AMD: R9xx Speculation

[Harison]

I as a consumer just want my product run as good as possible and i have to say that on Nvidia's side i am a lot more secure in that case.

Its admirable when Nvidia or AMD helps devs to make games better, thats NOT the issue. Problem starts when one manufacturer pays devs to specifically harm other graphic maker, along with its customers. Just because Nvidia is doing it, do you really want AMD to start doing it too?

Like we joked when AMD was helping Dirt team to get DX11 implemented, if done in Nvidia way, those features would be Radeon only, with Geforce cards locked out.

On a more serious note, I'm glad AMD isnt doing back-stabbing stuff, but they could indeed increase their devs help budget, which was probably an issue before due to lack of funds, AMD was in red for a long time.

 

[Unknown Soldier]

Thx god, their is one company who wants Tessellation in games before 2012.

I want Tessellation in games but I prefer playable frame rates over non playable frame rates thank you very much.

 

[Rangers]

Very nice, didn't expect it to be beating the 470. No doubt Nvidia got ahold of these cards a few days ago, saw the writing on the wall, hence the recent price slashes.

Cant wait to see how Caymen does as well, as great as 6870 performs already, wow. I wonder how high they can core clock it. If they get anywhere close to 900 mhz it will be a monster.

 

[Kaarlisk]

Most dual-link panels these days support DP as well. You're really looking at the 30" and the Dell 27" and the newer models carry DP.

New monitors are not what I'm thinking about. On the other hand, it's really good to see DP taking hold

And sorry for not saying this earlier. Congrats on a great product that really deserves its name!

 

[Zaphod]

You apparently didn't get to the tesselation bit of Hiiiiiiiiiiilbert's review before it was pulled.

Can't escape teh Googlies:

http://webcache.googleusercontent.c...ru3d.com/article/radeon-hd-6850-6870-review/1
...
http://webcache.googleusercontent.c...u3d.com/article/radeon-hd-6850-6870-review/23

Edit: Apparently you can if you're really popular and get reindexed quickly. It was there, and now (apart from 1-3, 23) it's not.
Edit2: Aaand it's back. Might just be dependent on which cache server you get.

 

[Dave Baumann]

New monitors are not what I'm thinking about. On the other hand, it's really good to see DP taking hold

Well, the number of panels that actually require dual link are limited - single link serves all the way up to 1920x1200, and its only beyond that where dual link is required, so you are mainly looking at the 30" panels (120Hz panels). While I think a few of the 30" vendors still have dual link the first generation the Dell refreshed their panels fairly quickly the 3008 has been on the market for a few years now with DP inputs.

 

[Harison]

It was slow, but I managed to open all 23 pages in seperate tabs, in case they decide to pull it.

After quick glance: matches 5870, beats GTX470 in some benchmarks.

Thanks thatdude90210for the link, and DarthShader - go go upload pages, I also was opening in Tabs and after 14th page noticed they puled off the review

 

[Psycho]

Does the single link dvi mean that we can now use both dvi AND the hdmi? (similar to a cheap single link dp adapter for hd5xx0)

 

[thatdude90210]

Heh, someone else found this but it's amusing. Google cache has all of it:
http://webcache.googleusercontent.c...ru3d.com/article/radeon-hd-6850-6870-review/1

Just change the number at the end, 1 to 23 for each page.

 

[DarthShader]

Already done, not by me though.

nothing can be hidden on the internetz
guru3d review - http://www.mediafire.com/?m5qgbb35ysutw1y#2

NOTE: post here if this is illegal. i will remove the link

Dunno if this is complete though.

 

[John021]

DAMN!

 

[digitalwanderer]

I believe you mean, "DAMN!".

 

[Bouncing Zabaglione Bros.]

Prices aren't too bad for launch day.

http://www.overclockers.co.uk/productlist.php?groupid=701&catid=56&subid=1866

Yeah, not bad at all, especially considering it's launch day pricing and prices may well go down in the face of Nvidia's price drops. The 6870 seems to be the same or a bit less performance than 5870, for less money than a 5850. That's a lot of performance being pushed down the price ranges.

 

[Alexko]

DAMM!

That can't be right. The 6870 can't have the same TDP as the 5850 and draw 50W less… What's going on there?

 

[compres]

New line looks solid, ETA for 6900 is November? nVidia must be in panic mode, no wonder the usual suspects are spamming FUD here today.

 

[no-X]

They don't have the same TDP. I think it's a typo in HD6800 specs, because HD5850 specs always stated 170W (only the HD6800-PDF states 151W for HD5850).

 

[neliz]

They don't have the same TDP. I think it's a typo in HD6800 specs, because HD5850 specs always stated 170W (only the HD6800-PDF states 151W for HD5850).

LolwuT?

HD5850@AMD.com

Speeds & Feeds

* Engine clock speed: 725 MHz
* Processing power (single precision): 2.09 TeraFLOPS
* Processing power (double precision): 418 GigaFLOPS
* Polygon throughput: 725M polygons/sec
* Data fetch rate (32-bit): 209 billion fetches/sec
* Texel fill rate (bilinear filtered): 52.2 Gigatexels/sec
* Pixel fill rate: 23.2 Gigapixels/sec
* Anti-aliased pixel fill rate: 92.8 Gigasamples/sec
* Memory clock speed: 1 GHz
* Memory data rate: 4.0 Gbps
* Memory bandwidth: 128 GB/sec
* Maximum board power: 151 Watts

TDP doesn't equal consumption (nvidiwatts) of course and I dont' think guru3d does accurate measurements but just wall measurements.

 

[Alexko]

They don't have the same TDP. I think it's a typo in HD6800 specs, because HD5850 specs always stated 170W (only the HD6800-PDF states 151W for HD5850).

It says 151W here: http://www.amd.com/us/products/desk...5850/Pages/ati-radeon-hd-5850-overview.aspx#2

Even so, that would explain a ~20W gap, but not 50W. It must be a typo in Guru3D's chart or something…

Edit: damn, too slow.

 

[3dilettante]

The TDP is a value used to spec out the maximum dissipation for a thermal solution for the product line. It is more a statement that the chip may dissipate a given number of watts in some load situation.

Individual chips can vary, and this is apparently an average over time.
Superior power management could allow for lower numbers in non-peak situations, but there could still be times when the chip spiked to similar power levels as its predecessor.
This assumes the methodology is sound.

 

[DavidGraham]

Tessellation :
http://www.itocp.com/htmls/97/n-1497-21.html
http://www.itocp.com/htmls/97/n-1497-22.html