• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

R650 & R700

S

ShadowOfMyself

Diamond Member
R650

Our sources confirmed that ATI is working hard to release its R650 chip. The 65 nanometre chip is getting closer to launch day by day and aims for some high clocks.

This won't just be a simple die shrink of an existing marchitecture as ATI plans to redesign some parts of the chip. The major architectural difference is that R650 will have more texture power as this was recognizes as one of the weakest points of R600.

ATI aims for higher clocks, closer to 1,000MHz and obviously it will improve the usage of the available bandwidth.
Click to expand...

R700

We got confirmation that the R700, a chip that is scheduled for the 1H of 2008, is going to be manufactured at 55 nanometre. The 55 nanometre process is an optical shrink of the 65 nanometre process and it is the logical step for the graphics guys. The CPU guys will go from 65 to 45 nanometre instantly, but the CPU guys didn't use 80 nanometre either as they decided to go from 90 to 65 nanometre.

Our knowledgeable sources also confirmed that R700 is a multi core design and DAAMIT calls it Multi Core Unified Architecture. This sounds like a couple of cores stitched together in a high end chip and a single core for a small chip.
Click to expand...

Let the speculation begin 😀

Well, I for one am happy they are increasing the texture power, and also start to think its really pointless to buy HD2900XT with such a better card coming in 6 months, which gives them ample time to perfect their drivers

I know I know "you can play the waiting game forever" but what I mean is with all these drivers issues, there really is no point in buying the card now, since by the time they are fixed R650 should be launching
 
Just FYI: don't ever use "marchitecture", it's not only not a word, it's not even a legit abbreviation. If you delete letters from a word you denote that with a hyphen or apostrophe. Not to mention they're barely saving any time or space by doing that. They're just being stupid. Oh, right, Fudzilla. 😀
If you want to replace "micro" with a single letter you use the special "u"-like symbol that means "micro".
Idiots...
 
Originally posted by: yacoub
Just FYI: don't ever use "marchitecture", it's not only not a word, it's not even a legit abbreviation. If you delete letters from a word you denote that with a hyphen or apostrophe. Not to mention they're barely saving any time or space by doing that. They're just being stupid. Oh, right, Fudzilla. 😀
If you want to replace "micro" with a single letter you use the special "u"-like symbol that means "micro".
Idiots...
Click to expand...

LOL I am with you there, I hate the words they make up... Also all the MADATI DAMMIT and whatever... sounds stupid and pointless

But eh, nothing better for speculation than Fudzilla 😀
 
Oh God, not already!! Can't we have just a LITTLE break from the "What if..." threads?


Pretty please?
 
Originally posted by: Creig
Oh God, not already!! Can't we have just a LITTLE break from the "What if..." threads?


Pretty please?
Click to expand...

But but... this forum looks so dead when there isnt any speculation going on 😛

Oh well, Ill let the crowd decide
 
Originally posted by: ShadowOfMyself

LOL I am with you there, I hate the words they make up... Also all the MADATI DAMMIT and whatever... sounds stupid and pointless

But eh, nothing better for speculation than Fudzilla 😀
Click to expand...

I think AMD should step up and announce that ATI will now be renamed from

ATI Technologies Incorporated (Yes, it is a recursively stupid name)

to

AMD Technologies Incorporated (maybe even give the "A" some AMD green)

As to speculation, I will speculate that viral marketers and "user group members" will derail this thread with bone-headed comments.

(I also speculate that some member who hasn't been here for two years will claim they started a thread on this subject two years ago)
 
Up not to bring up the NV30-->NV35 analogy but it does sure sound like it. This isnt necessarily be a bad thing since NV35 was actually much better than the former. Increase in TMUs will be great, but are the number of shaders being increased to?

ATi just cant add in more texturing units because of the way R600 is designed. There are 4 clusters, each with 16 vec5 ALUs -well 16 shader units with 5 sub scalar units per shader unit. Each clusters also house 4 "TMUs" (filtering units, address units, etc etc), and each cluster is then tied with one quad ROPs. So if they want to increase one part of their architecture, other parts will have to increase at a proportional rate to keep the ratios between the architecture the same unless the design layout is changed.

I wonder if they will bring back MSAA hardware support, instead of having them done on shaders though. R650 sure does sound like what R600 shouldve been. If they are aiming at 1ghz core clock, that would be one helluva of an engineering achievement.
 
for R650 do you think they might fix the broken MSAA part ? its been done on the shaders now isnt it? which lets you go all programmable on the jaggies ass, but it would be cool if they got a normal MSAA engine working so that for just standard AA modes you could get small performance hits instead of the some times large ones we've been seeing.
 
Originally posted by: yacoub
Just FYI: don't ever use "marchitecture", it's not only not a word, it's not even a legit abbreviation. If you delete letters from a word you denote that with a hyphen or apostrophe. Not to mention they're barely saving any time or space by doing that. They're just being stupid. Oh, right, Fudzilla. 😀
If you want to replace "micro" with a single letter you use the special "u"-like symbol that means "micro".
Idiots...
Click to expand...

Actually, the IT Tabloid definition of marchitecture is short for "Marketing-Architecture", as the description of the architecture is what the marketing department sells techs on. Not "micro-architecture".
 
Originally posted by: ShadowOfMyself
Click to expand...


Let the speculation begin 😀 [/quote]

I speculate that they "have the card working and could ship it right now" but they are waiting on the mid-range so they can ship them all at once...

😉

I couldn't help it..

 
Originally posted by: SunnyD
Actually, the IT Tabloid definition of marchitecture is short for "Marketing-Architecture", as the description of the architecture is what the marketing department sells techs on. Not "micro-architecture".
Click to expand...

So it's still an inappropriate abbreviation as it lacks a hyphen or apostraphe or other designation that there are letters missing between the "m" and the word "architecture".
 
Wasn't there some stuff on the inquirer about R700 having a modular design?With different number of units making up the high-end,midrange and budget segments?

R650 may very well perform better than R600 but its gotta be cheap,quiet and power efficient to stand a chance against the nvidia card of that time?(G90 possibly?)
 
That R650 looks interesting - the drivers would be better by then and the process shrink should cut down on the power consumption - but they need to bring it out quickly (in the next two or three months) for it to have any real impact.
 
is getting closer to launch day by day
Click to expand...

I don't think anybody is going to fall for that again.

It should also be noted that all that "information" is from Faud of the Inq. So take it with a dump truck of salt.
 
Originally posted by: ShadowOfMyself
Originally posted by: Creig
Oh God, not already!! Can't we have just a LITTLE break from the "What if..." threads?


Pretty please?
Click to expand...

But but... this forum looks so dead when there isnt any speculation going on 😛

Oh well, Ill let the crowd decide
Click to expand...

If you're bound and determined to go through with this, then at least PLEASE, PLEASE, PLEASE do NOT link to any Inquirer or Fudzilla article that contains the word "confirm". As in "Our sources confirm" or "We got confirmation", etc... Because the only thing those two sites have actually ever "confirmed" is that they're clueless idiots who haven't figured out where to find the spellcheck button.
 
Originally posted by: Wreckage
is getting closer to launch day by day
Click to expand...

I don't think anybody is going to fall for that again.

It should also be noted that all that "information" is from Faud of the Inq. So take it with a dump truck of salt.
Click to expand...

actually thats TRUE.
unless amd plans to drop out of the graphics market completely we are getting close to the R700 launch day by day. be it launched tomorrow or the expected ship date of 2019
 
I'm not an Nvidiot, but IMO the r600 is almost as bad as the nv30, and it seems like the refresh will need to have some redesign done, besides the obvious die shrink and speed bump. Right now Ati has no hope of competing in the high end sector against the 8800gtx, and in the future against the 8900gtx. I'd like to see HW resolve for AA added back to the ROP's (or fixed if it's broken in the r600), and increased texturing capability, because once again Ati is lagging in this area. Also, while the AF quality of the r600 is not bad, it's still the same as the r520, and not as good as Nvidia's.
 
Originally posted by: Drayvn
When did the CPU "guys" skip 80nm on their CPUs?
Click to expand...

When they went from 90nm to 65nm.
CPU's always skip. 250 -> 180 -> 130 -> 90 -> 65 -> 45 -> 35(?)
GPU's 250 -> 220 -> 180 -> 150 -> 130 -> 110 -> 90 -> 80 -> 65 -> 55(?) -> 45 etc

RAM chips also do the intermediate processes.
This is one reason why I personally never say the idea of ATi using AMD's manufacturing capabilities as something that was ever going to happen.
 
Until they integrate with Fusion or whatever. =)

Besides since CPUs are a generation ahead of GPUs in size, maybe they could make a second use of each CPU fab level for subsequent GPU production.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top