If You Can’t Beat Them, Join Them
It’s pretty hard to contest that AMD’s Hammer are better than Intel’s offerings.
There’s basically two reasons for that. It would not be too far off to say that today’s Hammers are yesterday’s XPs built with SOI and having an integrated memory controller.
Most of the improvements Hammers have over the older XPs is due to these two factors, and of the two, the integrated memory controller is probably the bigger contributor.
Indeed, the integrated memory controller in Hammers is probably the biggest CPU performance booster we’ve seen in the last five years.
Some of these articles indicate some doubt as to whether this integrated memory controller will include desktop chips, but given how Intel designs chips and the need for any boost they can get for the desktop, I think it’s safe to say we’ll see it there, too, around the same time: 2007.
Common Interface for IA-32/IA-64?
The announcement about integrated memory controllers also contained word that Intel plans to have a common interface for both children of PIV and children of Itanium 2.
This isn’t really news, Intel mentioned it a while back, but what may strike people as being news is that Itanium isn’t dead.
No, it isn’t; it’s just sleeping. 🙂
Seriously, Intel never declared Itanium dead, and is still developing away. You just don’t hear about it because it’s been put in the high-end niche for a while.
Even when the prospects for Itanium looked brighter, Intel didn’t plan to put it on the desktop until 2007.
I’m not going to even venture to guess what may happen in 2007, but if a mainstream-sized IA-64 can wallop an IA-32 descendant, I think we’re going to see one there.
In an age of multicores, I wouldn’t be at all surprised to see hybrid chips capable of handling both, and I don’t think AMD has any licensing rights to the Intel IA-64 standard.
I suspect such an effort would come later than 2007, likely when 45nm process technology becomes available, but I wouldn’t put it past Intel to push Itanium at that time.