NV30- the fan will last how long, we must dust it how often?

Just like to comment on this.

The XPC's generally have two expansion slots, so if you can manage to get everything you want integrated, there's no problem.

 

For htpc/pvr setups, those two slots are used up by 1: Better Video Card (AGP Slot) 2: TV Tuner/Capture Card (PCI Slot). That is if video card does not have tv-capture or tuning capabilities. The better video card is optional, but the tv-tuner/capture card is always an add-on and installed.

 

How is a massive, "dustbuster" bolted to the side of videocard a "non-relevent" aspect to consumer hardware?

The only thing that could allow someone to overlook such a monstrocity, or better yet, downplay or discount it's existence would be the infamous rose-tinted glasses.

I was never aware that all the guys over at MadOnion were thermal engineers. By your account, one obviously must be a thermal engineer in order to know that slapping a large heatsink/fan onto a processor improves heat dissipation, which in turn allows clock speed to be increased. I'll look towards all those folks with >400mhz 9700 Pros at MadOnion with all new respect knowing that each and every one of them is a professional thermal engineer that has conducted months of air flow and heat dissipation tests in order to achieve their results.

The problem starts when that non-standard exhaust system takes up the passenger seat.

We don't even know if an NV30 will be released that cures cancer, repairs the ozone layer or restores lost forrestry from over logging either. This is the luxury of creating hypotheticals in the absence of facts. The only facts or information we do have is the dual-slot monstrocity that was presented recently from which real, live photos of the unit have been published on many websites.

I only see one person building hypotheticals in the absence of facts here. And being "gifted" a free 9700 Pro doesnt really add much value by stipulating yourself as an owner. Someone who either downplays Fords or champion's Chevys doesn't add any tangible value if the Ford they drive was given to them. This doesn't figure the perception of actually choosing a product for use based on it's merits, other than the merit of it being handed out for free.

I agree the thread has taken a course a bit on the extreme side, as MANY motherboards share PCI slot 1 with the AGP slot as far as resources, so this even further narrows the problem of the slot being taken up. The real subject of this thread (and the subject champions for a particular IHV would fight tooth and nail to try and cloud with FUD) is the size and scale of what has happened to this NV30, and to speculate this was an "original" design decision is totally unfounded.

The source of criticism is simply that many of the touted benefits of going to a .13 micron design have been perceived to go "out the window"- as size and cooling on this .13 micron product from the release photos illustrate no such advantages. The new entry of "effective" bandwidth has also been invented as a tool to try and market this product as well, especially given the ~200mhz clock speed advantage this unit already has over 9700 Pro. A whole new piece of FUD has been invented just to "fit" the marketing of this card, as well as a standard to bolt cooling that takes up a secondary slot (and quite possibly crowds and reduces airflow to neighboring 3rd slot) as well and it's damn reasonable for some folks to comment on all of this.

Many of the concerns here are valid, legitimate concerns given the lack of substantiated information. Those of space (will it fit in XYZ case, and will it have proper clearance with other PCI cards), the longevity of the "dustbuster" plastic cooling system strapped on the side of this beast, and of course how loud such an extreme cooling measure will be. And no, the same concerns would exist regardless of what IHV brand name is stamped on the reference PCB. To state these as "irrelevent" or cloud these concerns in FUD is only symptomatic of one thing..

 

HAHAHAHA! That was a good one!

 

Isn't it rather ironic that it was DemoCoder that was targeted at?

 

Well, DC--*chuckle* since you are obviously not a "thermal engineer" yourself, I wonder if you perchance belong to the Guild of Thermal Engineers....from whence your self-appointed authority surely comes?...

Let me just state for the record that unlike you I do not believe the nv30 Dustbuster requires a "Thermal Engineer" to decipher. I think its mysteries are actually very mundane.

Oh, of course--none of these things play the slightest role in this particular heatsink-fan combination....*chuckle* It's all therefore "nonsense"....You might wish to pause a moment and reflect on the fact that some people who have had what is generally termed "experience" with heatsinks and fans for a number of years might make observations that to the inexperienced mind seem like nonsense.... It also might be worth considering that such people might actually have a legitimate basis, therefore, for making statements that remain nonsensical and mysterious to you. Just a thought...

Before I get into a bit of well-deserved sarcasm here, I would like to suggest to you that just because a device "takes into account" operating conditions like dust, etc., does not mean that device then becomes impervious to those conditions. In effect, air conditioners and heat pumps in homes, for instance, "take into account" things like "dust," etc. This is WHY they have filters and also explains WHY those filters MUST be cleaned regularly--on the advice of the manufacturer. In fact, if you fail to clean said filters, etc., as a part of routine maintenance in many cases it is considered negligence and your manufacturer's warrantly is terminated.

So, think real hard about a "filter" and what it's presence would mean if such a filter is in fact present (that's what we were talking about.) If I have to spell *this* out for you I would drop my application for Thermal Engineer--I don't think you'll make it...

Secondly, I know that you prefer to think that nVidia's mass of computing power has been hard at work for the last couple of years calculating all of that advanced and very difficult data concerning the precise placement of the nv30 Dustbuster fan--and that this task was one of herculean proportions that surely has had to have been in motion since the first day work on nv30 began. I understand that this is your interpretation. However, I think that's the biggest load of crap I've heard in a long time!...

Anyhoo...


[sarcasm on]

Of course--to suggest that any of these companies ever consider that they are competing with each other is absurd, isn't it? Why, the very notion that nVidia would think it would have to compete with another company making 3D graphics chips is patently ridiculous!

And of course all of these companies--even though we really are talking about one company and one product--all of these companies exist in a rarified, competition-free environment where competition with other companies just doesn't exist, and where none of these companies ever do anything that might be construed as competitive, right?

And certainly nVidia, simply because it is several months behind ATI in shipping any DX9-Ogl2.0 hardware, months behind ATI in getting these products to market, and simply because investors wonder what's going on while nVidia reports losses and is late with a new product line, late for the third or fourth time relative to this one specific product line--none of that would ever give nVidia the slightest reason to act as though it was in anyway "competing" (perish the thought) with ATI for dominance in the 3D chip marketplace.

Therefore, since nVidia doesn't consider itself in competition with ATI, it had no intention of ever caring whether or not its products would be seen as competitive (what "nonsense"), and mounting a gigantic, Dustbuster fan on the nv30 card was simply something planned from the start which has no relation whatever with ATI's R300 parts--despite the incredibly positive press ATI has been getting for months about this GPU. Despite all of that, nVidia could care less whether its nv30 is considered a performance competitor with ATI's shipping DX9 GPUs. But the fan...ah, the fan...now *that* was something of supreme importance to nVidia from day one!

Oh, Certainly.....how clear it all is now....

[sarcasm off]


Look DC, if you really think nVidia designers were so lame as to deliberately design something like this fan into their premium nv30 product, from the beginning, without a care in the world as to whether ATI's R300s ran circles around it.....well, I think you should post haste apply for the position of Thermal Engineer somewhere.... In other words, you don't seem to have a clue in this matter, and you appear to be doing absolutely nothing except "defending" the use of a huge fan when you admittedly have no idea of the background leading to the decision to market such a product.

What some of us have been discussing here are hypotheticals--that is, reasonable, experienced, educated GUESSES as to what something like this is doing in the reference design for a consumer market 3D card.

You ought to realize that the lifespan of the nv30 will be very short in nVidia's history just as will the R300 be in ATI's. However, if nVidia's nv30 isn't perceived as having at least some performance advantages over the ATI R300-9700 Pro reference design, nVidia will have a very tough row to hoe. To even think that nVidia is deaf and blind to the competition from ATI the R300 products present is...utter nonsense. nVidia is not a company piloted by utter fools... They know exactly what they have to do with nv30 to *beat* ATI's R300 9700 Pros in the marketplace, and of course that is to beat them in performance.

OBVIOUSLY....clocking the nv30 to 500MHz is what nVidia figures it needs, and this apparently cannot be done WITHOUT the dutbuster fan onboard--REGARDLESS of how much nVIdia would prefer not to have to include such in the nv30 reference design (for lots of good and evident reasons even an aspiring Thermal Engineer should be able to comprehend)....

I don't know how much simpler it can be, really. The proof is in the pudding and the FAN--the BEHEMOTH FAN--is the pudding. It rarely gets any simpler than this.

 

Excepting this whole idiotic discussion is about how that fan is some sort of technical boat anchor that is sure to break unless pampered every couple of weeks, not that the fan exists and is necessary.

Again, there's nothing but the blatant accusations of people who don't know what the f'k they're talking about saying that the thing is poorly designed and will break within a week, leaving the chip melted.

 

Guys no hardware company with the size and reputation of Nvidia is going to put a flagship board on the street that hasn't been tested rigorously for usage. They've obviously put the thing under stress tests.

If it only works for 2 weeks, who do you think is going to take the biggest losses in the deal? The consumer? Lol.

I'm going to eagerly await Sharkfood and WaltC's engineering report in 9 months when their astute observations will be put to the test.

Perhaps its wise though to see the actual thing run in the first place? (just a thought)

 

No, they're accusations poorly disguised as questions. Which is why I called them blatant.

 

Hah, yeah.. I was thinking the same thing.

 

Hehe, I believe WaltC is Checkmate - one of the regulars from the old 3dfx gamers forum (Voodoo Lounge, anyone?). I, and others, had many arguments with him - the same with Sharkfood (another annoying individual) - however, in this instance I have to agree with them.

Yes, very true, couldn't agree with that more. nVidia obviously feels they NEED such a fan to gain a performance advantage over ATi, or why bother? Wouldn't they settle for a conventional fan if that could achieve the desired result?

Well, from this consumer's perspective, a huge card with a monstrous fan is not only overkill, it reminds me of the voodoo5 5500 and its large size compared to nVidia's products of the same period. 3dfx were slammed for their huge card with 2 chips - and mocked even more for the ridiculous v5 6000 (4 chips) - so it would be very hypocritical to say that size is irrelevant now, when it's nVidia's turn to produce a gigantic specimen.

Bigger is not always "better".



edit: typos...

 

I've not followed this discussion for some time, but if people are questioning when this heat-sink solution came about I think its reasonably obvious from that it was fairly late on.

NV's reference board shots sent out just prior to the launch clearly have markings for where they envisaged the heatsinks of the board to be, hence where the board should be clear of tall components. It seems they had planned for small RAM-sinks and what looks to be two designs of core HSF on that board, neither of which match the cooling solution they've announced it with.

Now, I don't doubt that we'll probably see a lower end version using cooling solutions that match the outlines on that reference board, but its also fairly clear that the solution they have shown it with is not the initially designed solution for the boards, and probably not what they had planned for the initial chip speed/heat projections. I'd say they realised they could reach 500MHz with aditional cooling once they got the chips tested and this was a reasonably late decision.

 

demo,

while i do agree that pc enthusiasts tend to go for the larger space towers, consider the following (taken from http://www.cgshaders.org/)

a quick look up at the prizes section (at http://www.cgshaders.org/contest/contest-results-02.php):

 

I've never commented on the physical robust-ness of the fan (actually never occurred to me, really.) I think a discussion of why it's there in the first place is far more interesting.

 

Well, I'm just hoping that the yields are high enough such that nVidia sells a number of lower-cost chips that don't require the external air supply.

 

Okay, this topic will not fade away, so I might as well join the fun...

In the big picture: This year has shown a somewhat disturbing trend in power dissipation from the high-end GPU's and CPU's.

The all-new P4 3.06 HT have typical thermal power rating of 81 W while the Athlon XP 2800+ stays at more 'common' 62 W. I don't know what the Radeon 9700 Pro's power dissipation is - but is damn hot, that's for sure.

And now enter the NV30. The point of the special fan might very well be that nVidia cannot otherwise be sure that people have sufficient air flow in their case to keep the GPU and ram cool with a standard fan.

Whether their cooling solution is good or not is a bit difficult to judge, but I'm inclined to point out that the air intake seems somewhat on the puny side and thus they have use a very high-speed fan. It'll be loud, like it or not.

Anyway, I have taken the consequence of all this and are now going the water-cooling way (since I'm want a very low noise system). Wish me luck... 8)

 

8) LoL...I could admit to being Checkmate--I could, yes--but then I'd have to nuke your home town!... Hmmmm...."annoying"...yes, has a nice ring...


...

Exactly. Seems really obvious, doesn't it?

...

Yes, and 3dfx was also mocked for its attitude about AGP texturing, and for the extra onboard power connector for the V5. The V5 6K would have been an 8-pixel-per-clock 3D card two years ahead of the 9700 Pro. In the beginning they were also mocked about their implementation of FSAA (with some calling it either "blurry" or a gimmick.) Now with 128-mb 3D cards common and external power connectors being deployed to facilitate stability, and FSAA capability a fact of life, I might even say in some technical matters 3dfx was visionary. It's just too bad they couldn't have carried that innate understanding about 3D technology over into their business endeavors. Ah, well...

But I digress...the single most disturbing element of this giant fan bonanza that is nv30--to me--is the fact that all signs point towards the nv30 GPU being possibly overvolted but definitely overclocked to hit that 500MHz target. That essentially, nv30's performance claim to fame is dependent on a giant fan to facilitate factory overclocking to 500MHz (if this turns out to be practical on a sufficient scale, I presume). I'd have had more respect for nVidia if they had clocked at 400MHz-425MHz and used more normal cooling--and left the overclocking to the OEMs and after-market hobbyists.

...

Sigh...I wouldn't know...lost mine in the war, you know...

 

Yep, good observation on those reference board pics. It would seem as if they'd gotten enough initial samples of the chip back to determine that they weren't going to clock to target speeds as their original models might've indicated, and so in order to hit those speeds a drastic (and expensive) thermal solution was required. I think the probability is excellent that had ATI never produced the 9700 Pro that the nv30 we'd have seen would have been clocked much lower and marketed without the giant fan. (IMO, of course.)

 

Good luck! I've toyed with the idea of water cooling, but to tell you the truth it's all just a bit...wet...for me