R520 Running

MuFu said:
Well, nothing confirmed, but based on recent speculation here, Dave's excellent <a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&k=front%20page" onmouseover="window.status='<a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&k=front%20page" onmouseover="window.status='<a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&k=front%20page" onmouseover="window.status='front page'; return true;" onmouseout="window.status=''; return true;">front page</a>'; return true;" onmouseout="window.status=''; return true;">front page</a>'; return true;" onmouseout="window.status=''; return true;">front page</a> ramblings, Inq/Digitimes etc, past trends and common sense:

0.09u (low-k) @TSMC
Taped out about two months ago so first silicon probably back early this month
SM3.0
FP32
16P/8V
Shader core still based on R3x0/R4x0
Native PCIe and Rialto-bridged AGP versions
Max core clock in the region of 600-700MHz
Die size 250-300million transistors
256-bit, GDDR3/4-compatible memory interface
256-<a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&k=512mb%20memory" onmouseover="window.status='<a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&k=512mb%20memory" onmouseover="window.status='<a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&k=512mb%20memory" onmouseover="window.status='512MB memory'; return true;" onmouseout="window.status=''; return true;">512MB memory</a>'; return true;" onmouseout="window.status=''; return true;">512MB memory</a>'; return true;" onmouseout="window.status=''; return true;">512MB memory</a> at launch (GDDR3@1.2-1.4GHz?)
Possibly higher MSAA modes (8x?)
Multiple GPU capability
Bundled with Gabe Newell - "I like pie whilst I frag" - <a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&k=t%20shirt" onmouseover="window.status='<a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&k=t%20shirt" onmouseover="window.status='<a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&k=t%20shirt" onmouseover="window.status='t-shirt'; return true;" onmouseout="window.status=''; return true;">t-shirt</a>'; return true;" onmouseout="window.status=''; return true;">t-shirt</a>'; return true;" onmouseout="window.status=''; return true;">t-shirt</a>

hmmm....

24P 8V 600mhz.
FP32/SM3.0+ 512MB Ram
New Form of AA not just a sample bump.

Basically

Pro version is 16 P 6V 256mb
XT versio is 24 P 8V 512mb
 
Hellbinder said:
MuFu said:
Well, nothing confirmed, but based on recent speculation here, Dave's excellent &lt;a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&amp;k=front%20page" onmouseover="window.status='&lt;a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&amp;k=front%20page" onmouseover="window.status='&lt;a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&amp;k=front%20page" onmouseover="window.status='front page'; return true;" onmouseout="window.status=''; return true;">front page</a>'; return true;" onmouseout="window.status=''; return true;">front page</a>'; return true;" onmouseout="window.status=''; return true;">front page</a> ramblings, Inq/Digitimes etc, past trends and common sense:

0.09u (low-k) @TSMC
Taped out about two months ago so first silicon probably back early this month
SM3.0
FP32
16P/8V
Shader core still based on R3x0/R4x0
Native PCIe and Rialto-bridged AGP versions
Max core clock in the region of 600-700MHz
Die size 250-300million transistors
256-bit, GDDR3/4-compatible memory interface
256-&lt;a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&amp;k=512mb%20memory" onmouseover="window.status='&lt;a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&amp;k=512mb%20memory" onmouseover="window.status='&lt;a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&amp;k=512mb%20memory" onmouseover="window.status='512MB memory'; return true;" onmouseout="window.status=''; return true;">512MB memory</a>'; return true;" onmouseout="window.status=''; return true;">512MB memory</a>'; return true;" onmouseout="window.status=''; return true;">512MB memory</a> at launch (GDDR3@1.2-1.4GHz?)
Possibly higher MSAA modes (8x?)
Multiple GPU capability
Bundled with Gabe Newell - "I like pie whilst I frag" - &lt;a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&amp;k=t%20shirt" onmouseover="window.status='&lt;a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&amp;k=t%20shirt" onmouseover="window.status='&lt;a href="http://www.serverlogic3.com/lm/rtl3.asp?si=11&amp;k=t%20shirt" onmouseover="window.status='t-shirt'; return true;" onmouseout="window.status=''; return true;">t-shirt</a>'; return true;" onmouseout="window.status=''; return true;">t-shirt</a>'; return true;" onmouseout="window.status=''; return true;">t-shirt</a>

hmmm....

24P 8V 600mhz.
FP32/SM3.0+ 512MB Ram
New Form of AA not just a sample bump.

Basically

Pro version is 16 P 6V 256mb
XT versio is 24 P 8V 512mb

is this what you think it will be, or what u 'know' it will be.
 
Hellbinder, is this new AA method the same thing as the R420 as 4 RV350 cores combined? 8)
 
The Baron said:
Hellbinder, is this new AA method the same thing as the R420 as 4 RV350 cores combined? 8)

no its the huge amazing new feature that the 9800xt would bring.
 
hovz said:
The Baron said:
Hellbinder, is this new AA method the same thing as the R420 as 4 RV350 cores combined? 8)

no its the huge amazing new feature that the 9800xt would bring.
ah yes, OOOOOHHHHHVVEERDRIVE!

I wish Hellbinder wouldn't hype things.
 
The Baron said:
Hellbinder, is this new AA method the same thing as the R420 as 4 RV350 cores combined? 8)

Heh. That looks a hell of a lot better in retrospect than those who were flogging "8 extreme pipelines" at the same time. Unless someone has repealed the 4x4=16 rule recently and I missed the memo.
 
kemosabe said:
The Baron said:
I wish Hellbinder wouldn't hype things.

At least give the guy credit for toning it down a lot lately. What was so ludicrous about the specs he just posted? :?

People just like to bash him because he's been wrong in the past and has hyped things up. *Shrug* His guesses about these sort of things are as good as mine.
 
jb said:
ChrisRay said:
His guesses about these sort of things are as good as mine.

Or he took what SirEric said here ( http://www.rage3d.com/board/showthread.php?t=33796563) and made a guess about the improved AA methods?

Perhaps, I thought this was a speculation thread anyway. I dont see why everyone has to jump on the guy with sarcastic rumors and remarks. Sure he's been wrong in the past. But who hasnt?
 
digitalwanderer said:
Hellbinder said:
New Form of AA not just a sample bump.
For all the newer ATi cards? Or just the new ones coming out? :| (Oh, and thanks for the info and a happy new year to ya! :D )

If there will be a new form of AA it would only for the new ones coming out
 
MuFu said:
Bundled with Gabe Newell - "I like pie whilst I frag" - t-shirt

I fail to see why making fun of his weight has any relevance to this topic. In fact Tim Sweeney's ass kissing to nvidia would be more relevent.
 
kemosabe said:
The Baron said:
I wish Hellbinder wouldn't hype things.

At least give the guy credit for toning it down a lot lately. What was so ludicrous about the specs he just posted? :?
24 pipes, eight vertex shaders, okay. 600Mhz, I really doubt. Add in SM3.0 and FP32 throughout the pipeline, and that clock speed looks more and more ludicrous. And then an entirely new AA method, not just a sample increase? No. Not going to happen. The transistor count would be too insane for the level of cooling ATI is probably willing to put on retail cards.
 
tEd said:
If there will be a new form of AA it would only for the new ones coming out
Nuts! :(

Can I politely and curteously ask why you know/think that? I thought ATi had programmable AA so I figured they might be able to do something new and make it useable on older cards too. :? (I'm not arguing with ya here, I'm just curious. :) )

As for Baron &amp; Hovz, could you lay off HB a little? Do you have any idea on just how much false info can be shoved down someone's throat in pre-release season? At least HB was cool enough to share something with us that he thought was on the up-and-up....I've had quite my share of people trying to get me to post up bogus info and still do. :rolleyes:
 
digitalwanderer said:
tEd said:
If there will be a new form of AA it would only for the new ones coming out
Nuts! :(

Can I politely and curteously ask why you know/think that? I thought ATi had programmable AA so I figured they might be able to do something new and make it useable on older cards too. :? (I'm not arguing with ya here, I'm just curious. :) )

Its only "programmable" in the sampling positions that it takes. There are, however, large numbers of transistors dedicated to making this form of MultiSampling that is in place. When "a new form of AA" is talked about it would imply that this is not MultiSampling, at least in its current form, and would need new algorthims / logic to achitece it, which means different transistor layouts - i.e. this is a hardware thing.

At least HB was cool enough to share something with us that he thought was on the up-and-up....I've had quite my share of people trying to get me to post up bogus info and still do. :rolleyes:

Just don't get hung up on the idea that it is on the "up-and-up".
 
Back
Top