More NV40 details emerge

[russo121]

I've seen this at theinquierer

"...It's even more interesting that Nvidia claims that its card will end up four times faster in Doom 3 and seven times faster in Half Life 2..."


http://www.theinquirer.net/?article=14058

[The Baron]

4x in D3 is probably accurate in some situations. 4x2/8x0 versus 6x2/8x2/12x1/16x1. nifty.

7x in HL2, I'll believe it when I see it.

[russo121]

This is only PR stuff..... but if it comes true....... whoooo..... and let's not forget r420.

[silence]

Originally posted by russo121

http://www.theinquirer.net/?article=14058


urm....L'INQ?.....
correct as usual....:cool:

[retsam]

is nvidia pulling out the hallucinogens again :D

[nForceMan]

Originally posted by russo121
I've seen this at theinquierer

"...It's even more interesting that Nvidia claims that its card will end up four times faster in Doom 3 and seven times faster in Half Life 2..."

http://www.theinquirer.net/?article=14058
Old news. click (http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=23489&pagenumber=7)

[russo121]

Originally posted by nForceMan
Old news. click (http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=23489&pagenumber=7)

Soo.... let's talk about nv50.......:p

[digitalwanderer]

Originally posted by russo121
Soo.... let's talk about nv50.......:p
Pfffft! The R500 will toast it. :cool: :p ;)

[nForceMan]

Originally posted by russo121
Soo.... let's talk about nv50.......:p

OK. NV50 will be n times faster than NV40 where n ∈ [2..10]. :naughty:

[russo121]

Originally posted by nForceMan
OK. NV50 will be n times faster than NV40 where n ∈ [2..10]. :naughty:

Translating that to numbers will be:

3dmark03 nv38 = 5000; nv40 3*5000 = 15000; nv50 (minimum) 2*15000 = 30000..... What a Number!!!!!:D

[nobie]

The Half-Life 2 benchmarks that were released showed absolutely terrible performance on a Geforce FX, IMO NV40 had BETTER be 7x faster.

[silence]

Originally posted by nobie
The Half-Life 2 benchmarks that were released showed absolutely terrible performance on a Geforce FX, IMO NV40 had BETTER be 7x faster.


i bet it will run game loading screen 7 times faster....even without optimisations:angel:

or maybe cut scenes?

[Malfunction]

Originally posted by silence
i bet it will run game loading screen 7 times faster....even without optimisations:angel:

or maybe cut scenes?

Thank goodness solid drivers will allow that and backwards compatibility then, sure would be a disappointment if the drivers weren't there to deliver the performance. ;)

Peace,

:afro:

[Ninja Prime]

7 times faster really isn't all that fast if you think about it... Under some situations the R380 is like 3-4 times faster than NV38 in Half-Life 2.

[killermmn]

http://www.samsungusa.com/cgi-bin/nabc/semiconductor/landing_pages/gddr3/gddr3_landing.jsp

check mebay the memory on nv40:D

[-=DVS=-]

Originally posted by Ninja Prime
7 times faster really isn't all that fast if you think about it... Under some situations the R380 is like 3-4 times faster than NV38 in Half-Life 2.

R9800XT = 3/4 times faster then = NV5950
NV40 = 7 times faster then = NV5950
that makes
NV40 = 1.5/2.5 times faster then = R9800XT :rolleyes: :p

Yeah this ram looks sweet
GDDR3 256Mb K4J55323Q 8M x 32 FBGA 1.8/1.8V 500/600/700/800
GDDR2 256Mb K4J55323Q 8M x 32 FBGA 1.8/1.8V 500/600/700/800
:drooling:

[Nutty]

R9800XT = 3/4 times faster then = NV5950
NV40 = 7 times faster then = NV5950
that makes
NV40 = 1.5/2.5 times faster then = R9800XT


If only it were as simple as that.

If we look at doom3..
http://www.hardocp.com/article.html?art=NDc0LDI=

NV40 = 4 times faster then NV5950
NV5900 = 1.22 times faster than 9800Pro.

therefore NV40 = 4 * 1.22 = almost 5 times faster than 9800Pro. :cool:


Lay off the rolley eyes until some true reviews come out. :rolleyes:

[Zeno]

Don't forget that one graphics card being 7x as fast as another doesn't necessarily mean 7x the framerate. The overall speedup will depend on how much the CPU is waiting for the GPU and vice-versa.

[Nv40]

what its more interesting is that the INquirer Comparisons of the Nv40 is againts the Nv35.. so Drivers performance doesnt help in the diferences since both cards most likely used the same. and in the same Nv paths. means that comparisons were apples vs apples.IF the numbers are more or less acurate.. not only the card will be deadly fast ,but also Very expensive.. to get those high numbers.. hehe



Nutty: If only it were as simple as that. If we look at doom3.. http://www.hardocp.com/article.html?art=NDc0LDI= NV40 = 4 times faster then NV5950 NV5900 = 1.22 times faster than 9800Pro. therefore NV40 = 4 * 1.22 = almost 5 times faster than 9800Pro.

you save me a post.. :) in Doom3 in the Fp16 Nv35 was sometimes ~Twice the frame rate of the R350.. in the early Id-nvidia released benchmarks. so 4xtimes faster than Nv35 could mean the Nv40 could be 6x times? if the benchmarks are done again with the same Hardware+drivers.

however the Hl2 numbers againts their last generation hardware Nv38 ,looks a bit to high for an new generation video card . i will be more than impressed. if the Nv40 match Ati R420 performance in Hl2 at full Fp32 precision with all the effects working ,the way valve wants it to be played . :)

[digitalwanderer]

<sigh>

I swear, I was trying like hell not to post in this thread...I really was.

</sigh>




You do realize you're using numbers from a very questionable benchmark to interpolate the performance on a part that doesn't exist yet for a game that hasn't been released?

It's kind of a stretch, ain't it? :|

[Rytr]

This thread is a stretch except the part about the cost and even that will stretch the ole pocketbook...:D

[-=DVS=-]

Originally posted by digitalwanderer
<sigh>

I swear, I was trying like hell not to post in this thread...I really was.

</sigh>




You do realize you're using numbers from a very questionable benchmark to interpolate the performance on a part that doesn't exist yet for a game that hasn't been released?

It's kind of a stretch, ain't it? :|

Mehh everything in rumor mill is questionable guesses no harm from pointelss speculations :rolleyes: :rolleyes: :rolleyes: :rolleyes: :D

[ClyssaN]

This kind of speculation only hurts nvidia imo... I really hope that nv40 becomes a great card but this hype can be bad, but hey, this is the rumor mill and we love rumors :)

[jbirney]

This reminds me of the good ol NV30 slides. Remember when those were show by NV on how much faster it was gonna to be vers the R300? Well we all know how accurate that was and thus gives you an indication how accurate this will be....

The good ol 3dmark2001
http://www.zive.cz/ilustrace/nv30_3Dmark.gif

Q3
http://www.zive.cz/ilustrace/nv30_quakeIII.gif

And I saved the best for last, the PS comparison
http://www.zive.cz/ilustrace/nv30_pixelshader.gif

That PS is really funny now consider how poor the FX shaders are. So keep this in mind before you blindly believe x2, x4, x7 ect :)

[digitalwanderer]

Originally posted by ClyssaN
This kind of speculation only hurts nvidia imo... I really hope that nv40 becomes a great card but this hype can be bad, but hey, this is the rumor mill and we love rumors :)
Oh, I'm not trying to dissuade anyone from speculating as it's one of my fave hobbies too...it's just that I agree with ClyssaN that getting your hopes up about those kind of performance numbers is just setting yourself up for a disapointment.

If it sounds too good to be true, it usually is....fooled once, shame on you; fooled twice, shame on me. ;)