With all the 1000x time speedups that IBM releases I'm amazed that we have to wait for computers at all, no matter what the task.
I really wished that these announcements would be at least accompanied by working prototypes showing the main claims in a realistic setting.
Since these are all things that may (or may not) be realized in the distant future when those reading the article will have long forgotten about it or will be safely deceased I suspect they are more of a marketing ploy than a true announcement of a breakthrough.
Generally I agree. This one caught my eye because it addresses such a basic problem that it seems more realistic to expect this one to actually make it from theory to practice.
Yes, but the actual release states that they plan to jointly develop this (3M and IBM), not that they even have this adhesive, let alone the tech to stack the chips and do the interconnects.
Really, this is an announcement of the development of vapourware, it doesn't come much more tenous than that.
IBM seems to be pretty desperate for the spotlight lately.
Basically the bogosity here is that while its cool that the glue can act as a heat spreader, the heat still has to go somewhere. This has been a bigger hindrance on 'stacked' silicon. How do you pull 250 - 500 watts out of a chip efficiently. Its a very hard problem. If you have good chops on how to do this you can make a mid six figure income designing power stages for fighter jet radar units.
I get the feeling that researchers and manufacturers always wait until we're close to falling short of Moore's Law to come out with advances like this. I guess they can push quantum computing back a few more years now.
While I am as dubious as the rest of you about how likely we are to actually see this, I appreciated learning that adhesives are one of 3M's 46 core technology platforms. Does anyone know what the other 45 are?
6 comments
[ 3.0 ms ] story [ 23.7 ms ] threadI really wished that these announcements would be at least accompanied by working prototypes showing the main claims in a realistic setting.
Since these are all things that may (or may not) be realized in the distant future when those reading the article will have long forgotten about it or will be safely deceased I suspect they are more of a marketing ploy than a true announcement of a breakthrough.
Really, this is an announcement of the development of vapourware, it doesn't come much more tenous than that.
IBM seems to be pretty desperate for the spotlight lately.
(which is now dead)
Basically the bogosity here is that while its cool that the glue can act as a heat spreader, the heat still has to go somewhere. This has been a bigger hindrance on 'stacked' silicon. How do you pull 250 - 500 watts out of a chip efficiently. Its a very hard problem. If you have good chops on how to do this you can make a mid six figure income designing power stages for fighter jet radar units.