.[/QUOTE] Perhaps. Still doesn't mean he's not an idiot.

[QUOTE=Roast Beast;21070207]Perhaps. Still doesn't mean he's not an idiot.[/QUOTE] Still can't admit he has a source, after all of this. Pretty sad. Also, XFX basically said that fermi was shit as well, and won't carry it. [url]http://www.legitreviews.com/news/7707/[/url]