With the progress and pure performance of HTT3.0 I wouldnt be suprised to see them give up on CSI and use it instead. Most recent roadmaps show 2008-2009 timeframe.
HT3.0 is up to 5.2GT/s and so on a standard 16-bit processor link that would yield 10.4GB/s. CSI is up to 6.4GT/s and on a 16-bit link would provide 12.8GB/s. Not all the features of CSI are known yet, so I don't know if it has valuable options like unganging, but CSI is very competitive bandwidth-wise. Latency is still unknown too. CSI is scheduled for release with Tukwila in 2008.
With the progress and pure performance of HTT3.0 I wouldnt be suprised to see them give up on CSI and use it instead. Most recent roadmaps show 2008-2009 timeframe.
You are right on this one, but the fact is that intel is too stubborn and arrogant to even accept that AMD beat them in the interconnect race.
They should really consider adopting HTT and give up the whole CSI fiasco since many enterprises like SUN, CRAY, HP and IBM (DELL will join sooner than later) are supporting AMD's initiative.
My opinion? You'd be hard pushed to find someone who's in the market for an Itanium (could just leave it there) who would think 'ahh, I'll get a Xeon first then save up/expand to Itanium'. Being able to use Xeon and Itanium on the same board is a non-event. Who would want that?
You are about to answer a thread that has been inactive for more than 6 months. If you still wish to proceed, please ensure that your posting is original and does not duplicate or overlap any prior responses to this thread.