Profit of NV40 parts

10817474486qLMOmeutS_1_15_l.jpg



Just remember that someone designed this sucker to have two of those and a speaker, there had to be a good reason for both of those being added.

P.S. one of the reasons I like Kyle's reviews especially the Motherboard ones is that he takes alot of really good pictures so you know what he is referring to in the text of the review.

P.S> taken from http://www.beyond3d.com/forum/viewtopic.php?p=276905#276905
 
I am willing to bet that indeed there are situations in certain games where the nv40 will demand more power. In such cases if the user only has 1 molex plugged in the card will crash. Then the end user will call support and ask why is my card doing it and tech support will tell them to plug in the other molex. Then the end user guy that did not upgrade his power supply will be stuck. The tech support will tell them they can wash their hands of it because the end user did not go with the min recommended specs. So for the most part it is not an absolute necessity but when the gpu is stressed... it probably does need the extra power.

Regarding the market share gains, it is likely directly attributed to NV dumping GF4MX and FX5200 cards on the market simply because they can. Even though enthusiast are familiar with them varieties most end users still have no clue what they are buying for graphics.

I am not buying their suggestion that they can produce the NV40 cheaper with higher yields. I just don't buy it. I guess if it is the case then they ought to have the NV40 widely available soon enough. Wait and see on that one.
 
Can't wait to see NVDA's explanation why what they're saying in this analyst talk is completely the opposite of [H]'s findings..
 
Florin said:
Can't wait to see NVDA's explanation why what they're saying in this analyst talk is completely the opposite of [H]'s findings..
maybe the final shipping boards will not require 2 molex connectors for stock clocks? If you look at the maximum specified power draw (110W) it should be IMHO no problem to get that out of 1 molex (+AGP).
 
Florin said:
Can't wait to see NVDA's explanation why what they're saying in this analyst talk is completely the opposite of [H]'s findings..
Oooooooh! Wouldn't it be something if nVidia just said [H]ard was wrong.... :devilish:
 
digitalwanderer said:
Florin said:
Can't wait to see NVDA's explanation why what they're saying in this analyst talk is completely the opposite of [H]'s findings..
Oooooooh! Wouldn't it be something if nVidia just said [H]ard was wrong.... :devilish:

[H]ard wasn't the only reviewer that did that test but they are the only one I can remember, I got to go through those 6880 previews again damn.
 
Sabastian said:
I am not buying their suggestion that they can produce the NV40 cheaper with higher yields. I just don't buy it. I guess if it is the case then they ought to have the NV40 widely available soon enough. Wait and see on that one.

That's not necessarily true though since it depends on when they actually started to produce the revision that's going out to customers. A better idea is to see how long it takes between first availability and mass availability.
 
Stryyder said:
[H]ard wasn't the only reviewer that did that test but they are the only one I can remember, I got to go through those 6880 previews again damn.
I e-mailed Kyle about it, he says they didn't go past the warning screen because they didn't want to fry out their only test board. (Can't say I blame them a bit)

He's doing some checking with some folks who have 6800s and I pointed this thread out to him.
 
digitalwanderer said:
Stryyder said:
[H]ard wasn't the only reviewer that did that test but they are the only one I can remember, I got to go through those 6880 previews again damn.
I e-mailed Kyle about it, he says they didn't go past the warning screen because they didn't want to fry out their only test board. (Can't say I blame them a bit)

He's doing some checking with some folks who have 6800s and I pointed this thread out to him.

Good work Dig
 
both Molex plugs are coming off of one wire from the PSU.

Remember this is what stryder said, that means that one molex would be able to provide the power w/o melting... b/c it is the same wire so it might as well be one plug. :) But perhaps the actual board does not allow the power to be transferred to run both the gpu and memory from one molex connector or something.
 
Maybe another way of looking at the dual molex statement is that the GeForce 6800 Ultra is an "overclocked" product at 400Mhz.

The NV40 was probably only meant to run at 350Mhz (or whatever the single connector card is running at), but changed when competitive rumours surfaced.
 
Sxotty said:
both Molex plugs are coming off of one wire from the PSU.

Remember this is what stryder said, that means that one molex would be able to provide the power w/o melting... b/c it is the same wire so it might as well be one plug. :) But perhaps the actual board does not allow the power to be transferred to run both the gpu and memory from one molex connector or something.

Personally I think that in most rigs with a decent power supply one molex = no problem.

Medium laden rig with two molex on the same rail no problem

Heavily laden rig with one molex or two on the same rale in FarCry cranked = problem.

They gave themselves a safety factor because when they crunched the numbers they theoretically had a voltage gap somewhere under certain conditions. what this means is that in most cases everything would be fine with one connector but a statistically high enough possibility existed in which there would be a problem hence the creation of the safety factor through the extra molex and the recommendation of a higher end PS.

You have a rig with a P4EE, two burners and a couple of hard drives in a raid array there will probably be a power issue somewhere.
 
For some crazy reason I just want to try taking a splitter molex and running one feed from the power supply to the splitter that is connected to both of 'em.

I know feeding of the same power supply line is pretty much the same thing, I just got this weird hunch that's bugging me.....
 
digitalwanderer said:
For some crazy reason I just want to try taking a splitter molex and running one feed from the power supply to the splitter that is connected to both of 'em.

I know feeding of the same power supply line is pretty much the same thing, I just got this weird hunch that's bugging me.....

I think the mad scientist is coming out in all of us but I certainly wouldn't want to do any of the things I am thinking about to MY $500 board.

LS
 
digitalwanderer said:
For some crazy reason I just want to try taking a splitter molex and running one feed from the power supply to the splitter that is connected to both of 'em.

I know feeding of the same power supply line is pretty much the same thing, I just got this weird hunch that's bugging me.....

We ran our benchmarks off one lead (set of wires coming) off the PSU that had two free Molex plugs. Never did we get any warnings, or see the clocks vary. Also, the PSU held up fine (430w), but it was not powering a bunch of extras either like you might in a real system.
 
BFGTech just tested a 6800U quickly for you guys. Here is what they said.

"We just ran 3DMark03 with one power connector with no problem. We are
letting it run for a few hours to see if anything happens.

The warning said the clock freqs would drop with only one power cable
hooked up, but checking CoolBits, they were the same.

Fascinating."
 
FrgMstr said:
BFGTech just tested a 6800U quickly for you guys. Here is what they said.

"We just ran 3DMark03 with one power connector with no problem. We are
letting it run for a few hours to see if anything happens.

The warning said the clock freqs would drop with only one power cable
hooked up, but checking CoolBits, they were the same.

Fascinating."
<neo>

Woah.

</neo>

(Two "woah"s in one day so far, it must be Friday...)

Thanks Kyle, great info and your posting it up here is very much appreciated. :)
 
Interesting feedback from BFGTech. It would seem that both plugs are very much needed for proper sustained operation of the 6800U.

"With one connector, we are getting some artifacting in 3DMark03 during
Battle of Proxycon and Trolls, but not much, some shimmering and black
lines. Played UT2004 no problem with one connector. "
 
FrgMstr said:
Interesting feedback from BFGTech. It would seem that both plugs are very much needed for proper sustained operation of the 6800U.

"With one connector, we are getting some artifacting in 3DMark03 during
Battle of Proxycon and Trolls, but not much, some shimmering and black
lines. Played UT2004 no problem with one connector. "
Fascinating stuff! Again, thanks and any more updates will be gladly appreciated. :) (Turning out to be a funner Friday than I thought! ;) )
 
Back
Top