2024 CPU Cooler Test Bonanza – Five Air Coolers Compared

Introduction – Welcome to my nightmare

Have you ever had one of those great ideas that you’re certain will improve things, but ends up being a massive headache? That’s where I’ve been for the last few months with this project. It started when Be Quiet! graciously provided a Dark Rock Pro 5 cooler for review. At the time I knew my test system was well behind the times, as well as being acutely aware that many of the most respected PC hardware review sites have been changing up their cooler testing methodologies to reflect more accurate representations about the true capabilities of the coolers. I felt like I needed to follow suit, but also to come up with something unique to PC Perspective, and not just copy/paste someone else.

Much of the issue stems from the fact that most recent Intel and AMD platforms just shoot for certain temperatures on their latest processors, then adjust power and clocks to stick at those temps. What this boils down to is you can have two coolers providing the same temperature, but the CPU will be operating at different power levels and clocks on each cooler, so older test methods, which just measure temperature, are not providing a complete picture of what is happening. It is also important to remember that both Intel and AMD have changed their IHS designs within the last two years and older contact plate designs have not performed as well on the new CPUs from either manufacturer.

I will not recount all the trials and tribulations I encountered while I attempted to find a decent set of testing parameters. I did go through five separate CPU stress testing programs, trying to find a test which would provide not only consistent loads throughout the test, but also consistent temperatures from test to test. I quickly discovered that Cinebench does not provide accurate and repeatable measurements as temperatures would spike, and then drop, between loops of the test. This was especially a problem at the lower wattage load. A couple of other software testing suites did not actually force full load of the CPU in the 250 watt and 275 watt tests. What I ended up with provides, what I feel, is a massive amount of data that gives a pretty complete picture of what each cooler offers in performance, but before we go too far into testing, let’s take a look at each cooler.

be quiet! Dark Rock Pro 5

be quiet! has been in the dual tower cooler game for a while (if the identifier of “5” didn’t clue you in). During that time the company’s design language and focus on silence and elegant styling has endured, and made them very popular in the PC enthusiast community. In previous reviews I have mentioned that one of my favorite coolers of all time was the Dark Rock Pro 3. The Dark Rock Pro 5 is another positive step in the evolution of that design.

Beneath its well-designed, magnetic top cover are two massive fin arrays to dissipate the heat from seven 6mm heatpipes. The cooler includes one 135mm fan, positioned between the fin arrays, and one 120mm fan, positioned above the ram. When fully assembled with the fans the Dark Rock Pro 5 has a total height of 168mm, and weighs just over 1.29 kilograms. The MSRP in the US is $89.90, but at the time of writing is selling for $79.90 at your major retailers.

As with every Be Quiet! Cooler I have reviewed, fit and finish on the Dark Rock Pro 5 is outstanding. Some may like the all black, stealth bomber aesthetic, while others may not be happy with the lack of any lighting at all. This really comes down to personal preference, but at the time of writing it’s the only option for the Dark Rock Pro 5. There is a new, higher level model offered by Be Quiet! called the Dark Rock Pro Elite, which does offer some (still subtle) RGB effects on the top cover plate, slightly higher speed fans, and replaces the 120mm front fan with another 135mm. These changes bring the Elite up to an MSRP of $99.90 US.

An interesting feature of the Dark Rock Pro 5 is a dual mode switch for Performance or Quiet modes. This switch limits the top speed of the fans. In quiet mode, the max fan speed is limited to 1500 rpm for the 120mm fan, and 1300 for the 135mm fan. In performance mode, the fans can hit their max speeds of 2000 rpm and 1700 rpm respectively. Bear in mind that your motherboards PWM sensor will only read the speed of the 135mm fan. Personally, I don’t see a huge benefit to this mode, as if you want to limit your fan speed for noise reasons, it can be done very easily in practically any UEFI.

One of the big knocks on Be Quiet’s flagship coolers has always been the mounting system, but this has been gradually improving with subsequent iterations of coolers. The Dark Rock Pro 5’s mounting system is the best that I have encountered on any Be Quiet! cooler to this point, but it’s not without complaint. In my humble opinion, the Intel mounting system has too many tiny pieces, including four tiny o-rings/spacers that could be incredibly easy to lose. More than once I dropped one of these during the install process, but was able to find them. That was lucky because there are no spares included. These o-rings are used to hold the four main screws to the backplate, and to keep the backplate spaced away from components on the rear of the motherboard. After getting those four screws installed, and the backplate in position, the rest of the install was as easy almost anything else on the market, and extremely secure once everything is mounted.

Thermalright Peerless Assassin RGB

I am fairly certain that if I had attempted to publish this type of cooler shootout, and not included the Peerless Assassin, I would be hunted down and beaten. In an effort to avoid that outcome I included two. This cooler has been the darling of air cooling enthusiasts for the better part of the last 18 months.

To be clear, as Thermalright offers about a dozen different models called “Peerless Assassin”, the first model tested here is the Peerless Assassin 120 White ARGB. It is based on the original Peerless Assassin, 157mm tall, dual tower, six heatpipe design. Completely assembled, this cooler weighs in at 1020 grams.

It’s a really attractive design, although, being honest, both the cooler itself, and particularly the fans, feel like a step down in quality compared to the Be Quiet!. In particular I noticed a slightly annoying, higher pitch from the fans that are included with this cooler, than any of the others. It is possible this is the result of the less rigid material of the fan blades that is necessary for the RGB effects.

One big reason the varied Peerless Assassin models have been so popular is the price to performance ratio. While pricing does fluctuate a bit, and some models (such as the White ARGB I managed to get) go out of stock often, all of the Peerless Assassin models can be found between $30 and $40 US dollars. When you compare that price to the performance (which we will get to later) it’s understandable why Thermalright has made such a dramatic return to prominence in the CPU Cooler market.

The mounting system on this model is slightly older than the other two Thermalright coolers in the shootout, so it differs slightly, but only in the backplate design. With this particular model, the main bolts are secured to the backplate by four plastic spacers. These are much easier to deal with than the tiny o-rings on the Be Quiet! and far easier to find if you happen to drop one (which I avoided entirely during this install). I did find that I had to fidget a bit with the cooler to get the spring loaded attachment screws lined up correctly, and this part of the experience was similar on all the Thermalright coolers. More than once I would tighten the first screw down a couple of turns, only to have it pop back off once I started tightening the second. It wasn’t a bad experience, it just took some extra care to insure the screws were lined up just right before starting to tighten them.

Thermalright Peerless Assassin 120 Mini

One might look at the inclusion of this cooler and wonder why it is here. The 120 mini is much smaller (135mm height) than any of the other coolers in this test, and has only a single, 120mm fan. To be honest, it had just released when I was working on this test, so I grabbed one to include and see how it stacked up.

As mentioned, it is much shorter than any of the other dual tower coolers, and lighter as well, with an overall weight of just 820 grams. Granted, having only the single fan skews the weight comparison, but adding a second would only total 940 grams. It still uses the same six x 6mm heatpipe configuration of the full size version, just in a shorter overall package. It will be interesting to see how this relatively diminutive cooler performs.

The mounting process here utilizes a different backplate from the 120 ARGB model. The new backplate utilizes simple click adjustments that are set to the indicated socket. Everything else was identical to the full size model. Overall, the fit and finish of the heatsink itself seems slightly better than the standard Peerless Assassin, and the fan feels to be much higher quality, with a heavier, stiffer frame and sturdier blades.

Thermalright Phantom Spirit 120 EVO

Finally, the last Thermalright cooler in this test is the Phantom Spirit 120 EVO. Much like the 120 Mini, the Phantom Spirit had just released as I was preparing to do this test, so I picked one up. The Phantom Spirit 120 EVO utilizes an entirely different heatsink design from the two Peerless Assassin models. Thermalright have added an extra heatpipe, bringing the total to seven, and matching the number of heatpipes on the Dark Rock Pro 5. The heatsink, at 157mm, is the same height as the Peerless Assassin, but with a more dense fin array. Without the inclusion of fans, the Phantom Spirit EVO heatsink is 60 grams heavier than the Peerless Assassin. With the fans, the total weight is 1250 grams.

The quality level of the Phantom Spirit EVO is much higher than either of the Peerless Assassin models tested. The heatsink is attractive and well constructed. Fit and finish are also getting close to Be Quiet! level. The fans are also a major step up. While these may not yet have the level of quality that you will find in a Be Quiet! or Noctua fan, they are certainly in the tier just below. They also feature a very tasteful use of RGB, with simple highlights on the frame.

The mounting process of this cooler was essentially identical to the Peerless Assassin Mini. I feel that it might have been slightly easier to get the tension screws lined up and started than on the other Thermalright coolers, but by this point, I’d gone through the process a couple of times, so it could have just been my own familiarity with the process, and not an actual fitment improvement.

When released, the Phantom Spirit 120 EVO was selling for $39.90 US. This has increased and pricing now seems to fluctuate between $42.90 and $49.90 US.

be quiet! Dark Rock TF2

Some of you may be thinking Didn’t he review this cooler almost three years ago. If you are thinking that, thanks for being with us so long. Yes, I did review this cooler in 2021.

The Dark Rock TF2 was the last model, in the last generation of be quiet! Coolers, before they updated to the LGA 1700 mounting. A kit can be purchased from be quiet! for many of their last gen coolers to adapt them to LGA 1700, so I wanted to include this model in this test to see how the older contact surface and heat pipe design compare to newer coolers on this larger Intel CPU. Detailed specs of the Dark Rock TF2 can be found in my review here.

Performance Testing

In my introduction, I went through a brief description of what I finally came up with for the new testing. There are two of the metrics that I wish to explain further, and explain the difference. These are the Average temperature vs. Peak Temperature.

One thing that I began to notice as I started gathering initial data points is that often, the peak temperature between two different coolers could be close, or even the same, but overall, one of the two would generally maintain a lower temperature. This is an example of a short temperature spike, which can happen for a number of reasons, but might not be the best metric to use for determining performance. I decided to start logging the average temperature, measured over three benchmark runs. System and ambient temperatures were normalized following each run, so this took a great deal of time for each cooler (especially as this was three benchmark runs, at each load level).

On the charts, if the CPU hits the thermal limit (100 C), the test would be stopped and the time it took to reach 100 C was logged. Average tests could not be completed due to this, and no testing at higher load levels was performed.

While I’m certain all of this is far from perfect, I feel like the final methodology I’ve come up with is a very comprehensive way for measuring a coolers capabilities.

One other note is that I would like to address is that 42 dB may seem high for a controlled sound level test, but this was measured on an open test bench, at a very short distance (20 cm). Even on the open test bench, at 42 dB, all of these coolers were practically inaudible from 3 feet.

I am hoping that the below explanation of the actual process is clear.

• Sound measurements are taken with a calibrated Dayton Audio UMM-6 measurement microphone, utilizing REW sound pressure level software.
• The Microphone is placed 20 cm (7.87 inches) from the edge of the cooler to minimize ambient sound contamination of the measurements, and fan speed was adjusted until a steady 42.0 dB was achieved
• Sound pressure level of each cooler is measured using the same parameters as above
• The system is loaded with a 10 minute Blender image render
• Each cooler is measured at loads of 100, 150, 200, 250, and 275 watts, at both 42dB and 100% fan speed
• At each fan speed, the test is run three times at each load level to determine the Average temperature
• Average temperature is measured only during the benchmarks’ run cycle
• Peak temperature is the highest temperature measured during any of the runs, and is not an average
• All tests conducted at a controlled ambient temperature of 23.5 C

Test System Specs:

• Intel 12900KS (used as CPU and Video Output to eliminate a GPU as an external heatsource)
• Gigabyte Z690 Aorus Master
• 32GB (2x16GB) G.Skill Trident Z5 DDR5-6400 Memory
• Sabrent Rocket 4 500 GB NVMe SSD
• Arctic Cooling MX-6 Thermal Compound

If you have looked at the accompanying charts, you know that is a lot of data, and I really wish I could have come up with a better way to present it. Spreadsheets are boring and bland, but for showing a lot of data, they function well.

When looking at the fan speed numbers, it is also important to go back to the fact that due to the Dark Rock Pro 5’s mismatched fan design, and the proprietary connector linking them, the reported fan speed is only for the middle, 135mm fan. The front, 120mm fan is spinning much faster in each instance.

At 42 dB, the fan speeds between the coolers are reasonably close, until the above factor is considered, then you realize that the front fan on the Dark Rock Pro 5 would be spinning somewhere around 1300 rpm, which is a much higher speed than the fan speeds of the competitors.

It is probably for this reason that, at this low noise level, the Dark Rock Pro 5 was the cooler to beat at every load level. At 250 watts the Dark Rock Pro 5’s average temperature was a full 2 degrees lower than its closest rival.

Almost without fail, that closest Rival was the Phantom Spirit 120 EVO, with the Peerless Assassin 120 ARGB right on its heels. All three of these full sized coolers were able to manage the 250 watt load, but failed at various intervals at 275 watts.

In the battle of the somewhat lower profile coolers, at 42 dB the two fan design of the Dark Rock TF2 easily outmatched the single 120mm fan on the Peerless Assassin Mini. One thing I did notice during these tests is that due to the top flow design of the TF2, it took much longer between tests to normalize the temperatures as the hot air just built up around the motherboard. This would only be a problem on an open test bench, or in a case with no exhaust fan positioned nearby, but I still thought it worth mentioning.

At max fan speeds, things start to get a little more interesting. The overall advantage has now shifted to the Thermalright Phantom Spirit 120 EVO, although the Dark Rock Pro 5 is almost equal, and the Peerless Assassin 120 is still very close behind.

These results were a good example of why I elected to measure both peak and average temperatures. The peak temperatures between the Phantom Spirit EVO and the Dark Rock Pro 5 were never more than 1 degree C different, and sometimes identical, however the average temperatures on the Phantom Spirit were always lower by up to 2 degrees.

One thing to consider here is that at this max fan speed, the Dark Rock Pro 5 was substantially quieter than either of the Thermalright dual tower models. Decibels are a logarithmic unit, and 3 dB is a doubling of sound pressure. It takes roughly about 10 dB for a human to believe a sound has doubled. The Dark Rock Pro 5 is not only quieter (imagine that) than the Thermalrights, but it has a more pleasant sound signature.

When looking at the max fan speed test results for the two smaller coolers, something interesting has occurred. The little, single fan Peerless Assassin Mini 120 has flipped the script and now outperforms the Dark Rock TF2. I will admit to some confusions as I was recording this data. It then occurred to me that the max speed of 1385 rpm on the TF2 just doesn’t flow enough air to dissipate the heat as well as the single 120 spinning at almost 1600 rpm, and pushing air through a much tighter fin array with more heatpipes.

Conclusion

We’ve got two coolers that either beat, or perform on par with the $90 be quiet! Dark Rock Pro 5, and either one can be had for under half that price. On the other hand, the Dark Rock Pro 5, even at max speed is extremely quiet and performs really great. The be quiet! level of fit and finish and quality are still on another level, but the Phantom Spirit 120 EVO from Thermalright is closing that gap, and beats the Dark Rock in pure thermal performance. It’s also going to be compatible with more cases.

Seriously??!! What are you still doing here? Go order the Phantom Spirit 120 EVO before the price goes up again.