Intel server strategy in shambles?

[Viditor]

http://www.theinquirer.net/?article=27192

An absolutely devasting accounting of Intel's server strategy for the next 3-4 years...
And for the life of me, I can't find a flaw in their reasoning!
The apparent loss of CSI (which was to be similar to AMDs HT) does not bode well at all for the upcoming quad core chips...imagine all of that bandwidth still being to the FSB!
Even dual FSBs (eg Twincastle) does not relieve it as it doesn't help interchip communication much at all. This really won't effect the desktop much, but for things like a 2P or 4P system with quad cores, it will be a very big problem IMHO.

 

[Pabster]

Time will tell. Certainly doesn't look very good.

 

[ZobarStyl]

This news, combined with their whole new "Performance Per Watt" marketing run makes me think they haven't figured out anything as far as getting ahead of AMD in performance is concerned. I'm sure they'll own the laptop market if they bring these negligible voltage procs to market, but the server people won't care; all that matters is performance. And if Intel has no quality product for servers for ~2-3 years then people like Dell finally have to give in to the Opteron server market, and that'll be huge for AMD. No forecasts of doom here, just saying AMD may be one step closer to a fair market, should they not sit on their butts as they are known to do in these situations.

 

[Vee]

I don't think CSI is gone.
I think Intel may be reconsidering tying up so much of their future support hardware plans on compatibility with the Itanic. The purpose of cost reducing the Itanic could hurt their x86-64 side.
The Itanic has cost Intel so much, and not only the K10 but also their own future x86-64 chips may well be Itanic-killers.
...Time to cast off the lines and let it sink... Slowly, of course! Might yet sell a few to unwary customers.

Edit: Or it could be just the opposite. In which case Intel is probably doing a mistake. Haven't they learnt not to listen to the market department yet, when it comes to technology?

 

[xbdestroya]

I was going to post on this, but hey here's a topic already!

It's not just the Xeon line that's getting hamstrung though - Itanium as well, and XBitlabs goes deeper into that aspect of the server disaster:

http://www.xbitlabs.com/news/cpu/display/20051024193043.html

 

[Wingznut]

Topic Title: Intel server strategy in shambles?
Topic Summary: it is according to the Inq...

Yes, because theinquirer would know so much more about such things than Intel. :roll:

 

[Viditor]

Originally posted by: Vee
I don't think CSI is gone.
I think Intel may be reconsidering tying up so much of their future support hardware plans on compatibility with the Itanic. The purpose of cost reducing the Itanic could hurt their x86-64 side.
The Itanic has cost Intel so much, and not only the K10 but also their own future x86-64 chips may well be Itanic-killers.
...Time to cast off the lines and let it sink... Slowly, of course! Might yet sell a few to unwary customers.

Edit: Or it could be just the opposite. In which case Intel is probably doing a mistake. Haven't they learnt not to listen to the market department yet, when it comes to technology?

Good point Vee...Intel didn't really say CSI was gone, just the chip that was supposed to have it. I must say that the Inqs speculation that CSI was the cause of the cancellation (ie they couldn't get it to work, so they cut their losses now) certainly has some merit. Intel have been very quiet about CSI for quite awhile...

 

[Viditor]

Originally posted by: Wingznut

Topic Title: Intel server strategy in shambles?
Topic Summary: it is according to the Inq...

Yes, because theinquirer would know so much more about such things than Intel. :roll:

Has nothing to do with knowledge, everything to do with spin or no spin...