Check out NV40 on HL2

[jsadie]

http://www.xbitlabs.com/articles/video/display/nv40_29.html

Beta code but the results are out of this world
I would say that this is quite an improvement over 5950 + 9800 XT

 

[Evildeus]

Well, i'm not sure (in fact i'm sure it isn't ) it's a good think to benchmark GC on beta demos of something that should be there. But yeah it's a great improvement something like 2-5 times faster in those particular settings.

I would prefer see the next big game release however :?

 

[hstewarth]

Just think will most likely HL2 not take advantage of PS 3.x shaders.. Just image what if performed and look like if it did.

 

[jsadie]

Well, i'm not sure (in fact i'm sure it isn't ) it's a good think to benchmark GC on beta demos of something that should be there. But yeah it's a great improvement something like 2-5 times faster in those particular settings.

I would prefer see the next big game release however :?

Yeah , i know.But it is extremely promising + i'm sure a big relief to Valve that the new NV cards are doing so much better than before

 

[Stryyder]

http://www.xbitlabs.com/articles/video/display/nv40_29.html

Beta code but the results are out of this world
I would say that this is quite an improvement over 5950 + 9800 XT

Not sure if anyone would actually rely on a benchmark or video perfomed on an unapproved stolen beta version. Thats just me though

 

[breez]

People seem to be really suprised by NV40, but it has double the pipes compared to 9800XT, more memory bandwidth and it's a new generation so I think this is just expected

 

[jsadie]

Not sure if anyone would actually rely on a benchmark or video perfomed on an unapproved stolen beta version. Thats just me though[/quote]

Dont look at the numbers.Look at the trend - Much better results than before

 

[jvd]

it says pure speed. How do we know this isn't running the mix mode persicion or even the dx 8.1 path for the nv40 ???

 

[Stryyder]

Not sure if anyone would actually rely on a benchmark or video perfomed on an unapproved stolen beta version. Thats just me though

Dont look at the numbers.Look at the trend - Much better results than before[/quote]

What that the 6800 performs better than the NV3x series on a hacked demo that hasn't been updated in a year and has missing textures and models.

Thats a big suprise and totally useless information

 

[jsadie]

it says pure speed. How do we know this isn't running the mix mode persicion or even the dx 8.1 path for the nv40 ???

We don't know

 

[jvd]

it says pure speed. How do we know this isn't running the mix mode persicion or even the dx 8.1 path for the nv40 ???

We don't know

so basicly just useless fud that proves nothing .

 

[Bjorn]

it says pure speed. How do we know this isn't running the mix mode persicion or even the dx 8.1 path for the nv40 ???

From what we've seen, that wouldn't be that much of an improvement for the NV40 though.

 

[Evildeus]

it says pure speed. How do we know this isn't running the mix mode persicion or even the dx 8.1 path for the nv40 ???

Do you see any DX 8.1 game that gives 5 times more frame rates? I think it answers easily your wondering

 

[Bjorn]

so basicly just useless fud that proves nothing .

You could of course argue that the publically released benchmarks also are fud at this point in time, for the same reasons.

 

[jvd]

it says pure speed. How do we know this isn't running the mix mode persicion or even the dx 8.1 path for the nv40 ???

Do you see any DX 8.1 game that gives 5 times more frame rates? I think it answers easily your wondering

I'm sure that the dx 8.1 shaders are much less complex in the game than the p.s 2.0 shaders used in the game. For a 16x 1 card clocked at 400mhz it should be able to run it much much faster than a 500mhz 4x2 card .

 

[Kaotik]

so basicly just useless fud that proves nothing .

You could of course argue that the publically released benchmarks also are fud at this point in time, for the same reasons.

I think his point was that as it's ran on stolen beta (or rather even alpha?), the performance figures may be way off the real course on all cards, compared to the final code.

 

[Bjorn]

I think his point was that as it's ran on stolen beta (or rather even alpha?), the performance figures may be way off the real course on all cards, compared to the final code.

My point was that the benchmarks that was released wasn't running on final code either.

 

[Kaotik]

I think his point was that as it's ran on stolen beta (or rather even alpha?), the performance figures may be way off the real course on all cards, compared to the final code.

My point was that the benchmarks that was released wasn't running on final code either.

I think I should start reading posts more carefully

 

[Stryyder]

so basicly just useless fud that proves nothing .

You could of course argue that the publically released benchmarks also are fud at this point in time, for the same reasons.

I think his point was that as it's ran on stolen beta (or rather even alpha?), the performance figures may be way off the real course on all cards, compared to the final code.

That was my general point

 

[onetwo]

It doesn't matter whether it's 'stolen' or not: The fact of the matter is NV40 ran the same game that was run on the other cards, and it ran it 2x-5x faster.

Helps that the game is an actual part of HL2.