Wich card is the king of the hill (nv40 or R420)

[martrox]

Even Dave seems to defend ATI alot I think even though Nvidia is attacked more......you can ban me now 8)

There's only one way to read this, Trinibwoy..... And, if you spent enough time here, you would never have posted this......

Many nVidia fans seem to feel that IF you say anything against nVidia, you're biased against nVidia. Not so..... and, right now and for the past 2 years, ATI has had the better available product - there are many here that swore by nVidia (that's by..not at! ) that now own ATI..... Just why is that? Are they now biased against nVidia because they didn't buy nVidia?

 

[Evildeus]

Of course they are biased How could they choose better performance, better drivers, better quality, in 2 words, better hardware

 

[trinibwoy]

Even Dave seems to defend ATI alot I think even though Nvidia is attacked more......you can ban me now 8)

There's only one way to read this, Trinibwoy..... And, if you spent enough time here, you would never have posted this......

Why not? Is it taboo around here? It is quite evident that he comments more often on ATI products and he seems to have a better relationship with ATI than Nvidia.

Many nVidia fans seem to feel that IF you say anything against nVidia, you're biased against nVidia.

Same for ATI fans so what's your point?

Not so..... and, right now and for the past 2 years, ATI has had the better available product - there are many here that swore by nVidia (that's by..not at! ) that now own ATI..... Just why is that? Are they now biased against nVidia because they didn't buy nVidia?

It's amazing how unbiased this forum is when nearly every discussion devolves into IHV branding and the requisite flaming thereafter. No secret that ATI has the better product and that's why I own one.

 

[trinibwoy]

Of course they are biased How could they choose better performance, better drivers, better quality, in 2 words, better hardware

I just looked at my initial post and I'm amazed at how innocent it was before one particular phrase was yanked out and jumped on......hilarious

In addition to Nvidia's past infractions I think the NV40's power requirements are the strongest negative affecting the polls. Then you have to factor in voter bias which I think is skewed on this particular forum and also the fact that in graphics wars it's usually compelling to pick a winner. And in this case if you were to pick a winner you would have to go with ATI given the preliminary tests and info that we have in hand at the moment.

have to factor in voter bias which I think is skewed on this particular forum

pfft find me an unbiased forum.

 

[Dave Baumann]

It is quite evident that he comments more often on ATI products and he seems to have a better relationship with ATI than Nvidia.

Thats not evident actually. Its about contacts.

 

[martrox]

Of course they are biased How could they choose better performance, better drivers, better quality, in 2 words, better hardware

OMG! I......I.......I'm speechless!

 

[DemoCoder]

Defend - or providing information that I know about? There's a difference.

How come you don't provide any positive information 'bout NVidia then?

Why does info provided when you jump into threads always seem to be positive for ATI but negative for NV (e.g. bad yields this, HDR that, no DX9 runtime speedup, etc) You defended ATI's performance in one benchmark based on driver issues and assert future performance gains, but seem somewhat nonenthusiastic about the concept of there being any big improvements in NV's driver performance) You said nothing about the performance of ATI's HDR, but then seemed oddly critical of it when NV added blending and filtering.

Hell, if you provided the same service for NV, I wouldn't even need to post my info. (and I don't believe I'm "more connected" than you are)

 

[DeanoC]

NV40 is a much better card for us at the moment, VS 3.0 (and geometry instances) and FP16 blending make all the difference for a developer working with next-gen level graphics. Maybe if you playing games then R420 based cards are better but for developers NV40 are best at the moment no doubt.

As always different horses for different courses.

I don't think that statement should be a surprise to anyone here.

Just out of curiosity, what would you have said if you were to go back in time and make the same statement, say, this time last year?

Last year ATI R3x0 were the best, decent fast PS 2.0, MRT support, floating point buffers that worked.

 

[Bjorn]

You said nothing about the performance of ATI's HDR, but then seemed oddly critical of it when NV added blending and filtering.

Does anyone know if Half Life 2 is going to have any HDR rendering ? (i think someone here mentioned that it wouldn't because of performance issues, supposedly on the R300 then).

 

[Heathen]

Why does info provided when you jump into threads always seem to be positive for ATI but negative for NV (e.g. bad yields this, HDR that, no DX9 runtime speedup, etc)

Could it be that info is actaully a) Of interest to people & b) true. Just because you can't take it doesn't mean it shouldn't be posted. I'd rather have the truth rather than some molly coddled view of the world where everyone's nice.

and assert future performance gains,

Well the DEC compiler guys are supposed to be some of the best in the business after all.

nonenthusiastic about the concept of there being any big improvements in NV's driver performance

Maybe because of past practice and the fact that nvidia's compiler guys are still somewhat of an unknown quality.

You said nothing about the performance of ATI's HDR, but then seemed oddly critical of it when NV added blending and filtering.

So, maybe by adding blending and filtering the speed just plain sucks. ATi's performance without the blending and Filtering may actually be quite good.

At the end of the day Demo, what you posted sounded pretty much like a whinge pure and simple. If Dave deosn't have the info he can't post it, accept it.

 

[Bjorn]

I'd rather have the truth rather than some molly coddled view of the world where everyone's nice.

And you're sure that the truth really is the truth ?

So, maybe by adding blending and filtering the speed just plain sucks. ATi's performance without the blending and Filtering may actually be quite good.

The funny thing is that HDR rendering was supposed to be a big advantage for the R300 vs the NV3X and there was some talk here about that fact. And afaik, HDR rendering using FP16 blending is supposed to be of higher quality and faster (if i'm not mistaken) then the method Valve was using for the Half Life 2 presentations back then. And much easier to implement also.

 

[Dave Baumann]

Why does info provided when you jump into threads always seem to be positive for ATI but negative for NV

Or how you perceive it.

(e.g. bad yields this, HDR that, no DX9 runtime speedup, etc)

Are the reports of IBM's customer yields being poor false? In which case, why are IBM suggesting it? Are there historical issues with NVIDIA on the 130nm process? If not, why have NVIDIA said it? Are you expecting wide scale improvements for < SM3.0 titles purely from DX9.0c?

You defended ATI's performance in one benchmark based on driver issues and assert future performance gains

I don't recall defending ATI's performance - I do recall pointing out a benchmark that points to an issue that would clearly suggest that there gains to be made with the memory bus because all theoretical measures (bandwidth, fill-rate, geometry rate) are significantly higher than the previous generation and yet the performances in most cases are lower. Given the parity in platform and all these other points I'm looking at what appears to be one of the most logical conclusions - apologies if this approach is a little too level headed for you.

but seem somewhat nonenthusiastic about the concept of there being any big improvements in NV's driver performance)

I raise the question as to whether the memory bus has changed significantly and I also point out that NVIDIA probably has a little better understanding about dual bank memory channel timings because they have more experience with them (more recently with 5700). Are these things not to be considered?

It is the case, though, that I know that the memory bus was one of the major focuses for change between R300 and R420, hence is one of the areas they are most likely to have a some learning to do - I don't know that the same can be said for NVIDIA since I've not heard there have been significant changes to it.

You said nothing about the performance of ATI's HDR, but then seemed oddly critical of it when NV added blending and filtering.

I suggested that the loss of FSAA and fill-rate will probably prevent this from seeing widespread use in this generation, and thats a belief of mine.

Hell, if you provided the same service for NV, I wouldn't even need to post my info.

Hell, if I just got on with a life outside of work, rather than dedicating to large portions of it to this site, then we wouldn't have anything at all. Sorry if the service isn't up to your liking however I only have a finite level of time and my girlfriend only has a finite level of patience I can't spend all day everyday finding out everything. Please, pay me £30,000 PA + benefits and then I'll be happy to make sure there is a level of service that purely satisfies you.

 

[Evildeus]

Of course they are biased How could they choose better performance, better drivers, better quality, in 2 words, better hardware

OMG! I......I.......I'm speechless!

Don't see why (because it's the truth, at least for the past 18-24 moths)

 

[Dave Baumann]

Does anyone know if Half Life 2 is going to have any HDR rendering ? (i think someone here mentioned that it wouldn't because of performance issues, supposedly on the R300 then).

It was demo'ed with it last year, however I've not heard anything since.

 

[Heathen]

And you're sure that the truth really is the truth ?

Any Vorlons round here? 8)

PS: Is it me or are most 3d card 'enthusiasts' also fans of Babylon 5?

 

[Evildeus]

At least, you weren't really quick on the 3dMark/6800 debacle DB.

 

[Bjorn]

I suggested that the loss of FSAA and fill-rate will probably prevent this from seeing widespread use in this generation, and thats a belief of mine.

I thought that the loss of FSAA wasn't necesarily true. Though there's of course always the SSAA option

And loss of fill-rate doesn't seem like a big issue with the R420, NV40X generation. The question should rather be, will the games still be playable at acceptable resolutions using those features ?

 

[Dave Baumann]

No, it is the case MSAA doesn't work with NV40's FP blending. The fill-rate may not be much too much of an issue, but there are 2X memory footprint and bandwidth costs as well. Shader rates may alleviate the bandwidth somewhat, but I'd wager a halving the bandwidth would still be felt, especially as it seems that hardware shader performance is outstripping software uptake. I just think that telling people now "give up half your performance for this" isn't going to happen for a generation or so, on a widescale basis.

 

[trinibwoy]

Why does info provided when you jump into threads always seem to be positive for ATI but negative for NV (e.g. bad yields this, HDR that, no DX9 runtime speedup, etc)

Could it be that info is ......

Well the DEC compiler guys are supposed......

Maybe because of past practice.....

So, maybe by adding blending and filtering.....

I just thought the lack of conviction in this particular post was amusing....with all the maybes

 

[martrox]

Any Vorlons round here? 8)

PS: Is it me or are most 3d card 'enthusiasts' also fans of Babylon 5?

"I really hate it when you answer like that......"

"Good......."