IBM Improves CPU Cooling

[Jimmy 2004]

New research by Big Blue could see CPUs running up to 40% cooler in the near future, without needing prices to rocket. In a paper released at the IEEE Semi-Therm conference, IBM published ideas designed to make thermal paste more efficient, therefore allowing for an improved cooling capacity leading to cooler processors. According to IBM, current thermal paste techniques can lead to 40% of heat given off by the CPU being absorbed by the particles in the paste, largely caused by the fact that the particles don't spread evenly, leading to the "Magic Cross" shown in the image below and to the left. IBM's new technology involves integrating micro-meter length trenches into the copper cap that sits above the CPU core (below on the right), which will allow thermal paste to be more evenly distributed and will lead to a third less thermal paste being required, as well as halving the pressure required to fit a CPU cooler. IBM still needs to finalise the research, but this relatively in-expensive method could be integrated into new CPUs and coolers before too long. For more detailed information you can read IBM's press release.


View at TechPowerUp Main Site

 

[Namslas90]

This kind of blows away the lapping idea!

 

[WarEagleAU]

well yeah it does, but its nice without having to do the lapping and getting the same results for less.

 

[Benpi]

Well, since nobody is buying their CPUs anymore, they only have the budget to find new ways to spread paste.

 

[HousERaT]

Sounds like good PR. I'll wait to see an independent reveiw before I start dreaming about 40% cooler running processors.

 

[Thermopylae_480]

Well, since nobody is buying their CPUs anymore, they only have the budget to find new ways to spread paste.

They helped design and manufacture the cell processor used in the PS3. The Power 6 series sounds like it could be impressive. Their processors are not where they get most of their money anyway. IBM is still a powerhouse in the tech world.

 

[Jimmy 2004]

They helped design and manufacture the cell processor used in the PS3. The Power 6 series sounds like it could be impressive. Their processors are not where they get most of their money anyway. IBM is still a powerhouse in the tech world.

And they helped make the CPU in the GameCube... and the Wii... and the Xbox 360... still a very active company - one which is benefiting every time you buy one of the major consoles (not the handhelds).

IBM is obviously huge in the server market, and always selling research and development.

 

[unsmart]

Could this be used with nanofoil and would the same concept apply to a heatsink base or is it only usable for lid attachments? I remember this being posted about four months ago when it was still in the early research phase, cool to see it getting this far Wish I could use it on my C2D right now, temps are getting a bit to high

 

[DBH]

How come this article has been posted again?? Swear it was shown end of last month / beginning of this month, or is there something different here

 

[OnBoard]

Now if you take a knife and do some crosses on top of the A64 heatspreded, will it do that and improve cooling ?)

DBH: this is a press release, the earlier one was research time news.

 

[Jimmy 2004]

How come this article has been posted again?? Swear it was shown end of last month / beginning of this month, or is there something different here

It's very possible it has and I've missed it. If you find another story like this one, post a link and I'll remove this.

Edit: you must be meaning this from October (can't find anything newer). Admittedly this is the same story but it has been developed a bit more and IBM has given more details this time, so I'll class it as an update

 

[unsmart]

Now if you take a knife and do some crosses on top of the A64 heatspreded, will it do that and improve cooling ?)

DBH: this is a press release, the earlier one was research time news.

There talking about the interface between die=heatspreader cap but it seems the concept would carry over to the heatspearder=heatsink. The only thing that I can see being an issue is human error in putting on the heatsink as apposed to the stamping approach of setting a cap. I can't ever put one on perfectly level.

 

[_SPEED]

This kind of blows away the lapping idea!

No-no! Lapping is still good. What they are talking about it the "glue" that goes between the silicon surface and the heatspreader, ie. the surface UNDER the metal cap you see on the CPU.

This is a key quote from IBM:

The image shows a cross-sectional schematic of the cooling architecture using the branched channel design. A highly viscous paste is brought between the chip cap and the hot chip in order to reduce the thermal resistance. Thanks to the tree-like branched "trenches" in the copper cap, the paste spreads very homogeneously and attains a thickness of less than 10 micrometers. With this technique, only half the pressure is needed to apply the paste and a twofold increase in cooling performance can be achieved.

Lapping the metal cap (top) and the heatsink (bottom) is still a good idea.

 

[_SPEED]

Now if you take a knife and do some crosses on top of the A64 heatspreded, will it do that and improve cooling ?)

DBH: this is a press release, the earlier one was research time news.

No-no! The IBM technique is the glue UNDER the heatspreader. Criss-cross on top of the heatspreadder is not as effective as lapping. Second, a knife will make scratches MUCH DEEPER than 10 micrometers. DONT DO IT!

 

[Zubasa]

Now if you take a knife and do some crosses on top of the A64 heatspreded, will it do that and improve cooling ?)

Go ahead and you will end up with a piece of Junk

 

[OnBoard]

Hehe, won' be doing it even if it would help, just asked And thanks SPEED for the explanation.