- Mar 27, 2009
- 12,968
- 221
- 106
Wasn't that determined to not be the real die shot?
(Reading on XS forums)
The basic layout might be correct. They might have fudged the details to add ambiguity to how big the actual die might be, but I doubt the changes are something unrecognizable.
Even having a clear picture of the die, it can be often 5-10% wrong because the measurement was off by mere few pixels. The upper and the lower cores have different dimension. Unless they plan to really do that(for example have the FPU on the upper core and make the lower core share it), the reason is probably due to editing.
The basic layout might be correct. They might have fudged the details to add ambiguity to how big the actual die might be, but I doubt the changes are something unrecognizable.
Even having a clear picture of the die, it can be often 5-10% wrong because the measurement was off by mere few pixels. The upper and the lower cores have different dimension. Unless they plan to really do that(for example have the FPU on the upper core and make the lower core share it), the reason is probably due to editing.
I just read JFAMD comments on semiaccurate and he pretty much siad the photo is very molested to not give away to much info.
I'm not an industry insider but what is sooo special about releasing a die of BD? Is this a proof that they can make it or it actually can show some design details? someone explain.
I'm not an industry insider but what is sooo special about releasing a die of BD? Is this a proof that they can make it or it actually can show some design details? someone explain.
Its not like AMD to keep this kind of stuff so close to the chest.
What exactly are they cooking.
http://www.xtremesystems.org/forums/showthread.php?t=258406&page=3
JF commented on page 2. I think AMD is amusing itself with all this free talk/advertising about the new architecture. They sure are rilling up the intel fanbois for sure.
Actually, we don't release die shots prior to launch. As for amusing ourselves, hardly. When the request came to me I originally said no, I would rather not have the die shot out ahead of launch if I can help it. But there was value in supporting our partner. All of the photoshop work was done to keep the competition from learning too much.
I doubt Intel cares what AMD is doing. Its not like Intel can't get one of AMDs cpus once they go out to partners, M/B makers. Its not like Intel will reverse engineer AMDs work .
Having a visual of a die allows some very smart people to begin estimating what they did, exactly. You can figure out how much space was dedicated to everything, figure out exactly how much bigger a bulldozer module is compared to a deneb core, and some other stuff.
I, however, just think they're pretty. So I don't worry too much about all that (I like to have definite info).
It is a confidence-building exercise.
....
we love our car analogies in these forums...car manufacturers go to great lengths to obscure/hide their prototype vehicles, and then along comes a detroit autoshow and they can't wait to show it off.
Does anyone have any guesses or theories on how large the die could be?
Would 150mm2 be out of the question?
My geuss would be that they might be around the same size as a phenom II x4 die for the 4 module 8 logical core BD.Unless they decide to use the process node to make the cpu smaller which you could scale down from current dimensions. Its gonna be close would be my geuss but maybe a bit smaller. hard to say.
I know why you did it but its been fun to watch the intel loyalists get stirred up over it. Admit it you guys are getting a great big chuckle out of all this.![]()
tI know why you did it but its been fun to watch the intel loyalists get stirred up over it. Admit it you guys are getting a great big chuckle out of all this.![]()