GTX 480 and GTX 470 Reviews

Well the card is meant for serious overclocking (also with LN2) so thats minimum.

In addition, the source says that the 16 phase PWM should be able to handle 480 Amps.

 

how much ln2 would you go through in a week

 

The card wouldn't last a week under LN2 conditions. The ASIC would look like a moon crater landscape due to electromigration from overvolting...

Anyway, what a horrific waste of a video card. That much voltage regulation hardware must lead to even more wasted power than in a stock 480. The Radeon 5870 version of the lightning draws like 50W additional power compared to a stock card. I have a hard time imagining this card will do any better. Also, FTL leaking that much heat straight into the case. Urghh...

 

I wonder if you could attach a mini-steam generator and generate some electricity from all the heat that thing would put out.

Regards,
SB

 

Can someone please explain to me what the point of that card is? Is it expected to be cheap or something? Because you'd get higher performance from overclocking a reference HD 5970.

And if you're really dead set on single-GPU, well, Cayman XT would likely do better.

 

Your question was already answered in the OP.

 

Often, yes. Of course you'd need crossfire profiles and you'd need to be prepared to settle for AMD drivers in general.

I'm sure you're aware that isn't available yet, though.

Anyway, I agree that it doesn't look like a very useful card. A lot of recent MSI stuff seems gimmicky like this.

 

You mean that the point is merely to overclock it to hell, final performance be damned?

Sure, Cayman isn't available yet, but it shouldn't take more than 2, 3 months at most. Designing a card such as the GTX 480 Lightning seems like a lot of work. Just a few weeks before it's obsolete, is it really worth it?

Or is it more advertisement than a real product, perhaps?

 

Well, I'm not sure how any DX11 supporting card will be obsolete any time soon. I guess MSI itself wouldn't necessarily agree with that rather trollish statement.

 

It won't be obsolete per se, as it will still be a friggin' fast vidcard, but if AMDs 6800 series is a substantial performance upgrade - particulary from a power/performance perspective - then it will be essentially obsolete as the GF100 GPU pisses away electricity even on a reference design, and with 16 power phases and overvolting out the wazoo it won't exactly get any better.

 

By obsolete I meant "not the biggest-fastest-meanest beast anymore", which I think is what this card aims for. Obviously it should still run all games very well, and for a while.

 

Yeah I guess I just don't subscribe to the notion that power use/efficiency per square mm of die is a terribly important aspect of the overall GPU experience for end users. Somehow that seems to have become the end all of GPU discussion these days.

To me that totally pales in comparison to solid driver support (I have a 5850 and a Fermi, and with Nvidia application specific fixes tend to already be in the driver before the app is released, and once they are in, they actually stay in, and performance won't suddenly nosedive again 2 driver releases on), practical and user editable game profiles, image quality, and 3D support that is actually available and affordable. Not to mention, top performance without having to dink around with CrossFire.

 

That should be obvious given that it was launched at the finals of the Master Overclocking Arena. All participants use the same hardware so the only variable is how much you can overclock said hardware. i.e the fun and challenge is in the overclocking. A stock 5970 may score higher but it obviously won't be as engaging as overclocking some ridiculously souped up 480 on LN2 @ crazy voltages.

 

Actually the N480GTX Lightning wasn't announced, it was merely previewed. the release date is somewhere in October.

http://www.techsweden.org/nyhet/mS4V-msi-n480gtx-lightning/

 

I know, but developing an entirely new card just so you can get some publicity for an overclocking competition doesn't seem worth it unless you can sell it afterwards.

If the card is scheduled for release in October as Neliz says, how long is it going to stay relevant? A month? How much money can you make from that?

 

I think you're missing the point of the card. They don't care about profit or relevance or volume. It's just marketing.

 

It does matter when your home heats up ~3-4C just from turning your PC on.

Drivers and stability and all that matters too, but I can't say I ever had much in the way of problems with those with ATI, not even in the old Rage128 Pro days. That card rolled over and absolutely DIED when RtCW launched, but the game ran fine. It was just so slow as to be unplayable. After that I had a Geforce 3 (MAJOR pants-kick) until Doom3, when I switched to a Radeon 9700Pro, which also served me very well for many months until I got a new PC with a Geforce 6800 vanilla in it, then it was new PC again with twin 8800GTXes. Now I've had Radeon 4890s for a year and a half, they've been great too. Never any problems at all really other than making tons of waste heat. Not that the 8800s really were any better...

 

I'm just surprised they would go through so much trouble, mobilize engineering resources and spend that much just for marketing. But who knows, maybe it's not a lot more expensive than a traditional advertising campaign…

 

while taking up 3 expansion slots... are SLI bridges even long enough to contemplate putting 2 of these (never mind 3) in SLI ?