RV870 is NOT the codename for next-gen ati-parts

[MarcVenice]

I had an interview with Sasa today. And when I mentioned RV870 to him, he said it does not exist. Of course they are working on something, but it does not have the codename RV870. This has led me to conclude that any website, like hardware-infos.com, is making specifications up when they call it RV870. They surely would have known the 'real' codename if they have an insider.

Also, this is probably somewhat of a marketing ploy, and I don't know if Sasa was being overly optimistical, but he said AMD feels very confident about their next-gen part. If the rumours about Nvidia's gpu have a hint of truth to them, he is not worried at all.

He was pretty darn sure new parts will arive in Q4, or at least will be available for purchase in Q4. When I think about it, he didn't completely rule out an earlier release date. But he was confident new parts will at least be in stores in Q4.

He could not comment on tape-out, doh, nor could he comment on HD4890X2. He did say there will be an interesting keynote during Computex, but didn't say what it was about, but it was not about a product.

Now you know what I know, to bad it ain't much

AMD has contacted us, and are now neither denying or confirming the arrival of new gpu's in Q4. In essence, no comment. Take it for what it's worth. I can't comment any further on this.

 

[apoppin]

http://tweakers.net/nieuws/603...n-vierde-kwartaal.html

cool

 

[Idontcare]

Originally posted by: MarcVenice
I had an interview with Sasa today. And when I mentioned RV870 to him, he said it does not exist. Of course they are working on something, but it does not have the codename RV870. This has led me to conclude that any website, like hardware-infos.com, is making specifications up when they call it RV870. They surely would have known the 'real' codename if they have an insider.

The same was true of K8L, aka K10 eventually.

However that didn't mean all the early rumors regarding K10 were unfounded...just mislabeled by insiders and outsiders alike.

Just as an aside, several architects at AMD expressed puzzlement at the origins of the K8L name. There is an internal engineering code name for the project, but the marketing team is slightly behind and has yet to provide something catchy for the rest of the world. At least one architect at AMD indicated a preference for the name K8++, but it seems unlikely that anyone in marketing or PR would share this point of view.

http://www.realworldtech.com/p...icleID=RWT060206035626

 

[MarcVenice]

Well IDC, I could be wrong, but I still think if they know real specs, they also know the real codename? I just like to think of it as another nail in the coffin thats called credibility, for websites claiming to have specs.

 

[Keysplayr]

Is brightsideofnews next on your to-do list?
Really Marc, you come up with entirely too many conclusions with not enough information. Consider the possibility that Sasa is lying through his teeth for reasons unknown to us. You need to consider things a bit longer before arriving so quickly to these conclusions. In corporate biz, 1+1 doesn't always equal two.
Apologies in advance for appearing short with you. Not the case. Just discussing.

 

[Idontcare]

Did Sasa tell you what the real codename is?

If he didn't, and all he did was refute claims of the existence of an RV870 then for all we really know it does exist but he's hiding behind the "AMD won't officially comment on unannounced products" fine print.

But yes I do agree with the spirit of your thread, history tells us just how badly and off the mark the rumors regarding GPU's tend to be. In fact I have to scratch my head and think real hard of the last time I actually read a pre-release scoop on a GPU that actually turned out to be true.

I personally think it is intentional subterfuge by the GPU makers themselves. The websites aren't necessarily putting up fantasy stats on upcoming GPU's, they really are being fed info from insiders, but unfortunately they are just being played as a pawn in a greater marketing game of seeding false info to try and bluff the other guys trying to sift the tea leaves for competitive info.

Consider for instance that you really have to ask yourself why would AMD bother to spend salary money on Sasa to talk to you...they aren't doing it out of the goodness of their hearts to throw you an interview so you personally had something to do with yourself today...and if past rhetoric from AMD serves an example of anything it is "buyer beware" when it comes to taking "free" information about future products and their performance envelopes :laugh:

One would hope they are straight shooters, but there is an angle for them in this otherwise they wouldn't spend the money to waste their time doing interviews in the first place.

Which goes back to my original question - did he tell you what codenames do exist, or merely give you enough ammo to shoot down the other rumor sites?

edit: keys slipped in there before I posted, didn't see that, so I'm not the only one thinking the obvious here...

 

[MarcVenice]

Pfff, it's theo valich, former inquirer 'journalist'. Not on my to do list, their wild speculation kinda annoys me though.

Really, why would sasa lie to me about the codename? I know he is pr-guy, he will often not tell the whole truth, or answer things as vaguely as possible. When I mentionRV870, and he sais it does not exist, why would he be lying? He would also know that I'd be making a fool out of myself and the website I write for, if it's false. I doubt he'd want to get a bad rep with the press/us. If he didn't want to say anything about the codename, he would have said: no comment. It would have been the tenth time, so it's not like it would have mattered...

@ IDC, no he didn't. The other PR guy was telling him not to He flat out said RV870 is NOT the codename. No denials of it's existence. There is a codename, it just aint RV870. Oh, and he wasn't here just for kicks. The initial meeting was about computex-stuff, under nda. But we got enough time to ask any questions we had.

 

[Keysplayr]

Marc, it could be that the name is R775 or R880. And because you didn't ask for that model tag specifically, he could easily say it does not exist and NOT be lying to you in the slightest bit. They are just codenames after all, and have been known to change.

 

[MarcVenice]

Originally posted by: Keysplayr
Marc, it could be that the name is R775 or R880. And because you didn't ask for that model tag specifically, he could easily say it does not exist and NOT be lying to you in the slightest bit. They are just codenames after all, and have been known to change.

That's silly Keys. Of course it could be anything? I'm not saying I know what it is, I'm not saying its 'monkeylikebananasX2'. It just ain't RV870. That's all I'm saying.

In fact, I didn't even ask if RV870 was the real codename. I mentioned it, and he said it's NOT the codename. We were ofcourse talking about the high-end part. Really Keys, I know what he said, and what he didn't say. This isn't anything revolutionary, I just thought I'd share.

If you disagree with my conclusions, so be it.

 

[Keysplayr]

Originally posted by: MarcVenice

Originally posted by: Keysplayr
Marc, it could be that the name is R775 or R880. And because you didn't ask for that model tag specifically, he could easily say it does not exist and NOT be lying to you in the slightest bit. They are just codenames after all, and have been known to change.

That's silly Keys. Of course it could be anything? I'm not saying I know what it is, I'm not saying its 'monkeylikebananasX2'. It just ain't RV870. That's all I'm saying.

In fact, I didn't even ask if RV870 was the real codename. I mentioned it, and he said it's NOT the codename. We were ofcourse talking about the high-end part. Really Keys, I know what he said, and what he didn't say. This isn't anything revolutionary, I just thought I'd share.

If you disagree with my conclusions, so be it.

The "monkeylikebananasX2" would be my second choice for their codename.

 

[Idontcare]

At least you guys at ABT are trying to go straight to the horses mouth on this stuff, unlike the others in the field that publish whatever they get in an email from anyone claiming to be an insider. That is admirable of ABT even if it keeps you from expounding on the existing rumors for NDA reasons.

 

[heyheybooboo]

He ain't talking because it's a 40 nm SOI GPU, a variant of which will be used in the AMD Fusion CPU.

Hence, the new codename: (con)Fusion

 

[Grinja]

Originally posted by: MarcVenice

In fact, I didn't even ask if RV870 was the real codename. I mentioned it, and he said it's NOT the codename. We were ofcourse talking about the high-end part.

Maybe I'm taking things out of context ... isn't the highend (X2 version of RV870) part R800?

 

[Grinja]

Originally posted by: Idontcare
At least you guys at ABT are trying to go straight to the horses mouth on this stuff, unlike the others in the field that publish whatever they get in an email from anyone claiming to be an insider. That is admirable of ABT even if it keeps you from expounding on the existing rumors for NDA reasons.

:thumbsup:

 

[cbn]

Could it be possible that ATI is planning on using the ring bus patent Micron owns coupled with three or four 40nm RV740 GPUs? Very recently I read that one of the GPU companies was planning on using Micron GDDR.

If this were true there wouldn't neccessarily need to be a separate RV870 core right?

 

[cbn]

http://www.techpowerup.com/index.php?94330

Maybe I was wrong. It says here that Micron isn't even making DDR5 yet.

 

[Tempered81]

cypress & juniper/jupiter?

I heard these codenames for rv870

 

[MagickMan]

So, new GPUs the beginning of next year? Check.

 

[Idontcare]

Originally posted by: jaredpace
cypress & juniper/jupiter?

I heard these codenames for rv870

apparently the codename is evergreen

 

[OCGuy]

Originally posted by: Idontcare

Originally posted by: jaredpace
cypress & juniper/jupiter?

I heard these codenames for rv870

apparently the codename is evergreen

ATi says theyll be released Q4, that "article" says they are around already and could get out in July?


Someone is very wrong.

 

[Idontcare]

Originally posted by: OCguy

Originally posted by: Idontcare

Originally posted by: jaredpace
cypress & juniper/jupiter?

I heard these codenames for rv870

apparently the codename is evergreen

ATi says theyll be released Q4, that "article" says they are around already and could get out in July?

Someone is very wrong.

Well you are definitely right to be suspicious of BSN (bullshitnews) but in this case their language actually speaks some truth IMO.

Given the nature of TSMC's 40nm yield issues (it's a parametric problem, not a functional D0 issue per say) it definitely is a plausible situation that AMD could be ready now to start ramping volume on evergreen but until the yield issue is corrected there is simply no point to throwing more wafers at the situation.

So AMD could say Q4 and mean it, but if yield problem was corrected 2 weeks ago then AMD might be in the position to order full volume on evergreen and could have product in the channel within 2 months after that (or mid July in this hypothetical timeline).

Knowing TSMC though, I give AMD's stated timeline the greater likelihood of becoming reality than BSN's optimistic take on it.

 

[ronnn]

Must be the not caring that makes it possible to learn things. Nice simple language for us dummies ... thanks. :beer:

 

[cm123]

R870 - HD5870

why would though a focus group member or like (beta) person tell all right now???

This is the 1st (possible) non value (RV) part in awhile as well (for AMD/ATI)- did he share that with you?

I would (if we are just tossing everything out there to the wind) opt to disagree with Sasa, while AMD might have a part sooner than later, Nvidia next part is in fact a product to respect - either this guy is full of crap or telling you little to no truth at this point...

I would tend to accept a big new release coming this week from AMD and while they may have 1st dx11 part, will not beat Nvidia next, however they may in fact by that time be ready for yet another release that will stand on its own to the next Nvidia part...

end result really (again if we are guessing away), looks to be a very nice release of cards over the next 2 to 8/10 months...

 

[SickBeast]

Based on the estimated die size of the GPUs shown today, I'm not holding my breath for a performance monster. What we will likely have is a very good midrange GPU. I have a feeling it will match a 4890, or else beat it by 20-30%, while being the same price or cheaper ($150-200).

I'm sure that the economic downturn is going to lead to cheaper GPUs with less features.

 

[ronnn]

At this time many of us (with not large monitors) don't have a need for anything faster. Hopefully dx11 will bring along programs/games that change all that.