The Experiment: Understanding CPU Thermal Performance with Thermal Pads

When it comes to understanding the thermal performance of CPUs, stress tests and experiments can provide valuable insights into their behavior under various conditions. In this experiment, we aimed to test the thermal performance of a high-performance CPU using a simple yet effective method: swapping thermal paste for thermal pads. The goal was to see how the addition of these pads would affect the CPU's temperature and, subsequently, its performance.

The setup began with a basic stress test, where we ran Geekbench 4 at high temperatures to gauge the initial thermal performance of the system. With the CPU cooler in place, we were able to achieve temperatures around 55°F, which was not extremely high but still indicative of the system's ability to manage heat. However, as we progressed with the experiment, it became clear that this was just the beginning of a more complex and intriguing scenario.

The next step involved swapping the thermal paste for a thermal pad, specifically designed for CPU cooling. This decision would alter the thermal interface between the CPU and the heat sink, potentially leading to changes in temperature and performance. As we applied the thermal pad and secured it in place, we couldn't help but feel a sense of trepidation about what might happen next.

The Results: Thermal Overheating and Performance

As soon as we applied the thermal pad, the temperatures began to rise, and our system started to exhibit signs of thermal throttling. The CPU temperature quickly rose above 80°C, with peaks reaching up to 92°C. This was significantly higher than our initial results, indicating that the thermal pad was indeed having a profound impact on the system's thermal performance.

Despite these extreme temperatures, the CPU's performance remained relatively stable, with Geekbench scores hovering around 451 across multiple runs. However, as we continued to monitor the system, it became clear that this stability was not sustainable in the long term. As the temperatures continued to rise, the CPU began to approach its thermal limits, and it seemed likely that some form of thermal throttling would occur.

The Re-Runs: Testing Thermal Throttling

To further understand the behavior of our system, we decided to re-run Geekbench 4 and Center Bench multiple times. The results were striking: each time we ran these tests, the CPU's temperatures rose to extremely high levels, but its performance remained largely unaffected. This suggested that while thermal throttling was present, it was not severe enough to impact the system's overall performance.

However, as we continued to run these tests, it became clear that our system was approaching its thermal limits. The temperatures were rising rapidly, and it seemed likely that if we continued to push the system, we would eventually reach a point where thermal throttling would become more pronounced. Fortunately, with each re-run of Center Bench, the CPU's performance remained stable, but this stability was fragile and short-lived.

The Conclusion: Thermal Pads and CPU Performance

After completing our experiment, we were left with some interesting observations about the use of thermal pads in CPU cooling systems. While these pads can provide a significant improvement in thermal interface between the CPU and heat sink, they can also lead to issues such as thermal throttling if not managed properly.

In our experiment, we saw that both the thermal pad and the CPU cooler were still effective mediums for exchanging heat sources with heat sinks. However, we also observed that using these pads could lead to significant thermal throttling, which would impact the system's performance and potentially cause issues such as fan noise and overheating.

The Implications of Our Experiment

Our experiment highlights the importance of understanding the thermal behavior of CPUs in various systems. By manipulating the thermal interface between the CPU and heat sink, we were able to observe changes in temperature and performance that might not have been apparent otherwise. This knowledge can be valuable for designers and engineers working on CPU cooling systems.

Furthermore, our experiment suggests that while CPUs are designed to handle extreme temperatures, they can still experience issues with thermal throttling if pushed too hard. As a result, it's essential to manage the thermal interface between the CPU and heat sink carefully to ensure optimal performance and longevity.

Future Experiments: Pushing the Limits of CPU Performance

As we move forward with our experiments, we're planning to push the limits of CPU performance even further. In our next video, we'll be attempting to overvolt the CPU by adding several volts to its standard operating voltage. We're eager to see how this will impact the system's performance and longevity.

We invite our viewers to suggest topics for future experiments in the comments section below. Your ideas and suggestions are welcome, and we look forward to exploring new avenues of research and discovery with you.

"WEBVTTKind: captionsLanguage: enit's about time we add a second video to our will it fry playlist in the first video we microwaved a CPU definitely fried in this video I have an I3 7350k which should not exist it's a stupid processor don't buy it more details somewhere around here we're going to swap thermal paste which is kind of like the interface between a heat source and a heat sink with this right here a thermal pad now for those of you new to this scene thermal paste or thermal grease is sandwiched between typically a CPU heat spreader and a large CPU heat sink or a water block something that helps bridge the gap between the Heat source which is the CPU itself and what is supposed to dissipate the heat the heat sink the compound in here in other variants is supposed to be a great conductor of heat that's the point by the way to transfer heat from point A to point B as efficiently as possible but thermal pads which do the same thing in essence are not spec the same these are meant more for vrms vrm modules things they get hot but not you know 100° C hot so the mitigation technique is the same but you wouldn't want to use this between a CPU and a heat sink we're going to do it anyway in this video okay so in goes the CPU now we're going to run two tests one will be of course our control our control test will consist of just typical thermal paste application we're going to use some noctua N1 here it's good thermal uh thermal grease so we're going to slap some of this on then we're going to use the cryorig C7 to cool the CPU and then after we've run our tests we'll run a few benchmarks just CPU synthetics we'll record temperatures uh we will remove the cooler remove the thermal paste and then apply the thermal pad and in case you're wondering this is 3 mm thick so here we go with the control test first slap some of that on there all right in before the oh that's too much thermal grease bro or that's too little man you should have put more than that it's fine trust me it's fine don't forget to remove your warning label all right slap this one right on top there all right good to go that's literally all we need to do now let's uh plug it in and run some benchmarks this is a beautiful LED combo by the way okay cinch first remember this is with the thermal pace so temperature should be completely fine here running at stock frequencies by the way and the fan curve is just whatever the fan curve was in the system bios out of the box okay so max temperature with cin bench was right at 60° C so we will take note of that by the way if you're wondering cinan score at stock speeds 4.2 gz turbo boost with 454 CB let's go and exit out of that one now uh sure we can save that now you you're probably wondering why on Earth I'm running benchmarks and not just Ida 64 which is kind of like a CPU stress test that's because if this CPU does thermal throttle at all we will see uh actual like differences and score so that's just something else to reiterate here I'm not saying that it's going to happen but it's likely uh that it will it's probably going a thermal throttle when we swap the paste for those thermal pads all right and now geekbench here we go it's the highest temperature running geekbench 4 was 55° F now it is time to swap the thermal grease on top of the CPU spreader here for a thermal pad not sure if I said this already but don't try this at home big waste of time not worth it period we're going to clean off the thermal paste here and on the CPU Cooler and then just slap one of these pads on there I'll try to cut it down to size about the size of the die itself all right let's see something like that yeah that should do yeah not a bad size oh this almost this almost hurts it would hurt if I cared about this CPU CPU stupid though so I don't really care all right got the CPU Cooler back on there and uh plugged in to go ahead give it a go all right so right off the bat idle temperatures are a bit higher we were in the low 30s before now we're in the low 40s and cin bench here we go okay temperatures are going up and up uh we're in 70 72 74 75 76 this is not good 77 79 Jesus 80 91° C 90 please 92 uh-oh okay so things are freaking hot right now 92° C 94 we've almost hit T Junction if I'm not mistaken the that this is not good at all okay now I'm going to rerun Center bench because our score was pretty much the same here our score was 451 CB which was only four lower than our previous run so because the CPU is nice and toasty now I'm going to rerun CER bench to see if we can uh instigate that thermal throttling we were talking about earlier well uh on that run exactly the same score 451 again so uh we're not thermal throttling and if we are it's I mean it's barely happening uh because our score is basically the same it's kind of weird I assume that if we had kept running the C bench test over and over eventually it would have thermal thrott you can see this is basically both cin bench runs here uh so we would approach the T Junction and once we hit it it would probably throttle by about 5 to 10% to keep the temperatures uh around 90 95° C anymore than that and uh yeah not good for the CPU so it makes sense that we're barely throttling if if we didn't have any thermal pad at all any thermal interface between the CPU and the heat sink I imagine we would thermal throttle almost immediately all right in Max temperature for geekbench 4 looks like was about 69° C so that's about 14° higher uh than it was with our control test the thermal pad looks to be in pretty good shape though so there's that now there are a couple things we could take away from this weird experiment you might be wondering why the heck we even bothered trying it out cuz I was curious mainly but you can also learn a thing or two here the first is that both of these are still good mediums of exchange between heat sources and heat sinks the second is that you should use this for your CPU and not this I mean you could still use this I guess I mean you'd have a lot of thermal throttling issues you'd probably have a very loud system because your fan would be compensating for all that heat that's not being dumped from your CPU uh but the thing about CPUs nowadays is that they're so well engineered that it's difficult to force a CPU to overheat and explode or you know shut your whole system down there are so many fail safes built into these systems into these CPUs nowadays that it's going to be pretty difficult to uh to just fry one so with that I guess it didn't fry and that's kind of to be expected at this point but uh I have a few more things up my sleeve for a couple other videos in the future on this playlist I think I'm going to try overvolting the CPU maybe like three volts or something ridiculous see if we can kill the CPU then I don't know suggest things in the comments we'll see what we can do to kill the i 37350 k with that if you like the video be sure to give it a thumbs up thumbs down for the opposite click subscribe button if you haven't already and I'll catch you in the next video this is science Studio thanks for learning with usit's about time we add a second video to our will it fry playlist in the first video we microwaved a CPU definitely fried in this video I have an I3 7350k which should not exist it's a stupid processor don't buy it more details somewhere around here we're going to swap thermal paste which is kind of like the interface between a heat source and a heat sink with this right here a thermal pad now for those of you new to this scene thermal paste or thermal grease is sandwiched between typically a CPU heat spreader and a large CPU heat sink or a water block something that helps bridge the gap between the Heat source which is the CPU itself and what is supposed to dissipate the heat the heat sink the compound in here in other variants is supposed to be a great conductor of heat that's the point by the way to transfer heat from point A to point B as efficiently as possible but thermal pads which do the same thing in essence are not spec the same these are meant more for vrms vrm modules things they get hot but not you know 100° C hot so the mitigation technique is the same but you wouldn't want to use this between a CPU and a heat sink we're going to do it anyway in this video okay so in goes the CPU now we're going to run two tests one will be of course our control our control test will consist of just typical thermal paste application we're going to use some noctua N1 here it's good thermal uh thermal grease so we're going to slap some of this on then we're going to use the cryorig C7 to cool the CPU and then after we've run our tests we'll run a few benchmarks just CPU synthetics we'll record temperatures uh we will remove the cooler remove the thermal paste and then apply the thermal pad and in case you're wondering this is 3 mm thick so here we go with the control test first slap some of that on there all right in before the oh that's too much thermal grease bro or that's too little man you should have put more than that it's fine trust me it's fine don't forget to remove your warning label all right slap this one right on top there all right good to go that's literally all we need to do now let's uh plug it in and run some benchmarks this is a beautiful LED combo by the way okay cinch first remember this is with the thermal pace so temperature should be completely fine here running at stock frequencies by the way and the fan curve is just whatever the fan curve was in the system bios out of the box okay so max temperature with cin bench was right at 60° C so we will take note of that by the way if you're wondering cinan score at stock speeds 4.2 gz turbo boost with 454 CB let's go and exit out of that one now uh sure we can save that now you you're probably wondering why on Earth I'm running benchmarks and not just Ida 64 which is kind of like a CPU stress test that's because if this CPU does thermal throttle at all we will see uh actual like differences and score so that's just something else to reiterate here I'm not saying that it's going to happen but it's likely uh that it will it's probably going a thermal throttle when we swap the paste for those thermal pads all right and now geekbench here we go it's the highest temperature running geekbench 4 was 55° F now it is time to swap the thermal grease on top of the CPU spreader here for a thermal pad not sure if I said this already but don't try this at home big waste of time not worth it period we're going to clean off the thermal paste here and on the CPU Cooler and then just slap one of these pads on there I'll try to cut it down to size about the size of the die itself all right let's see something like that yeah that should do yeah not a bad size oh this almost this almost hurts it would hurt if I cared about this CPU CPU stupid though so I don't really care all right got the CPU Cooler back on there and uh plugged in to go ahead give it a go all right so right off the bat idle temperatures are a bit higher we were in the low 30s before now we're in the low 40s and cin bench here we go okay temperatures are going up and up uh we're in 70 72 74 75 76 this is not good 77 79 Jesus 80 91° C 90 please 92 uh-oh okay so things are freaking hot right now 92° C 94 we've almost hit T Junction if I'm not mistaken the that this is not good at all okay now I'm going to rerun Center bench because our score was pretty much the same here our score was 451 CB which was only four lower than our previous run so because the CPU is nice and toasty now I'm going to rerun CER bench to see if we can uh instigate that thermal throttling we were talking about earlier well uh on that run exactly the same score 451 again so uh we're not thermal throttling and if we are it's I mean it's barely happening uh because our score is basically the same it's kind of weird I assume that if we had kept running the C bench test over and over eventually it would have thermal thrott you can see this is basically both cin bench runs here uh so we would approach the T Junction and once we hit it it would probably throttle by about 5 to 10% to keep the temperatures uh around 90 95° C anymore than that and uh yeah not good for the CPU so it makes sense that we're barely throttling if if we didn't have any thermal pad at all any thermal interface between the CPU and the heat sink I imagine we would thermal throttle almost immediately all right in Max temperature for geekbench 4 looks like was about 69° C so that's about 14° higher uh than it was with our control test the thermal pad looks to be in pretty good shape though so there's that now there are a couple things we could take away from this weird experiment you might be wondering why the heck we even bothered trying it out cuz I was curious mainly but you can also learn a thing or two here the first is that both of these are still good mediums of exchange between heat sources and heat sinks the second is that you should use this for your CPU and not this I mean you could still use this I guess I mean you'd have a lot of thermal throttling issues you'd probably have a very loud system because your fan would be compensating for all that heat that's not being dumped from your CPU uh but the thing about CPUs nowadays is that they're so well engineered that it's difficult to force a CPU to overheat and explode or you know shut your whole system down there are so many fail safes built into these systems into these CPUs nowadays that it's going to be pretty difficult to uh to just fry one so with that I guess it didn't fry and that's kind of to be expected at this point but uh I have a few more things up my sleeve for a couple other videos in the future on this playlist I think I'm going to try overvolting the CPU maybe like three volts or something ridiculous see if we can kill the CPU then I don't know suggest things in the comments we'll see what we can do to kill the i 37350 k with that if you like the video be sure to give it a thumbs up thumbs down for the opposite click subscribe button if you haven't already and I'll catch you in the next video this is science Studio thanks for learning with us\n"