Slashdot Mirror
Nvidia Talks About Next-Gen Geforce, Plus Pics
Per Hansson writes "Techspot was at Comdex in Sweden a few days ago; we have now posted a small interview with Nvidia along with some high-res pictures of the Geforce FX on this page in our new comments system." This is one of the strangest looking video cards I've ever seen (and it isn't cheap), though it may look different by the time you can buy it in a box. Which is not yet, despite all the hype.
This is "market specific?" What market? Ill tell you this, They best not think people will go for a 2 slot card for "heat management". I do agree with the passive heat sinks on the reverse though, very good idea!
Obviously this is a first crack at the FX. I'd bet serious money that within six months of its release, a version will be ready that requires only one slot. Consumers hate incoveniences like this -- what if a cap on the motherboard gets in the way of one slot? Moreover, those who wait six months are more likely to be price-conscious consumers -- which means their systems are less likely to have gobs of space open (cheaper mobos = fewer slots).
Still, I want one. Now.
Make cheese not war 8:)
Honestly, there's gotta be a smarter way to cool that thing than a huge, ugly, entire-slot-using heatpipe. Either that, or developing a new way to crunch graphics numbers other than using a single chip..... SLI on one card using two slightly slower chips? Power consumption would go up, but you could use floppy power connectors in lieu of a new bus solution that provides more voltage grunt, and it'd be easier to cool.
Why don't manufacturers start doing dual-DVI outputs? Granted, most LCD's have a second analog input, but what's the point of having one DVI output then?
I wish they'd start putting dual-DVI outputs on them. Maybe one of the other companies that makes them (MSI, PNY, Leadtek, etc) will offer one finally. AFAIK they don't even offer a hydrahead adapter for the one DVI port to split to two (doubt its possible without a proprietary output like the Radeon VE's).
THE MAGIC WORDS ARE SQUEAMISH OSSIFRAGE
Unfortunately the amount of time I spend at work doesn't allow me to play many games, but seeing this article made me curious: is the demand for faster video cards still high enough to warrant so much development?
My reasoning comes from the simple fact that the human eye is unable to perceive extra frames beyond a certain number (I've heard this number to be around 25 or 30). I realize that owning a video card that averages at, say, 40 isn't ideal, because when you get into a lot of action you'll drop frames. But the statistics I see on the newer cards show that even the most complex games, on the highest resolutions, get WAY too many frames. The gamer is never going to notice dropped frames. And yet every few months a new card comes out, and lots of people get excited and shell out the bucks (enough to warrant the R&D I suppose).
I'm honestly not being cynical. Can someone explain why an average gamer (or even an extreme gamer) should buy the newest and best?
I've just put together a new computer an Nvidia nforce2 ASUS A7N8X motherboard, and you know how many PCI slots I use? None. My video uses the AGP, but then sound is on board (and it's good), usb/firewire/serial-ATA RAID/regular ATA, etc are all on board PLUS two NICs. Sure, I could add SCSI (but how many home users do?), or a TV tuner (already built in to my video card), or a variety of other things, but I really have no need for these PCI slots. I'm surviving quite well without them.
If I zoom in will I be able to see the blood pumping through it's wings?
Or a bunch of flowers.
Is each individual pollen grain to be described?
Will the water eventually splash?
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
The Inquirer has an article [theinquirer.net] that takes a look at the GeForceFX. Hopefully things won't turn out as they did for 3DFX.
Disclaimer: I have no idea about the economic status of Nvidia. But I do see them in pretty much every computer advertized, and they've generally delivered very successful products since the first Geforce chip, so I assume they got a strong finacial position. And if you can't solve it even if you got more money to throw after it than the rest, well maybe you deserve being dethroned. That's what competition is all about, isn't it?
Kjella
Live today, because you never know what tomorrow brings
What I want is for the hardware to support a realistic and comprehensive physics model in said Final Fantasy universe.
"A language that doesn't affect the way you think about programming, is not worth knowing" - Alan Perlis
This sounds just like what caused the death-knell of 3DFX: company bets the bank to make a monster video card that blows everything out of the water, and holds off on a whole scheduled version release (once every 18 months) to make this monster card... and blows it big time.
3DFX used to compete with NVIDIA. When NVIDIA released a new line of cards, so did 3DFX, or when 3DFX released a new line of cards first, so did NVIDIA.
When the GeForce2 cards came out, everyone waited for 3DFX to release their competitive line. About 4 months later, 3DFX released a couple Voodoo4 cards, but not much in the way of competition, and nothing spectacularly advanced above the Voodoo3's. However, they also let out news of plans to make a market breaker card, the Voodoo 5-6000, which would take up fall case length (and bump harddrives), have 5 fans on it, and require an external wallwart-style DC adaptor for power supply. It was a $600 card meant for the mega-gamers and graphic designers out there. This was a huge card... and their biggest flop, for once it came out, NVIDIA was already releasing the GeForce3's which had better specs and lower prices overall.
Now, Nvidia does something just like that. This card is double-height (the second slot worth is ducting for external air intake and exhaust) and is full case length. It's got monster specs, and has thrown off their regular 18-month cycle of new cards. This new one is $600 as well.
Sounds to me like some of the execs of 3DFX have gotten on the board of NVIDIA via the buyout, and are trying to make another Voodoo5-6000. I hope it doesn't end the same way, with this company going down the tubes as well.
Building in a cooling solution like that which is totally unrepairable by the end user is a great way to build in forced-obsolescence.
I think I'll stick with my radeon. If the fan quits, I'll just replenish the oil.
Kudos to Nvidia, though, for finding a way to force their users to buy new cards in the future! This'll certainly be the wave of the future, like fibreglass bodies on cars!
If you could be told what you can see or read, then it follows that you could be told what to say or think - BoC
it's called silicon real-estate.
it's also called packaging cost.
it's called data routing on the board (FR4 is very, very slow unless you use a LOT of traces, which is very, very diffcult).
I think it may also be called lower MTBF.
and how about "debugging is a pain?"
either way, though - don't expect "multi-processing" on but the most high-end incarnations - when they have squeezed out of every bit of performance per-chip.
My life in the land of the rising sun.
While all of the modern 3D chipsets have impressive frame rates for running 3D games they tend to suck badly for much of anything else.
The chips are very slow to switch from text to graphics and vice versa.
I had a board with a slightly older Nvidia chip set. I wasn't very satisfied with the stability of the Xfree drivers for it so I tried the Nvidia Linux drivers. Their driver took five minutes to switch between text and graphics modes.
Older chipsets were much more practical for day to day use; the super speed models remind me of trying to drive a AA fuel dragster to the office every day.
Actually, this goes out to all the video card manufacturers..
Why hasn't anyone put the GPU on the OPPOSITE side of the card yet? Every AGP card I see, the GPU is ALWAYS facing towards the PCI slots in the system where it.
A. Blocks out other PCI cards
B. The fan causes noise and instability if it is running too close
C. It exhaust the heat onto those other cards.
Instead of trying to put the carridge before the horse, why not just mount the GPU on the opposite side? There's no PCI slots to get in the way, and you could fit a HUGE cooling solution there.
Hey Nvidia if you want to hire someone with more common sense design tips like this i'm availiable. I'll slap your engineers with a cluestick for ya.
Once again, we're seeing a very elaborate (and ungainly) design for the next ultra high power graphics chip. As technology progresses, we're accustomed (as it should be) to the new products to be the same physical dimension (or smaller) as the older products. This is how we have such concepts as the PDA, the microATX form factor, just to name a few.
The graphics card arena has been a major exception to this for the last few years. It's one of the few industries that I can think of where the product is actually GROWING in size and becoming more combersome as the technology becomes increasingly faster and more complex. I believe this is a sign that, not unlike how we discovered in the Pentium II/III era, that card/based processor packages are poor product design that are a) larger than necessary b)gum up the works, and c) only enhance the problem of cooling, thus needing continuingly more complex cooling systems.
The current AGP(or PCI or whatever) bus expansion card methodology for video cards can be seen as going through the same problem, especially in the case of the GeForceFX. We've seen these problems previously in the designs for the GeForce3,4, made much fun of them in the case of the 3dfx Voodoo5 6000 cards, and even the latest ATI cards are requiring more power than the AGP bus can provide. Doesn't this show that there is an inherent flaw in the packaging design for this technology?
GPUs need to take the same road that CPUs have taken (and now restored since we now use socket based motherboard solutions again) and be sold soley as the graphics processor, with the memory substructures and soforth built onto the motherboard. This increases the efficiency and ease that the GPU can communicate with the central bus and the rest of the system. In addition, you will no longer need to build an elaborate cooling strucutre to make up for the lack of ventilation provided by the typical AGP/PCI card slot design.
Nvidia is part way there with the NForce already, building the graphics subsystem as a central part of the motherboard chipset and PC bus, but the flaw here remains (as in most integrated motherboard systems) that you are stuck with the technology. Of course, you can upgrade an NForce system with a full GeForce4 FX or Radeon if that is your choosing, but that just brings back the card problem. What needs to be done is to create a NForce type chipset with an FCPGA type socket for the GPU as well as the CPU, that way both systems are imminently upgradable (not to mention the potential benefits in creating a more efficient in-line cooling solution for the interior of the system) and thus our size problems begin to be alleviated.
-Julius X
remove "-whatkindofspamdoyoutakemefor-" from email to send
I thought I'd take a
picture and make a rebuttal to your statement. Gotta love digital.
In this pic there are 5 mobo's.
Intel 850GB
Some asus socket370 thing
Some soyo socket370 thing
Iwill BD100 slot1
Some intel socket370 thing
You will notice on the asus board I put a tape measure across as a reference.
Now out of the 5 boards sampled, only 1 has no space for heatsinks on the right
side. Also to note this board is a slot1, which is no longer in production.
On the other hand, every single semi modern board in this picture has more
than adequate room for heatsinks on the right side.
So unless these newer cards are going into an outdated system, putting the
fans/heatsinks on the right side shouldn't be a problem right? Simple enough
solution without having to resort to heat pipes/water cooling or peizo electric
cooling.
These shots were done with this card
autopr0n is like, down and stuff.