Anand has the details about r520,rv530,rv515

ATI will have X1800XL cards at the time of its announcements.

That? Okay. My comment on yields was aimed at R520XT (i.e. speed-binning). Tho I'm a bit confuddled why Trini would find it unusual that XTs are "hand-picked XLs". Wouldn't one say the same about the relationship of 6800GT to 6800Ultra and a whole host of other examples?

Unless, of course, you're suggesting they really are different spins and that's why there is a difference in availability dates. . . but I really hope not.

Edit: I suppose the other other explanation would be a desire to keep XT clocks secret as late as possible.
 
Last edited by a moderator:
Speed-binning has been a pretty standard way to get the "Ultra" "FX" "EE" or "XT" versions of chips for a very long time now.
 
geo said:
That? Okay. My comment on yields was aimed at R520XT (i.e. speed-binning). Tho I'm a bit confuddled why Trini would find it unusual that XTs are "hand-picked XLs". Wouldn't one say the same about the relationship of 6800GT to 6800Ultra and a whole host of other examples?

Unless, of course, you're suggesting they really are different spins and that's why there is a difference in availability dates. . . but I really hope not.

Edit: I suppose the other other explanation would be a desire to keep XT clocks secret as late as possible.

I'd guess it is this from Fudo:

Those cards are clocked at 500MHz core and 1000MHz memory and should be available in quantities at launch in first week of October.

One of the key things that you must bear in mind is that R520XL, Radeon X1800XL can not be overclocked well. Voltage leakage will prevent you to increase voltage and that will limit your overclockability. This voltage leakage was the key reason for such of delay of R520. Only the flagship Radeon X1800XT will overclock like crazy. So even if some board partners tried to overclock some of the Radeon X1800 XL cards it simply would not work as you could never even touch those XT speeds of 600/1400MHz, so ATI and others claim.


which would indicate that they will initally be different revisions, one which leaks and cannot go much faster than 500mhz and the other one that does not... I think in later post he was mentioning more than 600 up to 700?!?
 
Druga Runda said:
which would indicate that they will initally be different revisions, one which leaks and cannot go much faster than 500mhz and the other one that does not... I think in later post he was mentioning more than 600 up to 700?!?

:sigh: Yeah, it does suggest that. Man, I hope not. I know I wouldn't buy one. I'd buy an XT, or sit tight to wait for the new rev to move down to XL. And if they have any sense at all they better make sure that is clear in the reviews, or you can guarantee screams later.
 
geo said:
That? Okay. My comment on yields was aimed at R520XT (i.e. speed-binning). Tho I'm a bit confuddled why Trini would find it unusual that XTs are "hand-picked XLs". Wouldn't one say the same about the relationship of 6800GT to 6800Ultra and a whole host of other examples?

No that's not the part I'm confused about. The Ultra came out before the GT remember....

So if an XT is just a speed-binned XL, why do they need to stagger the launches - I mean, if things have been that bad so far, how many more XT's could they possibly get in a couple extra weeks? This is assuming that the chips are completely identical save for clocks.

I'm also wondering if the "speed-binning" could be more like "leak-binning" instead.

And I thought Dave was referring to this:

Inquirer said:
One of the key things that you must bear in mind is that R520XL, Radeon X1800XL can not be overclocked well. Voltage leakage will prevent you to increase voltage and that will limit your overclockability. This voltage leakage was the key reason for such of delay of R520. Only the flagship Radeon X1800XT will overclock like crazy. So even if some board partners tried to overclock some of the Radeon X1800 XL cards it simply would not work as you could never even touch those XT speeds of 600/1400MHz, so ATI and others claim.
 
:???: I thought the XL is supposed to be 550MHz core?

If it's 500, what's the Pro (which is supposed to be 500)?

Ah well, whatever.

Jawed
 
trinibwoy said:
No that's not the part I'm confused about. The Ultra came out before the GT remember....
6800U was pretty hard to buy back at launch - arguably worse than X800XTPE. Particularly as it was launched weeks before ATI responded.

Jawed
 
Jawed said:
6800U was pretty hard to buy back at launch - arguably worse than X800XTPE. Particularly as it was launched weeks before ATI responded.

Jawed

Ah, availability....yeah that could explain it. Still not sure what a couple weeks gives you though in terms of production capability.

And regarding the above Inquirer quote, why are they referring to the XL and the XT as two separate chips?
 
yep, I guess it depends on how fast they sell out, so give it a few months if you can and XL should become overclockable like the big brother.
 
Druga Runda said:
yep, I guess it depends on how fast they sell out, so give it a few months if you can and XL should become overclockable like the big brother.

Right. But what I'm worried about is the temptation to not tell anyone they are different revs of the chip. That would surely bite them on the ass and cause a great many returns and ill will as it becomes clear --and it would (it always does).
 
Druga Runda said:
yep, I guess it depends on how fast they sell out, so give it a few months if you can and XL should become overclockable like the big brother.

I'm thinking along that. I think the two weeks are the time needed for ATi to get plenty of the non-leaking X1800 made. The X1800's coming out before then are just the ones with the issues still. So with that in mind, it may be smart to hold off about a month in buying an X1800 and get one of those super overclocking bad boys. Of course, this is all total speculation on my part....

geo said:
Right. But what I'm worried about is the temptation to not tell anyone they are different revs of the chip. That would surely bite them on the ass and cause a great many returns and ill will as it becomes clear --and it would (it always does).

Returns on what grounds? The product is only rated for default speeds, nothing more.
 
Skrying said:
Returns on what grounds? The product is only rated for default speeds, nothing more.

Any made-up grounds necessary (umm, err, I get wavey lines on my screen --gimmee a credit!) if the customer didn't know in advance. These are pretty savvy folks at this level of the customer base, and I've hung around forums long enuf to feel confident it will happen. . .if it's a surprise discovery later on. I can pluck three or four examples off the top of my head over the years. . .
 
trinibwoy said:
Ah, availability....yeah that could explain it. Still not sure what a couple weeks gives you though in terms of production capability.

And regarding the above Inquirer quote, why are they referring to the XL and the XT as two separate chips?
The way I see it, ATI may have been using slow R520s (incapable of XT speed) to build up a launch inventory for LE, Pro and XL - so that the shelves will be literally heaving with them.

Bearing in mind that RV515 and RV530 are both slightly delayed too, I imagine X1800XT and all the low- and mid-range cards had to go through a final re-spin, which was discovered in the May-July window.

So XTs in reviewers hands at launch will be cherry picked XLs and essentially unavailable. Then after about the next 4-6 weeks from launch you'll finally find them appearing in stores, once the real XTs come rolling off the line.

That's my conspiracy theory, anyway.

An alternative way of seeing it is that ATI needs to pile up tens of millions of X1300s and millions of X1600s, so side-lined X1800XT for a bit, knowing that it's penetration is in the 1-3% region.

Jawed
 
trinibwoy said:
I'm also wondering if the "speed-binning" could be more like "leak-binning" instead.

And I thought Dave was referring to this:


Originally Posted by Inquirer
One of the key things that you must bear in mind is that R520XL, Radeon X1800XL can not be overclocked well. Voltage leakage will prevent you to increase voltage and that will limit your overclockability. This voltage leakage was the key reason for such of delay of R520. Only the flagship Radeon X1800XT will overclock like crazy. So even if some board partners tried to overclock some of the Radeon X1800 XL cards it simply would not work as you could never even touch those XT speeds of 600/1400MHz, so ATI and others claim.

This would imply that ATI made a load of early revision chips that are leaky and have design problems fixed in later revisions - enough to supply a whole couple of SKU's. I find it hard to believe given the delays and respins we've seen that ATI went ahead and made millions of known problematic early revision chips months back in the hope they could make them work as a low-performing XL product, while still chasing down the problem.
 
Bouncing Zabaglione Bros. said:
This would imply that ATI made a load of early revision chips that are leaky and have design problems fixed in later revisions - enough to supply a whole couple of SKU's. I find it hard to believe given the delays and respins we've seen that ATI went ahead and made millions of known problematic early revision chips months back in the hope they could make them work as a low-performing XL product, while still chasing down the problem.

Well I don't know how it works but I thought the IHV's only started ramping production after all the kinks had been worked out. So I'm not sure how ATi could have enough "bad" chips to fill the early XL/LE demand.
 
Well, I think we need a couple graybeard R3Ders over here to explain why this is going to cause problems, if that's where they are going. Not massive ones, certainly, but some, that their retailers and partners will not appreciate.
 
Back
Top