Wow, if it's truly silent, meaning passive, it may have somewhat low energy consumption after all. That would be awesome. :)

those 270W and whatever numbers are just plain and simple bollocks, a while back the 270W rumor got up from AMD Q&A Session video, where some reporter/journalist/whatever says "I saw a benchmark of R600 with 270W blahblah" - AMD never mentioned anything about it being true, and I think we can safely say that those benches people "saw" were nothing but BS back then?

Good luck getting API support for that mate :) Assuming that any of it is true in the least of course.

Passive? This is how "HIS ATI Radeon X1950 XT ICEQ 3 TURBO SILENT Heatpipe 256MB GDDR3 VIVO TV-Out/Dual DVI" (at least the name is not too long) looks like:


http://www.overclockers.co.uk/pimg/GX-067-HT_400.jpg

Or what people saw was work in progress never ment for public consumption.

Yes, there is "sound chip" in R600 and RV630 & RV610 - but it's meant for nothing but to get the sound out via HDMI with no extra cabling required, it won't be any damn EAX XXX monster, and probably won't be doing DDLive (though I'm not sure would it be even needed on HDMI either)

R600 is almost surely 80nm. And 270 W is very unlikely, as it exceeds the specs of the combined power sources.

But I am pretty sure they have their top model burn the maximum allowed (225 W), simply because they need the performance increase that gives them.

mm, that's a possibility, too.
In the same Q&A video, before that question in fact, they were asked how many watts for that teraflop box, they didn't give straight answer, but said "you're looking at about 200w per card"

As far as I get it, there is no real sound chip as such, but mostly something that does some post processing to make it easier to add sound to the HDMI output.

Nah, it would then still need the cables to get the sound from the actual soundcard / integrated sound chip to the card, so it has to have "full capabilities" if for nothing else, but for straight pass through sound (that's what the "real soundchips/cards" do with digital sound output anyway)

Where are the leaks, just a little Stalker/Obivion sneak peak would be nice.
There should be some perf. info to the press right now?

It-review.net has a fallup writeup on the R600 on its site.




More details.

"The 2600 series should be up and running up to 800MHz clock speed, using unique TSMC 65nm+ process. The 2400 series also has 320 SP's, and we're talking about superscalar ALU as well (dedicated branch execution units and texture address processors)."

Interesting. Just have to figure out if they adding speculation or getting feedback from Africa.

You are talking like you know it... are you so sure? ;)

Urhm... no... it doesn't have to come from the usual Inq/Fud article... In February - way before The Inq/Fud came up with it - a trustworthy source already mentioned a rumour of AMD doing "Sound Physics" next to "normal physics". At the time he didn't know for sure if it was true, but seeing as this word has been popping up it's head more frequently in the last couple of weeks... I wouldn't be surprised if R600 actually was capable of Sound Physics...

NocturnDragon, of course everything is possible, but it would eat quite a bit more transistors at that point, I mean, X-Fi for example has 51million transistors (which of ~20 million is logic, the rest cache), and even Audigy has over 4.6 million transistors.
So yes, I'd throw the possibility of being EAX monster etc out of the window, I can't see them spending even just that 4 million transistors just for the sound, especially on the lower end chips.

Valzic, 65nm was known for both RV610 and RV630, the 800MHz clockspeed rumor floated up few days ago I think, but wtf is that "320 SP's" on 2400 thing? :???:

Nah, drivers can do those things fine.

Then why do the current cards with HDMI include SPDIF-in to get the audio for HDMI ? :???:

Because they don't have the possibility to do the post processing?

What other "post processing" do you need except to make sure that the video & audio go out "in sync", which the current cards do with the digital audio signal they get via SPDIF-In?

Then Inq knows the term, but not how to apply it.. typical klok / klepel situation.

http://www.theinquirer.net/default.aspx?article=39042

To me this sounds like nonsense. "65nm+ process"? 320 SPs for 2400-series?

That you didn't know, all Radeon HD's will be the same, just the price will be different.. AMD really wanted the performance crown in all segments this time.. That's why they are selling millions to OEM's

Yeh, and in reality, the huge HSFs and extra powerplugs on 2900's are just for the imago there, i mean, it would look silly if most powerful card in the town would be looking like some lowend with short PCB and small 1-slot cooler, no extra powerplugs etc :grin:

Well, whatever you would like and GPGPU can do, I think.

LOL....:grin:... In related news, it was just revealed that every Radeon HD 2xxx, from 2100 to 2985.75 will have TWO R600, one 80nm and one 65nm one. AMD is shooting for 100% marketshare.