GN Dummy Heater Setup & Cooler Test Demonstration

This video offers a behind-the-scenes look at some of our new dummy heater and fan test equipment for CPU cooler reviews (and future fan reviews). This follows our cooler methodology update, found here: • Why Most Cooler Tests ...
We're showing the basics of setting up a synthetic power load on a dummy heater that we worked with a company to build.

[@gnsteve8846]

Watch our cooler methodology update here: kzbin.info/www/bejne/nJ63gH2GabCCr80

[@Allyouknow5820]

The actual mad lad. Went there and did it. Very happy to be a Patreon supporter because this is just absolutely insane !!! Saw Buildzoid's comment on the main video, and now I understand why he wouldn't do it.

[@pelor92]

As an Electrical engineer working mostly in the power electronics domain, but having also occasionally designed cold plates (for much higher power systems), I can say that this is by far the most scientific test system i ever seen for CPU cooler testing by anyone in the press, and I'm thorougly impressed, as I fully expect the data gathered on this system to be spot on

[@Im1CrazyCow]

Howdy Filippo, This is why we named Steve "Tech Jesus" He is in a Whole Other League when it comes to the Technical side of Tech to where other Channels do fluff Pieces , Steve Digs & Dissects things to Verify the Manufactures claims or Bullshit claims they try to make.....Steve was a smaller channel and most thought oh hes just here to make money or get free games at the time.... Im older so I saw something in him and followed a Young man Closer to My sons age than Me But its all good.........AS I said Most Saw Steve as a Kid trying to get free shit...he had a few things Hit but only 2k views & 1 or 2 hit 500k+ views etc .......Until he took Intel to Task and called BULLSHIT on these 2 Videos (below) About Intel Flat out Intentional deception to make Ryzen Look Bad...Most called intel Liars A 3rd party (Principle Technologies to muddy up the real facts) but Steve in his side step Humor tends to give them an "Out" which they hardly ever take... But then He hits them with facts or Questions they never see coming that everyone missed but aka "Tech Jesus" .... Here is when Steve Blew up from under 50k to where he is Now....this is when what we like to say was the Awakening of the Tech God... People are going back and watching some of his Old stuff so its funny to hear them spout off stuff on his Live Streams...I have to Bite my Lip BC most times they are wrong But they want Steve attention so bad i just let them Ramble. ... kzbin.info/www/bejne/emLQe4B_lNqUl5o .. kzbin.info/www/bejne/p6vWmZuofs9le7s The first video will give you the meat of the Issues & the 2nd will give you the "GOT You Moment !!!" Since this series Steve has been on a big incline BC even the other So called BIG Tech Channels Knew Steve was Good BUT this even Made them Realize ....Steve is in a whole other world of Tech and its Inner workings.......the best thing is if Steve does not know it he will learn it ......& if he cant he will get the Person who knows the Tech the best to explain it......he went from a goofy Kid swinging a bat in his driveway to "Tech Jesus".....I could Not be More prouder of him...Heck I still see him as Short Hair Steven......who played Star Citizen or tried too...lol Last Bit.....Patrick and Andrew & others I wont name BC they would rather not.....are a great Group of Young men whom alot of kids Could really Look up-to in today's Fake society of online posts of lies or just fake BS.... & Here is a Young man Whom does not Bring his BS to the world and hes had a few life Struggles but ...hes learn Life is what you fight for BC no one is going to give you shit.........Plus hes a great teacher of Tech.....that alot of the young kids dont realize how much Steve & GN are actually teaching them!! Sorry so long but Hes a great Young man and Ive been one of the few who saw it 11 years ago so ...... I tend to push him out there...........heck he didn't believe me about him going past 100K B4 xmas in 2018.....this was B4 the Intel Videos dropped....He still owes me on that bet..... its these Videos where Steve lets his hair down persay and shows how smart he really is most dont see much or dont care to BC its not about a Product. Ok Ill shut up now...... Cow}:-o)

[@-eMpTy-]

I mean, a few outlets have used dummy heaters in the past but with the addition of the fake CPUs it takes it a whole nother level.

[@willowjiggy2339]

this is the beginning of the end for the TDP and the beginning of the GNI (GamerNexusIndice)

[@estamnar6092]

@Erotikstudio Winkler GmbH Ive been saying this for years, 100% load at stock settings. The problem with that are the same that Steve coveres in the methodology video, about the Mobo, bios, Windows, ect doing bizarre stuff for no reason. Different mobos with different VRMs and loadline calibrations, ect means even with our ideal of 100% at stock would still be relative, and everyones mileage would vary.

[@everope]

@Erotikstudio Winkler GmbH 100% load means almost nothing. Just changing the FFT size in prime95 will give very different results in terms of power draw and temperatures, while both still technically being 100% CPU utilisation. The CPU changes the voltages and clock speeds of every core thousands of times per second, so it all depends on the scheduling and type of workload. Not to mention boost behavior that is enabled out-of-the-box. And even so, they will always favor marketing over accuracy. Putting a cooler that can only dissipate 65 watts on a 65 watt rated CPU will never be enough in real world use. It's best to just check out review testing results from different sources. Although I do sometimes use the marketed TDP to compare performance-per-watt as a ballpark figure. A 9900T will be much more efficient than a 9900k of course, and that is reflected in the TDP rating as is.

[@SWIRFTV]

"thermal watts are different then electrical watts" - AMD engineer

[@Hostilenemy]

Shouldn't call it a dummy heater. Heaters have feelings too! A constant burning sensation and a desire to be cooled by the thickest of heatpipes.

[@sp-xx8ty]

Thermal paste test too!!!! Mx4, pk3, nt-h1, kryonaut...

[@raulsaavedra709]

Awesome setup, needless to say uber professional. Some companies ought to learn quite a bit of HW testing methodology from GN

[@h.kurnia8175]

In fact his setup is derived from the testing done by the cpu cooler manufacturers

[@MTBScotland]

When you said dummy heater I thought that was to keep jay warm!

[@TheAnoniemo]

Now that you have this very accurate cooler testing, will you investigate what the variance between same model coolers is? I say because fans usually have a +-10% RPM range from one to another, and other variances in the production could influence performance in a way that could spoil your testing if a user could gets a product that performs 5-10% different from the review unit. E.g. Kitguru had two of the same model Noctua coolers that seemed to perform significantly differently when they tested a Zalman tower a couple weeks ago.

[@gnsteve8846]

Fan variance isn't relevant if you normalize for noise, but the rest is still interesting, yes. I doubt that two of the same model would perform "significantly" differently if controlled properly, but we can look into it, yes.

[@colajunkie13]

@@gnsteve8846 Since your time seems to be the critical factor in all of this, i'd imagine you'll only look into ones that have a reputation for high variance? How would you determine that? Other people's testing seems to be lacking (to be nice) compared to yours, so that's probably not a really good indication.

[@wertywerrtyson5529]

I never knew that reviews were so inaccurate until you explained it. Your dedication to accuracy is admirable. You might be the first to review computer parts so scientifically.

[@DangaRanga]

This whole setup is awesome from a data generation perspective. Would make a legitimate test platform to compare different brand pastes and different brand liquid metals. Would be very involved, but would be cool to throw on an ln2 pot and chart exactly how different pastes react at different temps. You guys are the best kinds of nerds

[@keineahnung8696]

I don't know if it is worth looking into it, but testing for changes in different climates (higher room temperature and humidity) might also give interesting data.

[@mikebutler9332]

This appeals to both the computer hardware enthusiast and the physicist in me. Great stuff.

[@BryceAC]

*No mention of tater tot performance??* :P Thanks for putting this insane amount of time, work and money into getting fantastic testing data. Any chance of a video going over some of the design process? (early stages, revisions, challenges with manufacturing, future plans, etc.) Cheers :)

[@spinshot6454]

OUT OF THE BOX THERMALS

[@tommihommi1]

I'm pretty sure your testing is more sophisticated than what some manufacturers use, at least the CPU die simulation.

[@mystical4442]

You really are offering some data that no other tech reviewer or website does. Godspeed, Steve.

[@jtm94]

This is actually impressive. Glad someone has put in this kinda legwork.

[@doublecrossedswine112]

Amazing. I hope at some point to see what the NH-U12A perform like on a 3950x

[@XFourty7]

Congrats so much on getting here. Thought you were beyond dedicated before, but now...? :O

[@ravishah2255]

So, I tried contacting you before about testing coolers because you complained about TDP being useless. But you got this figured out very nicely. There are few things which I would like to add though. First, there are just 3 main practical numbers that matter in testing coolers. How much temperature difference X a cooler needs across its ends, to conduct power Y, by creating noise Z(or corresponding RPM) from fan to increase the effective surface area between fins and air. The best way to represent these numbers is either a table where rows heading is watts and columns are noise level in decibel or X-Y graph where Y-axis is temperature difference and X-axis is Watt and there are multiple curves connecting data points corresponding to noise levels like 35 db, 40 db, 45 db, 50 db and max RPM. This will give a complete profile of the cooler. (Ideally, coolers should print this info as a table on the package instead of TDP) To do this you might need more steps in power levels like 50, 75, 100, 125, 150, 175, 200 watts. It's not necessary to have clean numbers for power levels, but it'd be great if you have the same power numbers for them across different coolers, ideally at common TDP ratings. To do a comparison between coolers, you can just make same graph, but instead of different noise levels, you put curves corresponding to the different coolers at the same noise level. This way of representing data is much better than just temperature numbers on a bar graph and a single graph can completely compare cooler performance profile given a fixed noise level. Second, the temperature difference has to be measured between two ends of the cooler. One end is air temperature because the effective fin-air surface area changes with the fan speed. For heat-sink surface temperature, we can make a hole on IHS and touch the center point of the cold plate with a thermal probe, which will bypass thermal paste, mounting pressure and other factors. I heard you mention a hole, but it seems like the probe wire is lying between the two plates. The bottom line is that I think you should try to measure the temperature of the center of cold-plate directly to eliminate other factors. If you're already doing it, great; if not, maybe you can try to implement it. I'm also unsure about how the internal arrangement of the die will affect the thermal performance, but if IHS does what its name suggests, it should have a negligible effect, I'm sure you'd check this yourself. To note, directly measuring cold-plate temperature will remove the uneven surface of cold-plate as a factor and it's important to talk about it. Perhaps you can find out the effect of this by comparing this with your other test-bench or have another temperature sensor on the hot plate of the dummy heater to figure out the combined effect of thermal material, mounting pressure and the evenness of the surface. You also have a surface testing tool for this matric, so this would be 4th most important number/factor for reviewing cooler (along with mounting pressure given by mounting bracket). Surface evenness is also something a user can fix by himself using sandpaper, maybe make a video to test how well sandpaper can fix the uneven surface of the cooler heat sink and find out the best method to do this. Just a future video idea. Lastly, there are more factors that decide the temperature of CPU apart from cooler like software gore, case airflow, thermal paste and intel-colgate/shouldering job inside the IHS. (Not including the obvious ambient temperature.) It'd also matter where the thermal sensors are located for a CPU model. However, these things should theoretically have an almost linear or predictable effect on temperature difference, so you can perhaps create a matrix to scale the performance across two testing methods. It would be a nice machine learning project. This can give more directly relevant data for a potential PC builder looking to purchase a cooler. It'd also be possible to make a small app that combines the case reviews, thermal paste reviews, CPU model and cooler reviews to calculate a final expected temperature of a build. I'd love to work on a machine learning model that calculates it if you can provide some training data and in this case, community members can also contribute their data if we develop a well-defined guideline for doing this test. You can also ask for thermal paste application pictures to ensure they aren't using too much or too little paste. Machine learning can remove bad data in its model, so we don't need to worry about bad data much. Overall, you have done a great job of designing the experiment. I hope my suggestions are useful to you and I'll look forward to seeing more videos.

[@thehumangerm]

Really great I know this was a ton of work. It is appreciated.

[@draeh]

I'm looking forward to TIM testing. I never know what type to use and whether there is a difference in what should be used on a CPU versus a GPU if there even is one. Thanks for all the testing work you do. I especially appreciate the data that backs up your reviews.

[@eric4946]

Was waiting for this stuff, really the right way to double check results

[@peterpain6625]

Pretty cool test setup. Might want to put a kind of air-tunnel behind it though to force exhaust upwards. Or it'll bleed back into the intake at higher rpm. Thanks for sharing.

[@Lurker-dk8jk]

Natural convection should help take care of that already.

[@All4Grogg]

Keeping it controlled is most important especially when testing relative results. That even includes seemingly tiny differences such as closing the door to the room or climate control intakes and exhausts. Placing the bench under a fume hood would make a pretty "big" difference. Fume hoods vented outdoors WILL change in airflow significantly with pressure differentials. Don't tell my idiot coworker that though, he's too busy drawing lines on the frame.

[@umeng2002]

Finally, real cooler comparisons and reviews.

[@coredumperror]

That's so cool, Steve! That cost number just blew my mind, though. Yikes.

[@Michael-OBrien]

Thank you for [planning on] doing stepped power increases on the coolers that will be tested. Heat pipes cap out, as I’m sure you know, and in a non-linear way. As such there is a rounding off or plateau that occurs as you reach QMax of the pipes.

[@bentosan]

This is so awesome Steve, well done this is a great contribution to the community

[@slantern6229]

I'm interested in the actual enclosure for the heater. Is it off the shelf or is it bespoke engineered?

[@VuNguyen-wo8bp]

Steve, please please please consider testing cpu waterblocks. You already have a MO-RA from some overclock streams, which should remove radiator capacity from any testing, so cpu block performance can be isolated. Brand new blocks from Optimus and EK are begging for accurate benches!

[@estamnar6092]

When a tech press group has a fuggin lab in it... love it

[@cabansinleaf8867]

Thanks for all the hard work you do guys, I personally have learned so much by watching you. This type of content is worth it's weight in gold

[@itstheweirdguy]

This is why I watch your videos. Keep it up.

[@jons2447]

Hello, Steve; Thank you for all you do. FANTASTIC! AMAZING! ASTOUNDING! STUPENDOUS! I could go on on but I think you get my point. I'm really impressed w/ the GREAT job, please keep up the GREAT work. Actually I've been looking for data that compares coolers based on heat flow. Your setup could be used to compile a database of cooler characteristics. Especially the amount of heat / power that each cooler can remove. Such a device if accurate, could totally revolutionize the cooler industry. I've subscribed because of this. I know your time is precious to you but maybe starting w/ most popular coolers you could test for heat transfer. Add other known factors & you got a cooler ranking unlike anything else available. Wish you the best w/ this, I think you're going to upset a lot of carts. Have a GREAT day, Neighbor! P.S. I know you're busy so taking a few minutes before each video to make a list of key points will actually save you time & help you to stay focused. Best.

[@MajorETX]

Impressive is an understatement

[@motor-head]

Now all the other hardware reviewers can just comment on how pretty a cooler is or how nice the box it comes in is. Because all of their "test" data is crap compared to yours.

[@Tigerskunk]

This reminds me of the power supply tester HardOCP had made. I think they where making a cpu cooler tester too before I lost track of the magazine and website.

[@ElettroStef92]

honestly I'm more interested in the analysis of the economic heatsinks compared to unnecessary € 200 liquid and rgb stuff . Really great job guys

[@ryanfav]

As bespoke as his particular setup is, If anyone else is after a setup similar to this, you can manage it for much less costs, steve just likely had to eat the R&D costs. This setup is using resistors to heat the IHS, however the same could be done with linear mosfets acting as constant current loads for much less, and is not the most complex circuit, e.g. 4 mosfets soldered directily to the IHS, so you can make it act as a 4 core device, in reality with the size of the IHS you could tightly fit 8 of the size of mosfet you would need, if you wanted to expand it (most of the cost of each circuit is the mosfet and the sense resistor, would be about $8-35 per "core", depending on precision) How I would have approach it would be something like a SDS1204X-E for the PWM channels, and something like a EEZ H24005 PSU for the heater supply, with another device for the thermocouple logging, both support SCPI, and the power supply software EEZ studio even lets you use python scripts to completly automate both the supply and any other SCPI devices. and offload the data at the end, as well as live veiw. Say you had each linear mosfet constant current load setup for a 5A constant load, you would have the power supply sense lines per channel connected across the mosfet, this rules out most of the complex math, and this video showed me that it was indeed the same, so you have whatever voltage you set * 5A to get the power of that "core", so in this example 5A* say 30V = 150W, there will be some small losses not coupled into the IHS, e.g. how well you insulate it, and heat flowing down the power supply wires, but its not hard to make it say 5.03A to compensate it mostly out.

[@tacticalcenter8658]

Thanks for clarification on the heater with the Chiplets and io die. That is a question I asked in the prior video. So this is great.

[@BenK12345]

that's a fancy hot plate you got there. looking forward to some numbers :)

[@JakusJacobsen]

Please, do some water block testing with that. There are some absurd claims made by some manufacturers about their products.

[@dougler500]

Thanks for the explanation of this. I'm really looking forward to seeing this method being used in your regular content. Very, very impressive.

[@joehorecny7835]

I would be interested in seeing the difference between using thermal paste, and not using thermal paste, which would be a mostly destructive test with a regular CPU but would be safe with your thermal simulator. Great Job and I'm looking forward to your future usage of this new tool.

[@-eMpTy-]

What do you mean with "destructive"? All modern CPUs will first throttle, then shut down when certain temperatures are reached.

[@aznwierdone]

Just wondering if there's a reason you guys have taped off the make and model of the power supply? Love the property id tag Also always really neat to see test setups! Keep up the great content, test system engineering, and journalism!

[@TheEntireUniverse]

You should publish a formal white paper about your CPU testing methodology. Maybe even submit it to an academic journal.

[@TCGView]

I love the Mad Scientist vibe I am getting from this.

[@All4Grogg]

Next will be an anechoic chamber for sound testing.

[@beedee95]

Amazing. Want to see more of this

[@dikbozo]

I am curious if the AM4 could be substituted for Threadripper or EPYC? I imagine yes and also for the various intel CPUs. Looking forward to more from this device.

[@lasouris1993]

Will you test the scallability of fan speed at a fixed heat load ? Great content !

[@MikeFic]

This is bananas.

[@agenericaccount3935]

Should seek a volunteer artist to do something with the walls. It's a well equipped lab, it should have some visuals too 😁

[@todayonthebench]

The dummy load should have likely used 4 pin connectors instead, so that you can regulate the voltage over the resistive element internally, as to also take out the wire resistance between the connector and the resistor. This is though a very small detail and you have already done a good job doing the sensing at the connector. So mostly a tip for the next load you build. Otherwise this is all a relatively nice setup and honestly beats any PC based test bench to be fair. (Since motherboards are laughably inaccurate as far as precision goes.)

[@gnsteve8846]

There's a sense line in there to help deal with most of the differences created along the wire.

[@todayonthebench]

@@gnsteve8846 Yes, it is good that you use sense lines. But the thermal test jig itself seems to only have two pinned connectors on it. This means that there is wires internally between the connector and the heating element. This means that the voltage drop over these wires aren't taken into consideration. Though, the contact resistance of the connector itself is also introducing some drop. The difference generally doesn't matter for relative measurements though, since this resistance would remain fairly constant. (Though, contact resistance does vary each time you replug it.) Using a four pin connector and running the sense lines all the way over to the heating element inside the box would though have been more ideal. Though, since you run at sub 3 amps, then you aren't going to see much voltage drop, and due to the high voltages you do use, you shouldn't expect that little voltage drop to have a major impact non the less. But mainly it is the engineering side of me pointing out a fairly minor detail. Just wanted to let you know. Also, nice to see a tech reviewer take reviewing this seriously and actually doing thermal testing properly.

[@micaiahflores1592]

"ill be fine" everyone on a 1000 ways to die

[@ItsKingMyles]

cant wait to see the numbers, noctua is the best clickbait

[@tropicbliss1198]

Nice

[@jcoc8877]

Wicked!!!

[@tst6735]

TechJesus, this is damn nice!

[@jmn8498]

Would it make sense to test the coolers without fans, too? In my mind, that would get a reading of how good is the heatsink itself.

[@julianangell342]

Genuine question for CLC testing: for noise-normalized thermals, how do you balance the pump and fan speeds, e.g. decreasing the pump speed to let fans run faster and cool the radiator more, or the inverse of lowering fan speeds to increase pump speeds and spread heat through the loop more effectively?

[@Pheatrix]

Is it possible to easily rearrange/change the heater units on the dummy heater? Or is it just that arrangement and if you want another "die" arrangement you have to buy another unit? Do you know how much results would change with another "die" arrangement? TL;DR: How "good" will the results be for other CPUs with this dummy heater? I doubt AMD will have the same die arrangement forever so is this "future proof" or would have to buy a completely new dummy heater for the results to be meaningfull for anything else thant Ryzen 3000 CPUs?

[@Sqtgdog]

Absolutely NOT trying to server shame, or judge you guys, but why isn't all that rack mount equipment, ya know, rack mounted?

[@brandonv8721]

Ever think of purchasing an electronic load? You calibrate equipment or assume drift is minimal?

[@Zarcondeegrissom]

"Ignore the government sticker", lol. Steve and crew are GNNL (Gamers Nexus National Laboratory), confirmed, lol.

[@Moliuan]

Do you plan on doing more PSU testing?

[@matonted]

I'm curious, why some of the equipment brand and model names are covered up? Is it for some internal reasons or you're not allowed to showcase them in videos?

[@Xoron]

Have you tried setting up the LASER-tach on a GorillaPod? I could imagine hand shaking making the reading more inaccurate.

[@berndeckenfels]

Multi channel modular lab bench supply, but it can't do the power calculation? ,)

[@EricM81]

Steve, this is awesome. Please test thermal interface materials (TIMs). Pastes are terrible. There are indium heat-springs (metal foil that plastically deforms and fills gaps with 40 psi of pressure), or you can polish the IHS and cooler flat. I don't buy the need for insulating pastes between micron flat surfaces. I think you will find this video from Tech Ingredients very interesting: kzbin.info/www/bejne/bn6yhYCnbsxrhM0 I'd be willing to help (and I won't need to run to the store to buy 5k sandpaper or diamond spray :).

[@mvanderplas2275]

Did the mini electric cooking plate come with tiny frying pans to make mini omelets?

[@Mythricia1988]

Is that power supply like surplus OrbitalATK equipment or something? :D

[@infango]

how do you test different IHS size ? can swap cpu IHS ?

[@darxustech2883]

Mainframe? 10:08 This seems to be a use of the term I'm not familiar with, so I'm curious what it means.

[@MaxHarden]

You need a small rack you can set on the table.

[@yanniskouretas8688]

There are tech reviewers and there is GN Tech Jesus .... Always learning ... To my knowledge only GN and Anandtech have this kind of bench for testing cpu coolers ...

[@eekpie]

Just to be annoying: Most resistors change resistance with temperature. Unless tracking wattage on result there might be a small error. Can that PSU work in constant wattage?

[@raulsaavedra709]

I don´t think that is too relevant. They can still properly (or as accurately as possible) compare cooler X under W load vs. fan Y at exactly the same W load pushed into their system. Both either with same fan speeds, or same noise levels, they can consistently push exactly the same W in each test for different coolers.

[@markjordan9059]

No officer, this is my F-16, I swear!

[@JGnLAU8OAWF6]

Would you automate this tests?

[@jintsuubest9331]

I don't understand the merit of having chiplet on the dummy heater, or having it placement like current ryzen 3000 chip. Intel for example haven't move to chiplet. If Intel did move to chiplet, the layout is not going to be same. We know some form of stacking is comming. AMD might come up with different layout with different interconnect. And much more. Are we just going to have to change the top whenever a new design come out?

[@gnsteve8846]

We have both.

[@bryandepaepe5984]

I'm sorry but every time I hear "the dummy heater" I can't help but to think of a cold Trump, maybe the thermal flux generator or TFG might sound better.

[@fragalot]

I'm a tad confused.. if you're giving the heating element 45 amount of watts, the same wattage a CPU consumes you think it will give out the same amount of heat output as a CPU would? I don't think that's apples to apples.

[@-eMpTy-]

Yes, it will give out the same amount of heat. Or at least very very close to it. If you'd put a PC with a 250w power consumption (from the wall) in a closed off room and run it for an hour, it would yield in the same result as running a 250w space heater in that same room for an hour.

[@fragalot]

@@-eMpTy- I don't buy that for one second. If you know ohms law then you'd see that can't be correct. Will a 250 watt light bulb give off the same heat? Incandescent or LED? It depends on it's efficiency. Resistance is what gives off it's heat, and if it's more efficient there's less resistance with more light and therefore less heat. Even resistors have a efficiency rating. Comparing it with a CPU isn't a proper analogue since it depends on what the CPU is doing. It's not like CPUs consume the same amount of power, and give off the same amount of heat at all times. If anything Steve is just simulating a CPU under full load by a CPU to test coolers, as if the CPU was consuming X amount of watts. If watts = heat output perfectly he wouldn't need so much equipment to test this. You can send 10 watts to a resistor with 5v DC and generate more heat than a CPU at 100% load if it has a very poor efficiency. I know this since I've built power supplies in high school (not switching PSUs, just 5/12v DC power converters).

[@-eMpTy-]

@@fragalot I don't really want to put the effort into explaining why that is but maybe this thread will enlighten you: superuser.com/questions/148070/where-does-power-consumption-go-in-a-computer

[@agenericaccount3935]

A look under the skirt 👌🏻

[@amerkiller1995]

5th

[@irishgiant5150]

Bro, don't say cat b-roll and then not show me a stinking cat.