By: Maynard Handley (name99.delete@this.name99.org), February 23, 2014 6:55 pm
Room: Moderated Discussions
Doug S (foo.delete@this.bar.bar) on February 23, 2014 12:09 pm wrote:
> Jason Lee (jasonlee.delete@this.nospam.com) on February 21, 2014 10:49 pm wrote:
> > Through silicon vias would be needed because of the wide interface between the cache chips
> > and the processor chip. Through silicon vias are said to be expensive but Intel currently
> > charges up to $2500 per processor for their dual-socket server chips (Ivy Bridge-EP) and up
> > to $6600 per processor for their quad-socket server chips (Ivy Bridge-EX). Through silicon
> > vias are not expensive compared to the price of a multi-thousand dollar processor.
>
>
> Or do something like this?
>
> http://www.thruchip.com/index.htm
>
I thought that tech looked familiar... SUN was working on this a few years ago, and guess where Ditzel comes from...
The paper I read about it (years ago) was clearly real, in the sense that they described the tech in detail, what it was and wasn't good for, and the problems they anticipated.
But it's an empirical fact that it hasn't taken off.
Why?
Once possible answer is that heat is the problem. The bare surface of a chip is so valuable for removing heat that that overwhelms the packaging advantage of putting RAM (or anything else) there. At some point the economics (and a lack of other options) may mean the industry goes this route (along with the additional steps that may be necessary, like dumping the entire assembly in pumped halon or oil to get the heat to the surface fast enough). but we don't seem to be there yet.
> Jason Lee (jasonlee.delete@this.nospam.com) on February 21, 2014 10:49 pm wrote:
> > Through silicon vias would be needed because of the wide interface between the cache chips
> > and the processor chip. Through silicon vias are said to be expensive but Intel currently
> > charges up to $2500 per processor for their dual-socket server chips (Ivy Bridge-EP) and up
> > to $6600 per processor for their quad-socket server chips (Ivy Bridge-EX). Through silicon
> > vias are not expensive compared to the price of a multi-thousand dollar processor.
>
>
> Or do something like this?
>
> http://www.thruchip.com/index.htm
>
I thought that tech looked familiar... SUN was working on this a few years ago, and guess where Ditzel comes from...
The paper I read about it (years ago) was clearly real, in the sense that they described the tech in detail, what it was and wasn't good for, and the problems they anticipated.
But it's an empirical fact that it hasn't taken off.
Why?
Once possible answer is that heat is the problem. The bare surface of a chip is so valuable for removing heat that that overwhelms the packaging advantage of putting RAM (or anything else) there. At some point the economics (and a lack of other options) may mean the industry goes this route (along with the additional steps that may be necessary, like dumping the entire assembly in pumped halon or oil to get the heat to the surface fast enough). but we don't seem to be there yet.