The LAST R600 Rumours & Speculation Thread

Status
Not open for further replies.
r520 was a fiasco, it launched what, almost half a year after g70? then a new card launched 2 months later? sucks to be someone who actually bought r520. i dont even recall it being much faster than g70, wasnt it a win some lose some?

I bought an X1800XT and i thought it was great. It did everything i needed it too, maxed all settings and was smooth. The R580 only excelled at more extreme resolutions if i remember correctly, beyond 1600x1200, but below that they were quite close in benchmarks.

It wasnt bad either, it had the best imagequality to date by a HUGE margine, it did AA + HDR, it was the first widely available 512mb card and the new memory controller eventually allowed some nice benefits with AA in OpenGL games while making bandwidth hits in pretty much everything else minimal. It was only about 15-20% faster then a stock, i stress the word stock, 7800GTX if i remember right, however the lead become much more apparent with HQ IQ.

The only fiasco that launch had, like this one, is a large gap with no competition in place, and getting rid of a product alltogether simply isnt the answer. We dont even know what the refresh of the R600 will have in store for us, and its release is pushed further back, into late Fall even, if they attempt to go 65nm on it to maximize yields though that depends entirely on the launch schedule for the next product.

By the way R520 hard launched in the first week of November with the X1800XLs and the XTs following at the end of the month.


http://www.hkepc.com/bbs/attachment.php?aid=497181

If you take the HKEPC configurations seriously, and compare that to what xbit reported for reasoning from AMD, then I think one of the areas you have to look at would be what role did memory configurations play in these decisions?

HKEPC is reporting that XTX is 1GB of GDDR4. Whereas XT and XL are 512MB of GDDR3. And, as we know, R600 supports 512-bit buswidth to external memory.

G80 on the other hand, has 768MB of GDDR3 on 8800GTX and 640MB on GTS, at 384-bit and 320-bit, respectively.

HKEPC also says that the GDDR3 versions would not be ready until April. So between xbit and HKEPC what I've been wondering is if AMD/ATI just decided why they should be at a cost disadvantage at launch with 1GB GDDR4 when they could still have more bw than the competition with less cost at 512MB of GDDR3 in April/May? It seems to me if XTX is even roughly performance competitive with 8800GTX, then the XT that HKEPC is describing should thoroughly trash an 8800GTS 640MB.


If you honestly believe they arent actually having a physical problem with the chips themselves then i can only say that is some really terrible marketing. Every week the 8800 series is left unchallenged the more people will give less of a damn when AMD launches. Alot of the sales of any series are from word of mouth, not based on what the latecomer does better and right now the 8800 has alot more street credit....actually all the street credit. Thats why the 7800 and 7900 cards were power sellers even after the X1800/X1900 came along and did things better. This was actually the good of paper launches because at least then people got an idea of what to expect. This though, would be a reverse paper launch, this would be the stay tight lipped until the last minute even though we're extremely late and losing prospective buyers all the time strategy.
 
Last edited by a moderator:
I think companies usually tell the truth about why they do something (if they address it at all), even if it is spun as self-servingly as possible. A delay to fix a physical problem doesn't fit with what they've said at all, so far as I can see.
 
what they've said doesnt exactly justify the extreme measures they took either, such as event cancellation. If they pass up mass press coverage at CeBIT as well i'd think it very strange.

If you have a good product you show it off, you dont stuff it in the cloak room, especially when you're a company who needs some positive sales and buzz.
 
If you have a good product you show it off, you dont stuff it in the cloak room, especially when you're a company who needs some positive sales and buzz.

I' am starting thinking that we have "R600" or upcoming another NV30, except it is revision 2. (Just maybe)
 
what they've said doesnt exactly justify the extreme measures they took either, such as event cancellation. If they pass up mass press coverage at CeBIT as well i'd think it very strange.

If you have a good product you show it off, you dont stuff it in the cloak room, especially when you're a company who needs some positive sales and buzz.

I'm sorry, I'm not following you here. Are you advocating a paper launch then?
 
I'm sorry, I'm not following you here. Are you advocating a paper launch then?

Doesnt have to be a classic paper launch to the letter, just show the hardware running some sexy looking demos and do it for the public, talk about some of the features, why not? The earlier they show something the better. Screw HardOCP if they're going to knock AMD for doing it let them. Showing something is indeed better then nothing at all. Hard launches this late really dont matter because they arent out to beat the 8800, they're trying to battle the clock (clock as in time, not mhz).
 
I think ATI should be worried, it might be predicted that the R600 will be faster than the 8800GTX, but Nvidia's partners might spoil ATI's fun soon enough(if the 8900GTX doesn't).

The G80 just seems to get faster and faster, so colour me impressed.

If the 8900GTX does have 25% more shaders, then it should give the R600 trouble I'd think.

Anyways, the R600's delay also has me scratching my head since it can't be driver issues.

US
 
Last edited by a moderator:
I think companies usually tell the truth about why they do something (if they address it at all), even if it is spun as self-servingly as possible. A delay to fix a physical problem doesn't fit with what they've said at all, so far as I can see.

Throw another conspiracy theory on the pile. No one really knows at all why they did this, other than that it can't be good.
 
A wizzard is never late. He arrives precisely when he means to :D

Jokes aside, while I understand that everyone wants to know the real reason of the last (unexpected) delay, it doesn't seem as important to me personally. It could be any kind of serious issue that slipped through anyone's attention until last minute or something that sounds way too simple or stupid for most to digest.

Whatever it is I personally prefer ATI to make its best possible appearance, rather than risking not to win the impressions they want to. A delayed product-line has to present distinct advantages of it's straight competitor even more so if it arrives that much later. Just like with any acquaintance the first impression is way too valuable to screw it up.
 
Don't mind me, just gathering my thoughts after 90pp.

http://www.hkepc.com/bbs/attachment.php?aid=497181

If you take the HKEPC configurations seriously, and compare that to what xbit reported for reasoning from AMD, then I think one of the areas you have to look at would be what role did memory configurations play in these decisions?
Hmmm, hmmm, hmmm. So the rumor at G80 launch that ATI was upset by it relates to memory configs and so BOM, rather than performance or even marketing? Good news, if true. But how does that square with the rumors that ATI was squirreling away GDDR4, possibly to NV's detriment, for R600's launch? Do they expect too much demand? Negative reaction to what may be a higher-than-G80GTX SRP? Seems like this could be easily countered with performance and availability. Surely they could launch with the inevitable GDDR4 model and follow up with more affordable GDDR3 ones without having to delay the press fest? Or is it that much more cost-effective for them to hand each of you a sample at the editor's day, rather than mail them out later? Or are they expecting so much demand of an attractively-priced GDDR3 XT &/ XL that they're stockpiling now?

So, basically, HKEPC (Geo's link seems to point to a deleted thread) echoes DT:

1GB of GDDR4 memory is the reference configuration for Radeon X2900 XTX. ...

Approximately one month later, the company will launch the GDDR3 version of the card. This card, dubbed the Radeon X2900 XT, features 512MB of GDDR3 and lower clock frequencies. ...

AMD claims the R600 target schedule will be a hard launch -- availability is expected to be immediate.

So they want a firm date, not an approximation. This still jives with Geo's sig, DT's piece, and serenity's excellent tidbit, as the launch date wasn't due to be publicized until the Ides. (Hmmm, what conxn to draw.... Will one of ATI's partners stab their PR dept in the back and leak [more] launch details? :))

I'm willing to believe the delay was to ensure a better launch, even a hard launch of this rumored and surely more desirable product. I'm just confused about this apparently undefined preview delay. Hard launches are great and all, but soft launches are fine as long as you stick to them. The problem isn't future dates, but future dates set in sand rather than silicon. Secrecy is great and all, too, but why not just come out and say why the delay, rather than leave a possible impression of disarray?

Or ATI's just not concerned about their 3D geek Q score taking a temporary, predictable, recoverable hit.
 
March 31 is the end of Q1 07, but the SEC says that a corporation is not supposed to make comments that could affect its stock price until revenues for that quarter are announced. So, even if R600 had to be pushed back two days from March 29, they wouldn't be able to announce it until the end of April. Q4 06 revenue call came on January 23rd, so that should give you a baseline.
 
March 31 is the end of Q2 07, but the SEC says that a corporation is not supposed to make comments that could affect its stock price until revenues for that quarter are announced. So, even if R600 had to be pushed back two days from March 29, they wouldn't be able to announce it until the end of April. Q4 06 revenue call came on January 23rd, so that should give you a baseline.

Thanks. At least that explains why they aren't talking about it, at least officially. Of course, they wouldn't officially talk about a product that hasn't even been officially announced yet anyways.

You do mean March 31 is the end of Q1 though, right?? ;)
 
Thanks. At least that explains why they aren't talking about it, at least officially. Of course, they wouldn't officially talk about a product that hasn't even been officially announced yet anyways.

You do mean March 31 is the end of Q1 though, right?? ;)

Calendar Q1 and Fiscal Q2? At least that was for Ati's fiscal calendar, not sure for AMD's one
 
R V 630

HKEPC has some details on RV630.

rv630yk3.png


rv630uvdcb0.png
 
Status
Not open for further replies.
Back
Top