Power and Innovation to Drive High-End GPUs in 2009

[Zillatech]

FYI ~

http://www.techpowerup.com/755...-End_GPUs_in_2009.html

 

[Tempered81]

A "5870 X4" in Q1/Q2 2009. Heheh, hope that turns out to be one rockin' 6 T-Flop Graphics card.

 

[MarcVenice]

Wow. So you reckon q1 2009 will bring us new cards from ATI ?

 

[cyphilis]

Wow... Looks like next year will be a "wait until next month" year. Why buy card "A", when card "B" comes out in a couple of months, and will obliderate card "A".

 

[OCGuy]

LOL @ the comment from someone:


"ATi will have DX11 support in Q1, nV wont have it until Q4, i'm buying ATi!"


Hmm...Windows 7 in Q1? Good luck.

 

[Tempered81]

Originally posted by: MarcVenice
Wow. So you reckon q1 2009 will bring us new cards from ATI ?

Well the graph in the OP's article says, "RV870 GPU (Q1)" and is in the (Q1~Q2) quadrant.

 

[MarcVenice]

Completely missed the graph ( it is rather small ) But yeah, pfff bah, this makes things even more complicated. I was about to pull the trigger on a HD4870, only had to fill in my bankaccount stuff, but I hesitated and cancelled it. I don't think I can really afford to buy a 200 euro videocard when my current videocard isn't exactly stopping me from playing all the new games. Right now, maybe I should wait 3-4 months and see what the 58x0 series bring ...

 

[Zstream]

Interesting, if they pull off the dual MCM I think microstutter will be partially eliminated.

 

[Tempered81]

I think you can buy RV870 & R800 in June/July 2009. So, that's still a good 8-9 months away. I was thinking of grabbing two open box ASUS EAH4850's for $279 shipped @ the egg, and then selling them off around end of spring '09 to get in on the rv870 cards.

 

[Cookie Monster]

Originally posted by: Zstream
Interesting, if they pull off the dual MCM I think microstutter will be partially eliminated.

How so?

Just because the chips are on the same package doesn't mean micro stutter will be gone. It all depends on how these chips communicate with each other. Core 2 quads (MCM, not native quad core) communicated via FSB. So in order for CPUs 0,1 to communicate with Cores 2,3, it goes to the FSB, to the Northbridge then to the FSB to Cores2,3.

I just dont know how they will pull this off. The chip, even at 40nm will be hot (based on temperature readings on the RV770 at 55nm) and to do an MCM approach has me worried about the cooling of this thing.

edit - DX11 already?!

 

[MarcVenice]

Originally posted by: Cookie Monster

Originally posted by: Zstream
Interesting, if they pull off the dual MCM I think microstutter will be partially eliminated.

How so?

Just because the chips are on the same package doesn't mean micro stutter will be gone. It all depends on how these chips communicate with each other. Core 2 quads (MCM, not native quad core) communicated via FSB. So in order for CPUs 0,1 to communicate with Cores 2,3, it goes to the FSB, to the Northbridge then to the FSB to Cores2,3.

I just dont know how they will pull this off. The chip, even at 40nm will be hot (based on temperature readings on the RV770 at 55nm) and to do an MCM approach has me worried about the cooling of this thing.

edit - DX11 already?!

What do you base this on? Why will those chips run 'hot'? All chips run hot, depending on how good you cool them. Heat output should go down or remain equal to that of a HD4870X2, they are jumping from 55nm to 40nm after all? It's going to be interesting to see how they cool it, but it should be possible, they also have a new cooling technique, something with vapor chambers, in the pipeline.

 

[Rhonda the Sly]

Hmm...Windows 7 in Q1? Good luck.

DirectX 11 is supposed to ship independently of Windows 7 (so it has Vista support) and be backwards compatible with DirectX 10/10.1 hardware.

I find it interesting how in the future we plan to run our graphic cards on "Power and Innovation" and governments on "Hope and Change." The next few years should be interesting as these fascinating new concepts play out.

 

[rjc]

Originally posted by: MarcVenice
What do you base this on? Why will those chips run 'hot'? All chips run hot, depending on how good you cool them. Heat output should go down or remain equal to that of a HD4870X2, they are jumping from 55nm to 40nm after all? It's going to be interesting to see how they cool it, but it should be possible, they also have a new cooling technique, something with vapor chambers, in the pipeline.

According to a vr-zone the first ati chip shrunk to 40nm hasnt shown much reduction in power consumption.

So for a RV770 shrink 256mm2 @ 55nm -> ~ 140mm2 @ 40nm. If the worst case happens and there is no or little power savings that will be around say 140W meaning the chip is emitting heat at 1W/mm2 which is quite huge.

By comparison depending on how you measure it the surface of the sun emits energy at ~20W/mm2 .

Good luck if your gpu fan breaks on that baby and you don't notice.

 

[videogames101]

Originally posted by: jaredpace
A "5870 X4" in Q1/Q2 2009. Heheh, hope that turns out to be one rockin' 6 T-Flop Graphics card.

Hell yeah!

 

[Concillian]

Originally posted by: rjc

Good luck if your gpu fan breaks on that baby and you don't notice.

Oh, you'll notice alright!

"What's that smell?"

 

[Denithor]

Originally posted by: Rhonda the Sly
I find it interesting how in the future we plan to run our graphic cards on "Power and Innovation" and governments on "Hope and Change."

Businesses have to make products that we buy to remain profitable and stay in business.

Government has no such limitations apparently and functions more like "American Idol."