Replies to post #24148 on Advanced Micro Devices Inc (AMD)

[chipguy]

Itanium - I'm sure their yield is good for a 421mm2 die.


Perhaps you didn't understand my comment about critical
area. It is entirely in the realm of the possibility that the
421 mm2 McKinley could yield better (i.e. higher percentage
of functional die after repair) than the 217 mm2 Willamette
made in the same process. Look at microphotographs or
die plots of both devices and the reason for this is staring
you right in the face.

That is not to claim that McKinley yielded better in practice -
the Willamette was made in >1000x higher quantities and that
has an impact of its own. But to say that I2 has manufacturab-
ility issues because of its size is simply not true.

[subzero]

"Prescott: I am sure that the 3rd revision of the core will be manufacturable"

So....how many core redesigns will AMD go through with their Opteron on 90 nM before AMD can make it manufacturable?
Ever wonder why the AMD brethren were so heartened that AMD demoed 90 nM Opterons late in 2003 - but has continually delayed the introdcution of them until the second half of 2004?
3 more core revisions?
4 more core revisions?
..................?

[confused]

Sgolds

The vast majority of the time your posts show a very balanced view. This response ignores several things or takes allegations that have no solid basis and acts as if they are true.

RDRAM was a superior technology to DDR and it was the failure to get an honest committment from the ram manufacturers that created to cost problem. They despised having to pay a royalty to Rambus. Leaving aside the issue of Rambus as a company the royalty was tiny, on the order of 2 to 3 % but it appears that lately the courts have been deciding in Rambus favor and the dram manufacturers may well have to pay some very large compensation to Rambus. I am not a Rambus fan, in fact I have no real interest in them one way or the other, but the legal consensus gives the appearance that the dram manufacturers coluded (spelling?) against the success of RDRAM.

The issue is Itanium's die size is a non starter and is only mentioned by people with a bias against Intel. It is a non issue .

I generally learn from reading your posts but this sounds like a party line from the crowd that cannot get it straight that AMD can do a good job without it meaning the Intel is screwing up.