Jump to content
Search In
  • More options...
Find results that contain...
Find results in...

Why Overheat your CPU on Purpose?

ColinLTT
 Share

There are SO MANY options for stress testing, but which is the best? We put not one but TWO systems through the wringer to find out.

 

 

Get HWInfo and GenericLogViewer here: https://www.hwinfo.com/add-ons/

Link to comment
Share on other sites

Link to post
Share on other sites

Why is the intro missing from last few videos?

Link to comment
Share on other sites

Link to post
Share on other sites

Interesting - dont particularly mind either way although I do miss the editor's 'blurb'. (Maybe that can go in end?)

 

But on-topic this was a great video idea, Thanks!

Link to comment
Share on other sites

Link to post
Share on other sites

6 minutes ago, swimtome said:

Interesting - dont particularly mind either way although I do miss the editor's 'blurb'. (Maybe that can go in end?)

 

But on-topic this was a great video idea, Thanks!

The blurb (called a Slogan internally) Is actually written by the writer (in most cases!) 

For this video, it was: Is it getting hot in here? 
 

Link to comment
Share on other sites

Link to post
Share on other sites

Nice video (for a change 😛).


Just one minor complaint, irrelevant really but it did trigger my OCD: that wobbly desk... hah.

VGhlIHF1aWV0ZXIgeW91IGJlY29tZSwgdGhlIG1vcmUgeW91IGFyZSBhYmxlIHRvIGhlYXIu

Link to comment
Share on other sites

Link to post
Share on other sites

33 minutes ago, ColinLTT said:

It's an editorial decision we're trying out. 

I did not even notice.

                                                                                      ██▓▒░    🆉 🆈 🆉 🆉 🆈 🆇    ░▒­­▓██

                                                                                        Rule 45: If there is not, there will be.

 

PC: ROG Flow X13 (5900HS, 3050ti, 16GB 4266Mhz) + XG Mobile (3080M 16GB 150W)
VR Headset: Meta (hmm) Quest 2
Phone: Samsung Galaxy Z Flip 3 5G + Galaxy Watch 4

 

Link to comment
Share on other sites

Link to post
Share on other sites

11 minutes ago, Biohazard777 said:

Nice video (for a change 😛).


Just one minor complaint, irrelevant really but it did trigger my OCD: that wobbly desk... hah.

Thank you for adding absolutely nothing to this conversation.

                                                                                      ██▓▒░    🆉 🆈 🆉 🆉 🆈 🆇    ░▒­­▓██

                                                                                        Rule 45: If there is not, there will be.

 

PC: ROG Flow X13 (5900HS, 3050ti, 16GB 4266Mhz) + XG Mobile (3080M 16GB 150W)
VR Headset: Meta (hmm) Quest 2
Phone: Samsung Galaxy Z Flip 3 5G + Galaxy Watch 4

 

Link to comment
Share on other sites

Link to post
Share on other sites

I don't understand how the power draw can spike like it does, while the clock speed stays flat. I thought they were directly connected.

 

41 minutes ago, ColinLTT said:

It's an editorial decision we're trying out. 

I like it so much, ty.

Though, kinda hate how I keep skipping too much of the content now, so used to skipping around 10 to 20 seconds.

Link to comment
Share on other sites

Link to post
Share on other sites

My HWinfo64 shows about 10c higher temps than occt (v4.5.1) while Linpacking... 

who do u trust ?  and who got the fakenews'?

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, Zyzzyx said:

Thank you for adding absolutely nothing to this conversation.

What is your contribution to this conversation? 🤣

VGhlIHF1aWV0ZXIgeW91IGJlY29tZSwgdGhlIG1vcmUgeW91IGFyZSBhYmxlIHRvIGhlYXIu

Link to comment
Share on other sites

Link to post
Share on other sites

Shoutout to Prime95

"And I'll be damned if I let myself trip from a lesser man's ledge"

Link to comment
Share on other sites

Link to post
Share on other sites

A lot of these indeed are not real world situations making their use kind of stupid - even for a full custom water setup are you really going to run some synthetic benchmark for 40 minutes that has nothing to do with what you intend to use your system for???

 

I run F@H and for me that's the best indicator for how good (or bad) a cooling and PSU setup performs. I have four GPU's along with my CPU/chipset on EK blocks with roughly a gallon of coolant going through the loop. These feed into 5 chunky 80mm thick alphacool radiators that are cooled by a combined 30 120mm/140mm fans in push/pull. To actually measure the performance I take the WATER temperature right at the point where the last block in the chain (heat source) enters the rad. I use a spare mobo temp header for the connection. I use the built-in Asus cooling controls for the mobo to set up the actual cooling profiles, and a desktop widget to monitor the actual temps. I also use GPU-Z to keep an eye on each GPU with the sensor tab open set to a 2.5 second interval.

 

Power draw is not an issue, since even with all four cards and the CPU at full tilt I'm pulling 1460 watts from the wall which is well within the 1600W spec of my T2 power supply. That stress does regularly include power spikes as well since each of the four cards will pause briefly at the end of each F@H job, something that can be clearly seen in GPU-Z's graphs.

 

Back to those Asus cooling controls - I have two custom profiles set up, a regular one and a summer one. Both regulate fan speed based not on CPU or GPU temps but on water temps, and that's because the water temp will always be lower than your GPU/CPU. The summer one I use when my ambient temps hit 30C which is pretty common where I live. Many cooling solutions would simply fail in those conditions yet I can run my system at full tilt and on 24/7 without a problem.

 

To give you an example, with a 30C room ambient my water temps stay below 52C, my four GPU's average 62C and the hottest core on my CPU is at maybe 70C. The key point here is the D5 pump I use which is rated for 60C max liquid temperature and I am well below that limit. This is also with the fans running in their mid RPM range - you can hear them but they are not at their maximum or overly loud. Since temperature rise is linear in both the ambient and water I would be looking at close to a 40C ambient to really hit the 60C pump limit, which still won't be the case for me as the fans speeds will climb to compensate for that temperature climb likewise. Maybe at 45C ambient my system would hit it's limit and switch off, but at that point it would be too hot for me to be sitting in that room, and it would probably have to be 50C ((122F) outside at that point as well. There are only a 2-3 places on the planet where it ever gets that hot, and you don't see people running balls-to-the-wall desktops there. Ain't gonna happen!!!

 

Modern server systems are expected to work in 45C ambient environments and many boards are validated to run at those extremes. Your typical desktop will never see such punishment, but a well built custom water system can definitely handle a summer room without air conditioning and keep your system running 24/7 with relative quiet.

 

Widget system.PNG

GPU Z quad SLI.PNG

Link to comment
Share on other sites

Link to post
Share on other sites

Isnt ist a thing that I could shorten the lifetime of my cpu If I Stresstest it like 48h. Its  maybe not extrem but for example it working now 9 years instead of 10?

Link to comment
Share on other sites

Link to post
Share on other sites

As a side note, the error with Intel Extreme Tuning seems to indicate that this feature was enabled (in Windows Security > Device Security).

 

image.png.a90f734c2916db48612091fccd43f1f3.png

 

It's nothing to do with a BIOS feature as suggested in the NUC video, it's a Windows security feature.

˙ǝɯᴉʇ ɹnoʎ ƃuᴉʇsɐʍ ǝɹɐ noʎ 'sᴉɥʇ pɐǝɹ oʇ ƃuᴉʎɹʇ ǝɹɐ noʎ ɟI

Link to comment
Share on other sites

Link to post
Share on other sites

"No real world application behaves like that" - The comedy being Prime95 is a real world application, it just has a benchmarking tool built in xD

Link to comment
Share on other sites

Link to post
Share on other sites

For a worst-case scenario full system load, I'd have liked to have seen GPU benchmarks thrown in there as well in combination.

 

Looking at Intel vs AMD temps, with the exception of Prime95, the rest of the temps aren't horribly far apart from AMD's temps. However, the Prime95 temps seem pretty disproportionately high for Intel. I wonder what's going on with Intel's execution units that they're drawing so much power when unchained?

My eyes see the past…

My camera lens sees the present…

Link to comment
Share on other sites

Link to post
Share on other sites

3 hours ago, ColinLTT said:

It's an editorial decision we're trying out. 

I usually skip through the intro and makes me more likely to skip during the video, without the intro I tend to skip much less, so I am glad you are looking into it

PC: Alienware 15 R3  Cpu: 7700hq  GPu : 1070 OC   Display: 1080p IPS Gsync panel 60hz  Storage: 970 evo 250 gb / 970 evo plus 500gb

Audio: Sennheiser HD 6xx  DAC: Schiit Modi 3E Amp: Schiit Magni Heresy

Link to comment
Share on other sites

Link to post
Share on other sites

5 minutes ago, Ubersonic said:

The comedy being Prime95 is a real world application, it just has a benchmarking tool built in xD

I knew of Prime95 from doing my Asc in Mathematics, but only after I started looking at LTT videos I used it as a benchmark. Its wild that most people don't know its original purpose 😅

PC: Alienware 15 R3  Cpu: 7700hq  GPu : 1070 OC   Display: 1080p IPS Gsync panel 60hz  Storage: 970 evo 250 gb / 970 evo plus 500gb

Audio: Sennheiser HD 6xx  DAC: Schiit Modi 3E Amp: Schiit Magni Heresy

Link to comment
Share on other sites

Link to post
Share on other sites

4 hours ago, Biohazard777 said:

What is your contribution to this conversation? 🤣

Feedback on the editorial change they made.

                                                                                      ██▓▒░    🆉 🆈 🆉 🆉 🆈 🆇    ░▒­­▓██

                                                                                        Rule 45: If there is not, there will be.

 

PC: ROG Flow X13 (5900HS, 3050ti, 16GB 4266Mhz) + XG Mobile (3080M 16GB 150W)
VR Headset: Meta (hmm) Quest 2
Phone: Samsung Galaxy Z Flip 3 5G + Galaxy Watch 4

 

Link to comment
Share on other sites

Link to post
Share on other sites

IME, OCCT Large is the best at figuring out CPU stability, quickly.  Linpack and P95 don't really work on marginally stable overclocks.

Workstation:  12900KF @ 5.2Ghz || MSI Pro-A Z690 DDR4|| EVGA FTW3 3090 1000W || G.Skill 3866 4x8GB || Corsair AX1500i || whole-house loop.

LANRig/GuestGamingBox: 9900nonK || Gigabyte Z390 Master || ASUS TUF 3090 650W shunt || Corsair SF600 || CPU+GPU watercooled 280 rad pull only || whole-house loop.

Server Router (Untangle): 8350K @ 4.7Ghz || ASRock Z370 ITX || 2x8GB 3200 || PicoPSU 250W, running on AX1200i || whole-house loop.

Server Compute/Storage: 10850K @ 5.1Ghz || Gigabyte Z490 Ultra || LSI 9280i-24 port || 4TB Samsung 860 Evo, 5x10TB Seagate Enterprise Raid 6, 4x8TB Seagate Archive Backup ||  whole-house loop.

Laptop: HP Elitebook 840 G8 (Intel 1185G7) + 3060 Thunderbolt Dock, Razer Blade Stealth 13" 2017 (Intel 8550U)

Link to comment
Share on other sites

Link to post
Share on other sites

@ColinLTT Do you have internal guideline on how to visualize data?

Beside general styling questions like the unit within the title or not formatting the parameter in the axis label as italic type there is an impressive amount of inconsistency.

For software I guess excel is good enough but there is paid software like Origin Pro or open source programs like GNUplot which do a better job of generating professional looking charts.

 

a collection just from this video

no axis title, no title:

image.png.4bba1303e008745678323c6a8ac03105.png

this time tittle but no axis label:

image.png.45d6081302be115e2ed7a05871c77aa3.png

 

completely different template but this time with title and x-axis label and unit for the x-axis. Also notice how this is the only chart plotting temperature above ambient while all the other  have temperature as parameter.

image.thumb.png.056b511ac2cccfda8af4d55116d597d9.png

 

program screenshot instead of excel (screenshots shows that actually a .csv was generated):

image.thumb.png.1fa4335da1c61651f5cb003e09883870.png

cut in half axis label and the other don't has it. Also right title features unit in the title while the left doesn't. Also every other unit now had ()  while this is now []

image.thumb.png.65e04d00e508cf1b21ed1f17be78ed77.png

Unit error (C = colomb/unit of charge) and now there is a axis label on x-axis (test) while previously on equivalent charts there where none. Also the same test has now a different name: e.g OCCT "CPU" - ... vs. OCCCT CPU:

image.png.86e7acfbf40b784af893bbc123948974.png

here is the Intel data again this time a different template and again no axis labels or unit:

image.thumb.png.65f5465db3cfc016757e7daf81ed7187.png

another one: this time () and [] mixed within the same chart:

image.thumb.png.6a1b3ef28d4b285ef6570886163a910e.png

 

Link to comment
Share on other sites

Link to post
Share on other sites

4 hours ago, Zyzzyx said:

Feedback on the editorial change they made.

Much wow. 😉

VGhlIHF1aWV0ZXIgeW91IGJlY29tZSwgdGhlIG1vcmUgeW91IGFyZSBhYmxlIHRvIGhlYXIu

Link to comment
Share on other sites

Link to post
Share on other sites

11 hours ago, ColinLTT said:

There are SO MANY options for stress testing, but which is the best? We put not one but TWO systems through the wringer to find out.

 

 

Get HWInfo and GenericLogViewer here: https://www.hwinfo.com/add-ons/

What settings were you using for the IntelBurnTest? (because the size of the matrix array that it is solving can make a difference in your results)

IB >>> ETH

Link to comment
Share on other sites

Link to post
Share on other sites

8 hours ago, Luscious said:

A lot of these indeed are not real world situations making their use kind of stupid - even for a full custom water setup are you really going to run some synthetic benchmark for 40 minutes that has nothing to do with what you intend to use your system for???

 

I run F@H and for me that's the best indicator for how good (or bad) a cooling and PSU setup performs. I have four GPU's along with my CPU/chipset on EK blocks with roughly a gallon of coolant going through the loop. These feed into 5 chunky 80mm thick alphacool radiators that are cooled by a combined 30 120mm/140mm fans in push/pull. To actually measure the performance I take the WATER temperature right at the point where the last block in the chain (heat source) enters the rad. I use a spare mobo temp header for the connection. I use the built-in Asus cooling controls for the mobo to set up the actual cooling profiles, and a desktop widget to monitor the actual temps. I also use GPU-Z to keep an eye on each GPU with the sensor tab open set to a 2.5 second interval.

 

Power draw is not an issue, since even with all four cards and the CPU at full tilt I'm pulling 1460 watts from the wall which is well within the 1600W spec of my T2 power supply. That stress does regularly include power spikes as well since each of the four cards will pause briefly at the end of each F@H job, something that can be clearly seen in GPU-Z's graphs.

 

Back to those Asus cooling controls - I have two custom profiles set up, a regular one and a summer one. Both regulate fan speed based not on CPU or GPU temps but on water temps, and that's because the water temp will always be lower than your GPU/CPU. The summer one I use when my ambient temps hit 30C which is pretty common where I live. Many cooling solutions would simply fail in those conditions yet I can run my system at full tilt and on 24/7 without a problem.

 

To give you an example, with a 30C room ambient my water temps stay below 52C, my four GPU's average 62C and the hottest core on my CPU is at maybe 70C. The key point here is the D5 pump I use which is rated for 60C max liquid temperature and I am well below that limit. This is also with the fans running in their mid RPM range - you can hear them but they are not at their maximum or overly loud. Since temperature rise is linear in both the ambient and water I would be looking at close to a 40C ambient to really hit the 60C pump limit, which still won't be the case for me as the fans speeds will climb to compensate for that temperature climb likewise. Maybe at 45C ambient my system would hit it's limit and switch off, but at that point it would be too hot for me to be sitting in that room, and it would probably have to be 50C ((122F) outside at that point as well. There are only a 2-3 places on the planet where it ever gets that hot, and you don't see people running balls-to-the-wall desktops there. Ain't gonna happen!!!

 

Modern server systems are expected to work in 45C ambient environments and many boards are validated to run at those extremes. Your typical desktop will never see such punishment, but a well built custom water system can definitely handle a summer room without air conditioning and keep your system running 24/7 with relative quiet.

 

Widget system.PNG

GPU Z quad SLI.PNG

The problem with using F@H for testing/benchmarking is that the workload varies significantly depending on the work unit that their servers send to your system/component.

Therefore; if you are trying to find out how hot your system gets (e.g. testing whether your thermal management solution is sufficient and efficient) and/or whether your system is stable, the variability with F@H makes a cross-platform comparison like this virtually impossible unless you purposely COPY (and retain a copy) of the work unit data so that you can start up the client with the former command-line tool (rather than with its vastly prettier GUI), which forces F@H to work on the same WU over and and over again.

 

 

7 hours ago, Tamesh16 said:

I knew of Prime95 from doing my Asc in Mathematics, but only after I started looking at LTT videos I used it as a benchmark. Its wild that most people don't know its original purpose 😅

According to the OCED PISA results, it is no secret that the general population of many countries are bad at math. There are some countries where they typically fair a little bit better, but there are also many that fair significantly worse as well.

 

For example, this video mentions that LINPACK solves the linear algebra system Ax=b, but very few people might actually know nor understand that it actually uses the Gaussian solver with partial pivoting to actually perform that solve and there might be even more people who DON'T know that the "rests" in between is actually the system checking the results for the vector, x, by computing the residual, where per the aforementioned linear algebra equation system, Ax=b, Ax-b should equal 0.

But in computer science (and also numerical analysis and matrices), Ax-b does NOT always give you 0, and the value that it returns is called the residual, and you want to try and minimise that residual as much as possible by balancing performance with accuracy/precision.

 

This is the range of results that you can get with the AMD Ryzen 9 5950X (on an Asus X570 TUF Gaming Pro WiFi motherboard, with 4x 32 GB DDR4-3200 unbuffered, non-ECC RAM):

 

5950X_16C_w_SMT.png.0261f67dcd4b2aa981005f0c2a4bea72.png

 

And here is the same test, on an Intel Core i9-12900K on an Asus Z690 Prime-P D4 motherboard, also with 4x 32 GB DDR4-3200 unbuffered, non-ECC RAM:

12900K_8P_w_HT_8E.png.1ce77d609e7b6ad8ecf1c5708efbeb79.png

 

And here is what you can get when you overlay the results between the two platforms over top of each other:

900823378_ComparisonofALeftSolvebfrom1024to65536on5950Xand12900KCentOS7.7.1908andWin1021H22022-02-22.thumb.png.30e6c6f26353e3965151ae1dcb9a4f71.png

IB >>> ETH

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share


×