Does Water Cooling your PC Also Cool Down Your Room? - The Workshop

[LinusTech]

Amazon: http://geni.us/BvX
NCIX: http://bit.ly/1RNNgXu

 

We water cool the CPU and video card of a computer to determine if that'll make your gaming environment more comfortable.. FOR SCIENCE!

 

 

 

[DominicNikon]

First time I'm early lol. (1 am in the morning)

[comicsansms]

God damn it. it is only has a 1080p option. I guess I'll wait.

[NumLock21]

WooHoo, I'm number tree!

[DND]

Holy mother of why 1080p option lowest  it would take me 3years to buff this.

[Orangeator]

Of course it doesn't, the heat has to go somewhere. (saying this before watching video)

[Hunched]

59 minutes ago, Orangeator said:

Of course it doesn't, the heat has to go somewhere. (saying this before watching video)

Yep.

I don't who this video is for, I doubt anyone smart enough to build a PC would be dumb enough to wonder about this.

This is like Kindergarten level science 

 

[Zodiark1593]

Perhaps Linus was irritated at rookie builders asking this particular question, and in his usual opportunistic fashion, made a video to direct them toward.

 

Or he's just trolling now. That avatar makes me lean toward the latter.

[TheWoollyMammoth]

I'm loving these workshop videos.

 

"Does x mean y in a situation z?"

 

-"No."

[Progressor]

I disliked the video. Not because Linus did something wrong or the video's quality, but because there are enough "people" that think this is how physics work, so much that Linus had to make a video to address it(this or the team is getting really out of ideas). Yes, I'm talking about you 10-20% idiots who voted with yes in the strawpoll. If you're an adult and you honestly voted with yes, this is all I got for you.

[Guest]

It took 166 years, but today before us we at last have the first law of thermodynamics finally being shown to be true in a youtube video. Rudolph Clausius can finally rest easy.

[PianoPlayer88Key]

I want to know how water vs air cooling computer parts affects things when the ambient temperature in the room is considerably higher.  For example ...

 

I'm in Southern California - more specifically, several miles east of San Diego, kind-of on the edge of the desert / mountains, I think.  Our house does not have A/C.

 

In summertime, I frequently see temperatures around 28-30°C (82-86°F), and at least a few times a year I've seen 32°C (~90°F) in the house.  Outside temperatures are regularly in the mid 30s C, or mid-high 90s F, and often will reach 38-40°C (100-104°F) or even a bit higher, usually at least a few times a year.  (We also have warm spells in winter sometimes - a couple weeks ago it was around 33°C (92°F) outside for a few days.)

Also, running Prime95 28.7 Small FFT on my Hyper 212 EVO-cooled i7-4790K this past summer one time, my CPU temp shot up almost immediately to 100°C (212°F) and throttled to 3.7 GHz.  (Running it in winter it still hits 100°C, but at 4.2 GHz with the occasional dip to 4.1 GHz.)  Running v26.6 (recommended by someone on Tom's Hardware because the newer ones stress the FPU too hard) yields temps more in the 80s C.

 

I remember several years ago we had one particularly hot summer.  The outside temperature, as I recall, was as high as 47°C (117°F) at least one day.  I don't remember what the inside temperature was, but considering how uncomfortable I remember it being (high humidity also, which is unusual for SoCal), I wouldn't be surprised if it was north of 38-40°C (100-104°F) in the house!   (I'm glad we didn't each in our family have PCs with an FX-9590 and two R9 295X2s then!)

 

How would water vs air cooling affect room temps when it's upwards of 100 Fahrenheit, or near 40 Celcius, ambient temperature in the room before factoring in the heat the PC is putting out?  Also what CPU (or GPU) cooling would you use in that climate to keep your CPU down to about 50-55°C (122-131°F, or cool enough to not burn your hand if you could theoretically touch the CPU) while running Prime95 28.7 Small FFT?   Also how would a place like Arizona (where my dad spent a couple summers in the late 1960s, and said there were many nights it didn't get below 100°F/38°C) have the temps affected with or by the PC?

[TheWoollyMammoth]

1 hour ago, PianoPlayer88Key said:

-snip-

 

How would water vs air cooling affect room temps when it's upwards of 100 Fahrenheit, or near 40 Celcius, ambient temperature in the room before factoring in the heat the PC is putting out?  Also what CPU (or GPU) cooling would you use in that climate to keep your CPU down to about 50-55°C (122-131°F, or cool enough to not burn your hand if you could theoretically touch the CPU) while running Prime95 28.7 Small FFT?   Also how would a place like Arizona (where my dad spent a couple summers in the late 1960s, and said there were many nights it didn't get below 100°F/38°C) have the temps affected with or by the PC?

Ambient temperature has no effect whatsoever for the need of heat dissipation on the hardware. So, it doesn't matter whether it's air or watercooled. The cpu or gpu would simply run cooler on watercooling because of the efficiency of the cooler. Heat would dissipate faster, but the amount of heat dissipated would be excactly the same.

[Modzy]

Linus I really hope you revisit this one day, use a real "room" or bedroom/office and use full custom water cooling. 

 

Your results do not match my real world results at all. My air vs water room Temps are hugely different. On air in an 2500@5.2 ghz/gtx 970 sli rig, I'd hit 30+ easy with door closed gaming after only 2hrs. When I switched to a full custom water setup with 2 360 rads and a 280 rad, high pressure pump and high speed fans, the rooms ambient stabilized at 26c. Based on pre testing ambient of 22c. 

 

I think using a shoe box room plus an AIO and an open test bench case was just a flawed method of testing. 

 

Larger normal rooms have much greater surface area with larger walls, ceiling and a lot more objects in the room to keep the ambient consistent.

 

 

[TheWoollyMammoth]

5 minutes ago, Modzy said:

Linus I really hope you revisit this one day, use a real "room" or bedroom/office and use full custom water cooling. 

 

Your results do not match my real world results at all. My air vs water room Temps are hugely different. On air in an 2500@5.2 ghz/gtx 970 sli rig, I'd hit 30+ easy with door closed gaming after only 2hrs. When I switched to a full custom water setup with 2 360 rads and a 280 rad, high pressure pump and high speed fans, the rooms ambient stabilized at 26c. Based on pre testing ambient of 22c. 

 

I think using a shoe box room plus an AIO and an open test bench case was just a flawed method of testing. 

 

Larger normal rooms have much greater surface area with larger walls, ceiling and a lot more objects in the room to keep the ambient consistent.

 

 

I believe there must be a variable here that we don't know about. If you generate x amount of heat, that's how much you generate heat. So does your watercooling system make your system more energy efficient? I doubt it.

[MSTRBKR]

Linus, why don't you buy jeans (or "pants") that fit you better?

[Acurus]

Are we being trolled? Why even make a video about this? That is just stupid. And now with the pizza heating pc?

Have LLT started April fools early this year?

[Counter-Strike Player]

 

Much funny. @LinusTech, please tell us this was a backup video made back in January.

 

6:10: "So there you have it guys, if the objective is to cool down the room then the only sane way to achieve that would be watercool all the computers in the room and put a radiator on the roof of your building. /s." Or, just get a free window AC unit from Fractal Josh.

[SolarNova]

My initial reaction as I started watching.

 

"Duh, of course it doesn't"

"What kind of halfwit thinks it would"

 

I voted on the poll and.....at this time 17% of people are ether halfwits and/or trolls  ....go figure.

[SaperPL]

6 hours ago, Hunched said:

Yep.

I don't who this video is for, I doubt anyone smart enough to build a PC would be dumb enough to wonder about this.

This is like Kindergarten level science 

 

It's actually more complex than you think, I'll explain it below

1 hour ago, SolarNova said:

My initial reaction as I started watching.

 

"Duh, of course it doesn't"

"What kind of halfwit thinks it would"

 

I voted on the poll and.....at this time 17% of people are ether halfwits and/or trolls  ....go figure.

I actually voted for yes and I am NOT wrong. The question Linus asked is missing the word SIGNIFICANTLY and also that it's not the water cooling that will cool the room but if the choice will affect final temperature. Note that in water cooling scenario temperature delta was 0.3 degree smaller than in initial scenario.

 

Explanation is quite complex and is based on something more than kindergarten science . The power draw of a chip is actually dependant on its temperature and the temperature is obviously an effect of how much it draws the power:

 

Source

 

This dependence is a feedback(closed loop?) but for current chips in normal conditions the effect is so small that it's simply countered by small throttling of the cpu, however overclocking may affect this countermeasure. At the same time you have turbo which means higher clocks when the temperature is low.

 

What I think Linus should do if he ever wanted to precisely check this out is to run the water cooling with a huge reserve tank and somehow lock performance of components to see if there is a measurable difference between the two scenarios.

 

Finally I wonder about the benchmark results for those two - if room temperatures were the same but gpu temperatures were significantly lower on water cooling then gpu might have been running on higher turbo clocks.

[TheWoollyMammoth]

53 minutes ago, SaperPL said:

It's actually more complex than you think, I'll explain it below

I actually voted for yes and I am NOT wrong. The question Linus asked is missing the word SIGNIFICANTLY and also that it's not the water cooling that will cool the room but if the choice will affect final temperature. Note that in water cooling scenario temperature delta was 0.3 degree smaller than in initial scenario.

 

Explanation is quite complex and is based on something more than kindergarten science . The power draw of a chip is actually dependant on its temperature and the temperature is obviously an effect of how much it draws the power:

 

Source

 

This dependence is a feedback(closed loop?) but for current chips in normal conditions the effect is so small that it's simply countered by small throttling of the cpu, however overclocking may affect this countermeasure. At the same time you have turbo which means higher clocks when the temperature is low.

 

What I think Linus should do if he ever wanted to precisely check this out is to run the water cooling with a huge reserve tank and somehow lock performance of components to see if there is a measurable difference between the two scenarios.

 

Finally I wonder about the benchmark results for those two - if room temperatures were the same but gpu temperatures were significantly lower on water cooling then gpu might have been running on higher turbo clocks.

You didn't factor in the higher power consumption of the water cooler,  which generates more heat than saves at the heat source. 

[LeapFrogMasterRace]

47 minutes ago, SaperPL said:

It's actually more complex than you think, I'll explain it below

I actually voted for yes and I am NOT wrong. The question Linus asked is missing the word SIGNIFICANTLY and also that it's not the water cooling that will cool the room but if the choice will affect final temperature. Note that in water cooling scenario temperature delta was 0.3 degree smaller than in initial scenario.

 

This not a controlled experiment and that is well within the margin of error and those cpus aren't even clocked the same in the graph they are using more watt energy because of that P=IV.

[lagittaja]

1 hour ago, LeapFrogMasterRace said:

This not a controlled experiment and that is well within the margin of error and those cpus aren't even clocked the same in the graph they are using more watt energy because of that P=IV.

Did you actually understand what those two graphs show?

Or did you even bother reading the article behind the link or in fact, actually just understand any of it? Since it seems like you did not.

 

~10% higher total system power consumption is "margin of error"?

How about you take a minute to comprehend what that means when you remove the motherboard/RAM/GPU/etc. from the equation and only look at dynamic CPU power...

In case your brain can't comprehend it or you're too lazy to actually read the article, the systems slice of the power consumption was ~130W and so with the 2600K running at 2Ghz@1.290V it means the dynamic CPU power went from 69W to 92W. That's a 33% increase.

Or for the 3Ghz@1.491V test it's 140W>165W = +17.8%.

The difference with the higher clockspeed+voltage test isn't as big because for that load, the NH-D14 passive cooled by the floor fan blowing air on it could not get the CPU temp lower than 67C vs the 47C CPU temp for the 2Ghz@1.290V load.

 

@LinusTech This needs a follow up Workshop episode where you take it to the extreme: aircooling with extremely high temperatures vs watercooling with as low temperatures as you can muster along with fixed clockspeeds and voltages. For both CPU/GPU of course.

[LeapFrogMasterRace]

23 minutes ago, lagittaja said:

Did you actually understand what those two graphs show?

Or did you even bother reading the article behind the link or in fact, actually just understand any of it? Since it seems like you did not.

 

~10% higher total system power consumption is "margin of error"?

How about you take a minute to comprehend what that means when you remove the motherboard/RAM/GPU/etc. from the equation and only look at dynamic CPU power...

In case your brain can't comprehend it or you're too lazy to actually read the article, the systems slice of the power consumption was ~130W and so with the 2600K running at 2Ghz@1.290V it means the dynamic CPU power went from 69W to 92W. That's a 33% increase.

Or for the 3Ghz@1.491V test it's 140W>165W = +17.8%.

The difference with the higher clockspeed+voltage test isn't as big because for that load, the NH-D14 passive cooled by the floor fan blowing air on it could not get the CPU temp lower than 67C vs the 47C CPU temp for the 2Ghz@1.290V load.

 

@LinusTech This needs a follow up Workshop episode where you take it to the extreme: aircooling with extremely high temperatures vs watercooling with as low temperatures as you can muster along with fixed clockspeeds and voltages. For both CPU/GPU of course.

At least I know the difference between a forum post and an article .

I know you are new but perhaps you should look at our community standards, on this forum we try to be nice to each other. 

[Martin86]

This is my first post, i've been following the channel on YouTube for years but I just had to register to say something about these new videos where they test different "Myths" out. The scientific quality is not that great. There are way to many variables that aren't controlled, which is also mentioned above by SaperPL.

 

At minimum they should have put a WattMeter on the outlet and logged wattage during the test. I know they have a unit they have used it in other videos.

You also have to fix clock speeds, voltage etc. The chips self-regulate performance based on temperature (throttling etc).

 

Also, if the system is running at lower a temperature, the electrical resistance in the copper decrease meaning the power consumption will decrease as well, outputting less heat into the room.