Furmi (GT300) - Page 4

Comixion · #37 11-04-09, 09:16 PM

I need to change my sig to say NVidia.

Comixion · #38 11-04-09, 09:37 PM

drool....all packed inside such a beautiful case.

http://en.expreview.com/2009/10/30/r...its-might.html

Nova · #39 11-04-09, 11:17 PM

I'm not sure if you guys noticed but there are no 5800's in stock anywhere. There's also a lot of people at futuremark and [H] waiting for an x2, so yea... Which brings up another point, I got a 5870! Yeah baby! I could sell this b1tch on ebay for $500.

Quote:

Originally Posted by Comixion

drool....all packed inside such a beautiful case.

http://en.expreview.com/2009/10/30/r...its-might.html

Apparently that card has 8 gigs of ram. 8 f*cking gigs?! I mean ************************, since it's gddr5 and all, might as well use it as system ram and use system ram as a secondary HD.

Comixion · #40 11-04-09, 11:37 PM

Quote:

Originally Posted by Nova

I'm not sure if you guys noticed but there are no 5800's in stock anywhere. There's also a lot of people at futuremark and [H] waiting for an x2, so yea... Which brings up another point, I got a 5870! Yeah baby! I could sell this b1tch on ebay for $500.

.

Or me for much cheaper.

Wait a minute...why sell it?

Nova · #41 11-04-09, 11:44 PM

Quote:

Originally Posted by Comixion

Or me for much cheaper.

Wait a minute...why sell it?

I'm not selling it, gtfo stop eying my card. Shoo, bad Comixion bad bad bad Comixion.

I'm just saying, these things are being bought up faster than they can be made.

Comixion · #42 11-04-09, 11:54 PM

Then you should follow suit and sell it. To me. Don't break the current goings-on.

Nova · #43 11-05-09, 03:55 AM

Just wanted to throw this out there. I'm not sure what it means, but I'm sure you will all figure it out.

3DGuy · #44 11-05-09, 05:44 AM

Quote:

Originally Posted by Avid6eek

I was under the impression that Fermi was on schedule for December...only during the past couple days has news of delays started to leak. Definetly not good.

Back when the 4870x2 was launched, there was lots of rumors about the integrated "sideport" interface, which was to allow fast communications between both GPU's on the card. While the hardware was there, it was never enabled due to some problems. I wonder if such a port, working this time, will be on the 5890 to allow for better scaling.

I've seen some of the 5890 Crossfire benchmarks and they are some seriously high numbers. I'm not sure a single 5890 could replace the 2x X2's though. However your on 24" screen so it could be possible to get similer figures. If you were on a 30" i suspect the X2's would have a bit more muscle overall.

Yea that sideport feature certainly has me interested. The bandwidth would be immense. I think its something like a 6gbs connection at the moment with sideport it takes it up to 22 or something in those scale terms (cant remmber actual figures)

I dont think it was a problem with it if i recal. I think it was just unnessicary to enable because teh bandwidth couldnt be used but it was a implementation with the future in mind. So i have no doubt alot of internal testing will have gone in to that.

Avid6eek · #45 11-05-09, 09:26 AM

With the 5890's performance is a bit of question. It has as many shaders as the 4870x2 and they are on the same GPU die so thre is no crossfire scaling issues, yet it still doesn't consistantly out perform it. ATI is either holding back performance with poor drivers, or the GPU simply doesn't have enough memory bandwidth. I don't think the engineers would have not provided it with the bandwidth needed for peak performance, so I'm hoping there will be a set of performance drivers in the near future.

Quote:

Originally Posted by Skorp

Fudzilla, I will sometimes accept, with a pinch of salt.

Semiaccurate is a disgrace to the internet.

Nvidia is too proud of a company...if these reports were wrong, they would come out and deny them. I haven't heard any denials yet.

dtech · #46 11-06-09, 03:04 AM

Can't wait for Fermi to surface!!! Just imagine Core i9 (6 cores | 12 virtual threads) + Fermi!!! ZOMG!!!

Skorp · #47 11-06-09, 03:14 PM

Quote:

Originally Posted by Avid6eek

Nvidia is too proud of a company...if these reports were wrong, they would come out and deny them. I haven't heard any denials yet.

Well, the site claims that the GTX 260, 275, and 285 were killed off months ago. They don't seem killed off to me. While the focus may not be on the 275 and 285 any more, thanks to the new Radeons, they're still about; and the 260 sells like hot cakes.

I'm still disappointed that we might not get Fermi this year. And I mean that in the way a father is towards a son; not angry, not sad, but disappointed. Ain't that harsher than being angry?

Zefram · #48 11-06-09, 04:27 PM

Quote:

Originally Posted by Avid6eek

With the 5890's performance is a bit of question. It has as many shaders as the 4870x2 and they are on the same GPU die so thre is no crossfire scaling issues, yet it still doesn't consistantly out perform it. ATI is either holding back performance with poor drivers, or the GPU simply doesn't have enough memory bandwidth. I don't think the engineers would have not provided it with the bandwidth needed for peak performance,

I think the issues is with the memory bandwidth. 4870x2 have 2X256-bit memory buses.....

GDDR5 theoretical bandwidth might seem humongous, but GPU performance doesn't scale that much compared to GDDR3 models. That tells me something about the memory efficiency on GDDR5's.

ATi engineer think about yields, they think about less complex GPU, thus the decision to go with 256-bit memory buses and using GDDR5 to cover back the bandwidth.

If nVidia stays true to their 448-bit (or more) memory bandwidth, and by pairing those with GDDR5, the memory bandwidth would be huge. It would ate ATi for breakfast