The ATI Radeon 5XXX Thread

[jaredpace]

I spotted this on Wiki - dunno where this pic came from and if it's valid, but FWIW;



edit:
Looks like fake, tbh.

I guess it doesn't let you link to quotes in closed threads.

Largon, Here is what I said about it:

...similar to dual core, yet not actually two dies in one package ala MCM). Perhaps there is no more real-time compiler in the driver and its all handled on the hardware level by the scheduler. Because the core of the chip is so modular & scalable with certain areas sharing parts of the die (ROPS + Memory controller logic), you are able to divide the specs in half (1600/256/80/32 to 800/128/40/16) and have two parts (rv870/rv830) and rv870 appears as two rv830's, yet it is still only one die. Hence term "dual-core" - more like "modular". Some people think, yes, but all GPUs are multi-core because each of the shaders is like a core by itself. Well, here we are dealing with two large arrays of 800 shaders, along with other standalone logic that communicates with either one of the two identical arrays. More specifically, one Rv870 die is composed of two Rv830-like 160x5 clusters/shader arrays (like two current-gen revamped rv770's in one die) sharing certain features, but connected via "internal crossfire" working in unison and the entire design is a continuation of R600 architecture. It is load balanced, efficient, and requires no crossfire software optimization (because it is hardware level communication); it works via SFR 8*8-32*32, and is bandwidth intensive. The board is using next-gen 5ghz Gddr5 to provide the required bandwidth. So, apparently they've slapped together two 40nm rv770's... so it's easy to see where "dual-core" confusion comes from. Cypress is like a 40nm 4890x2 in one die! -So, it is like a "native dual core" CPU...

To me it sounds like a "dual-core GPU", because even though there are 1600 shaders/stream processors, they are in two groups of 800. 10 SIMD x 2. That's all, the guy over at Hardocp forums called it "dual shader mega clusters" CJ at tweakers called it "dual shader" and Theo called "dual shader engine" at BSN. Dozens of posts over chiphell they call it "native dual core gpu". Occasionally someone comes into the Beyond3d thread and declares it to be "native dual core". One guy brought up a point earlier and showed a gt200 die shot and said, "well i guess you would call this quad core?" (because it looks like 4 quadrants of unequal size, or two equal halves?) Well you could cut it in half down the middle and seem to have two identical parts, but are those 2 parts rendering via SFR internally, and designed that way to be seen as a single gpu? Probably not. There are lots of threads on "multi-core" and "interconnect" that are way more technical than what I can understand. Either these people are completely mistaken, or it is actually using SFR internally, and sharing the MC. Remember, we still have a bet

Here is a quote from the technical discussion of the core at chiphell. (It sounds goofy because of the translation, but see what you make of it):

Since the House passed 100 to summarize it:

We all know that RV870 is a native dual-core, MEM share, TMU sharing, etc. These hardware innovations is just the beginning, or rather in addition to hardware of the new technology, RV870 brings us the updated software in-depth , for example, CrossFire mode, specifically can refer to the savage major technical paper, I did not explain the response to this (http://bbs.chiphell.com/viewthre ... & extra = page% 3D1)
RV870 by native dual-core hardware support, the biggest benefit is the CF model can be evolved from today's AFR for a more reasonable SFR (do not understand what AFR and SFR for us to look savage Great article,), SFR and thoroughly get rid of the PCI-E bandwidth bottlenecks leading to CF or SLI can not use the SFR of the constraints, CF no longer rely on driving on the optimization of 1 +1 is infinitely close to two and said, where everyone will be RV870 native dual-core architecture of the deep deep impress, at least in my case, native dual-core is my first time in years, felt playing card graphics technology, a major step forward.
Go back to Crysis on the results, in fact, the RV870's Crysis performance can be regarded as a strong reason why the bounds of reason, but it can be regarded as lucky, because to say whether it is before the CF or SLI, the efficiency of both in Crysis is not high, RV870 just drill the use of technology, this SFR loopholes and completed card crisis, the role of Terminator, but I hope that everyone calmly RV870 performance, in other CF or SLI more efficient GAME in, RV870 impossible to think about Crysis so lucky, though it in terms of the total, RV870 architecture as pioneered the use of dual-core technology, combined with the many part of the shared framework, the core DIE or the ratio of power consumption and performance will be the same Performance Products the highest, which is beyond doubt.

Lao Lao hastily said so much, would have wanted to test them on the RV870 article says, but I got fascinated by the RV870 architecture could not resist spraying a bit out, more content or waiting for me in the future release of the RV870 test, let us share the bar.
Quietly asked a small problem, AFR to SFR does not seem to call it evolution, can only be called after the performance to a certain extent the development of an inevitable trend of bar