News Microfluidic Cooling on Overclocked Intel Core i7-8700K Drops Temperatures 44%

[Admin]

A team of researchers with Microsoft and Georgia's School of Electrical and Computer Engineering tested one of the most promising future cooling solutions on an off-the-shelf Intel Core i7-8700K. The result? 44.4% lower temperatures, on an overclocked CPU!

Microfluidic Cooling on Overclocked Intel Core i7-8700K Drops Temperatures 44% : Read more

 

[Giroro]

I'm looking at the picture at the top of the article.
Did they delid the processor by grinding through the heat spreader?
Why?

 

[jkflipflop98]

I'm looking at the picture at the top of the article.
Did they delid the processor by grinding through the heat spreader?
Why?

Well I mean, they're scientists and not engineers.

 

[Aaron44126]

"Microfluidic Cooling on Overclocked Intel Core i7-8700K Drops Temperatures 44%"

sigh
Can you low a temperature by a percentage? How do you compute this? IMO this only makes sense if you are using Kelvin as your temperature scale, and in that case the percentage drop would be much lower...

 

[InvalidError]

One major problem with micro-channel cooling built directly into silicon: you are screwed the instant your cooling loop gets contaminated by anything and ends up clogging the channels.

That sort of stuff sounds like it would only be workable for HPC and other similar environments where you can afford to have an engineer or other specialized resource monitoring coolant loop chemistry on a daily basis to catch deviations before they cause problems.

Did they delid the processor by grinding through the heat spreader?
Why?

Why not? Delidding by breaking the adhesive requires special tools to apply carefully measured force, delidding with a grinder only requires steady hands or a 3D-printed fixture.

Can you low a temperature by a percentage? How do you compute this?

The most logical way to compute temperature drop as a percentage IMO would be to use room temperature as 100% "efficiency" and standard liquid cooling as the reference for 0%. If room temperature is 20C and core temperature with conventional liquid cooling hit 90C, then "44% better" would be 44% of the way down towards 20C from 90C, which would be 61C.

This way, no matter if you use Kelvin, Celsius, Fahrenheit or whatever other linear temperature scale, the outcome is the same.

 

[TJ Hooker]

"Microfluidic Cooling on Overclocked Intel Core i7-8700K Drops Temperatures 44%"

sigh
Can you low a temperature by a percentage? How do you compute this? IMO this only makes sense if you are using Kelvin as your temperature scale, and in that case the percentage drop would be much lower...

Looking at the original IEEE article, it seems the value that actually decreased by 44% is the "junction-to-inlet thermal resistance when compared to a conventional cold-plate operating with the same inlet temperature of 34°C". Not the operating temperature as the TH article states.

But I agree that referring to changes in temp using % doesn't usually make a lot of sense.

 

[Kamen Rider Blade]

Why can't be just use "- ##° C" improvement over Ambient?

 

[funguseater]

Is this really that great in practice though, perhaps they should have chosen a cpu that will perform better with this type of cooling, not just cooler at the same performance level? My what I consider standard 8086K hits 5.2 all core 24/7. And I use an air cooler.

Also all I needed to delid was a vice and a hammer, so no special tools required. 😉

 

[InvalidError]

Is this really that great in practice though, perhaps they should have chosen a cpu that will perform better with this type of cooling, not just cooler at the same performance level?

Due to how the coolant needs to remain absolutely pristine to prevent clogging of micro-channels etched into the die's passivation layer, I think the cooling technique would only be applicable to datacenters and supercomputers to reduce the amount of power needed for cooling: if you can almost eliminate the junction-to-coolant thermal impedance, then you can work with 40-50C coolant in summer instead of needing chilled loops.

 

[ccl13]

"Microfluidic Cooling on Overclocked Intel Core i7-8700K Drops Temperatures 44%"

sigh
Can you low a temperature by a percentage? How do you compute this? IMO this only makes sense if you are using Kelvin as your temperature scale, and in that case the percentage drop would be much lower...

The most logical way to compute temperature drop as a percentage IMO would be to use room temperature as 100% "efficiency" and standard liquid cooling as the reference for 0%. If room temperature is 20C and core temperature with conventional liquid cooling hit 90C, then "44% better" would be 44% of the way down towards 20C from 90C, which would be 61C.

This way, no matter if you use Kelvin, Celsius, Fahrenheit or whatever other linear temperature scale, the outcome is the same.

Yeah I was going to say the same. "Drop temperatures 44%" only really works when we talk about Kelvin, and anything in CPU temperature range with 44% decrease would make lot of dry ice around the CPU.

But if that's in deltaT it makes a lot more sense. But that is not "drops temperature", it is "drops temperature increase".

Anyways, the title is misleading and unprofessional.

 

[JamesJones44]

The big question is the commercial viability of such a solution? From the sounds of it, this approach doesn't seem all that viable for typical use cases. So while fun to think about, it's hard to get excited about such a solution.

 

[edzieba]

Reminds me of Aquasar, though I think there they etched the IHS rather than the die.

I'm looking at the picture at the top of the article.
Did they delid the processor by grinding through the heat spreader?
Why?

It appears they removed the IHS for the spin-coat process (fig.4 shows the delidded die after etching, and there is clearly no IHS present), then cut a hole[1] in it, and stuck it back on when they potted the cooling stack in order to act as a dam for the potting compound.

[1] With tinsips, or a particularly blunt spoon.

 

[Giroro]

Why not? Delidding by breaking the adhesive requires special tools to apply carefully measured force, delidding with a grinder only requires steady hands or a 3D-printed fixture.

Those special tools of course being a small inexpensive vise and Allen key, which can also be 3D printed. The "why not" being the near-certainty you destroy the CPU when it turns out your hands are less than steady.

Reminds me of Aquasar, though I think there they etched the IHS rather than the die.
It appears they removed the IHS for the spin-coat process (fig.4 shows the delidded die after etching, and there is clearly no IHS present), then cut a hole[1] in it, and stuck it back on when they potted the cooling stack in order to act as a dam for the potting compound.

[1] With tinsips, or a particularly blunt spoon.

It makes much more sense that they would cut the IHS after removing it. But it's weird that people who can etch 10um micro channels in silicon wouldn't be able to cut a better hole in a small chunk of copper, or find an easier way to make a dam.

 

[Eximo]

IBM pretty much has production versions of this for supercomputers they made many years ago.

They still have the coolest and strangest prototype. Power via fluid:

IBM is trying to solve all of computing’s scaling issues with 5D electronic blood

Animals use blood for cooling and power delivery. Why can’t computers, too?

arstechnica.com

 

[ccoonansr]

This was a late April Fool's piece? Yes? My delidded 8700k runs 5.2 ghz on air (Cryorig H7). It's 100% stable. Been daily driving this since last year. What am I missing ?

 

[TJ Hooker]

The title directly quotes the researcher's announced percentages in their paper, which was published in a peer-reviewed scientific publication.

The value of 44% from the paper referred to a decrease in thermal resistance though, not a decrease in temperature.

Edit: more specifically, thermal resistance from junction to inlet.

 

[Co BIY]

How much of the benefit could be had by just thinning the silicon down to the depth of the channels and soldering the heat spreader closer to the heat producing elements ?