A look at X800 and 6800 UK Availability

Here is also a list of different Pro ViVo's and their succes-rates. As you can see, none has failed. Though to be fair, there has been only 13 tested cards.

I've Sapphire's X800Pro ViVo and it required only a bios update to get the last pipes working. But, like already mentioned in this thread, getting the clocks to XT PE-levels is hard. My card is stable at 510/580 but core at 520 hangs the computer when running 3DMark. It seems that the reference cooler isn't that effective.
 
This is a nice bonus for Ati users, my assumption is that they would have bought the X800 anyway, so this is just the icing on the cake, rather than would have got the XT, but risked the X800 and got the bonus, if this was the case then ATi are losing some money.

Would it not be possible for ATi to have a lower clocked 16 pipe XT at a price somewhere between the X800Pro and X800XT thus getting the most out of these pseudo 16 pipe cards ? This would then give the GT a good run for it's money ( to put it mildly )
 
jvd said:
Its what a 33% increase in fillrate for the 4 extra pipes ?

If you can only get 500mhz with 16 pipes thats 8000 megapixels.

You would need to clock it to 570ish mhz with 14 pipelines to get the same performance . of 8000mega pixels So of course the 16 pipes would be the best bet.

I follow your math, but I thought it's 12 pipes in the pro. :) Then it would need a devilish core speed of 666 MHz to compete 16 pipes.
 
No the yields of the 16 pipeline cards are so high they are rebranding them as 12 pipeline cards to keep up with the demand of the x800pro and the fact that there isn't enough ram for x800xt pe speeds
 
jvd said:
No the yields of the 16 pipeline cards are so high they are rebranding them as 12 pipeline cards to keep up with the demand of the x800pro and the fact that there isn't enough ram for x800xt pe speeds

I think we have a winner....
 
jvd said:
digitalwanderer said:
I guess what I was asking was which would give you the better performance increase if you had to choose; higher speeds or the 4 more pipes?

It depends .

Its what a 33% increase in fillrate for the 4 extra pipes ?

If you can only get 500mhz with 16 pipes thats 8000 megapixels.

You would need to clock it to 570ish mhz with 14 pipelines to get the same performance . of 8000mega pixels So of course the 16 pipes would be the best bet.
Thanks JVD. :)

DaveBaumann said:
Put the pieces together...
I must not have had enough coffee yet, I think I actually understand your hint! :D
 
I think we have a winner....

Except some (perhaps all that softmod?) have the 1.6ns RAM. In the UK some (club3d) even had "XT platinum edition" on the heatsink.

They seem to have decided *after* making the cards that they will only run at Pro speeds!!!

EDIT: Here is a linky. Scroll down and look at that pic of an XT Pro!
 
Right. Only thing is is that 1.6ns ram rated to run at 560 and is ati getting a warrenty on the ram from the manufacturer (or is club3d getting it)

That is the whole thing

1.6ns ram is rated over diffrent mhz at diffrent voltages.
 
Here are the "pieces" as I see it:

There are suddenly a lot of VIVO XT pros on the market without hardware locked 12 pipes.

These "unlocked" pros have expensive 1.6ns RAM on them.

Some have shipped with XTPE on the heatsink.

These chips are bios locked to 12 pipes but left ATI as non locked chips (no laser cut bridges)

A lack of XTPE boards available.

Is it not to much to conclude that these chips left ATI as PEs and only after AIB partners had started making them into cards did ATI phone up and say "sorry lads but that batch of chips we sent you last week does not pass all our quality tests for running at XTPE speeds, if you can just lock 4 pipes in the bios and sell them as Pros, thanks"

Hence no *real* XTPEs at this time.

If this is not on the Inq Tomorrow, I will be dissapointed.

Don't flame me its just a theory. BTW I've had an XTPE on order since the 14th May, so I was pretty upset to see XTPE cards on the market as XT pros when there are no XTPEs in the UK.
 
Canada would be a good answer since it's somewhat colder there than the rest of the world, but even I would doubt that.
 
Hey Dave, do you happen to know at what point in the process ATi does there speed binning?

A+ for effort :)

My assumption was that they were speed binned incorrectly and the mistake realised later (this is WRT my "wild theory") sort of like the Intel P3 1.13. Where they were binned is irrelevent for this theory, only that a mistake was made. Which only makes Dave's post even more cryptic. Is he saying they are binned at place A, there is no way that sort of mistake could happen. Or is he saying they are binned at place B so a mistake could have happened?

Of course the non-threaded nature of the boards only helps Dave's vagueness, maybe he was responding to another post? :eek:
 
Back
Top