I intentionally overheated my NVMe SSD (and proved your cheap motherboard heatsink is fine)
NVMe SSD prices have skyrocketed, so it’s more important than ever to protect your expensive investment. The 2TB NVMe PCIe Gen 4 I bought for $100 is now literally triple the price. Lately, I’ve been wondering if the original NVMe heatsink that came with my motherboard was good enough to protect the drive from its worst enemy: heat. So I did some tests to find out.
Ignoring NVMe SSD temperatures is problematic
Heatsinks are not optional on modern SSDs
Modern NVMe SSDs are extremely fast. PCIe Gen 5 SSDs like the Samsung 9100 PRO can achieve read speeds of up to 14.7 GB/s and write speeds of up to 13.4 GB/s, which is absolutely mind-blowing when you think about it.
Even older, budget-friendly PCIe Gen 4 models with slower QLC NAND flash, like my Crucial P3 Plus, can achieve up to 5,000 MB/s sequential read and 4,200 MB/s sequential write.
All that speed generates heat, especially from the controller, which is the brains of the SSD that manages a whole list of processes.
So, to stay cool and avoid catastrophic failure, it is highly recommended to install a heatsink on your NVMe. This is a large piece of metal that, using thermal interface material (i.e. thermal pads), draws heat away from the controller to keep it cool.
Without it, even older, cost-effective NVMe SSDs risk premature failure caused by prolonged overheating. Additionally, heat causes NVMe to slow down, leaving a lot of potential performance on the table.
Fortunately, most modern gaming motherboards include at least one NVMe heatsink, and even if they don’t, you can easily get something like the ARCTIC M2 Pro or Thermalright TR-M.2 for a few dollars to keep your NVMe cool.
How I Deliberately Pushed My SSD to the Extreme
Stress tests are the only way to check how high temperatures can go
The only way to know if the original heatsink that came with my ASRock B650M PG Riptide could handle my NVMe SSD was to run a stress test and closely monitor the temperatures. Setting up the test is pretty simple, so if you want to replicate the test at home, here’s what you’ll do.
First, download and install CrystalDiskMark, which we will use for stress testing, as well as HWiNFO, one of the best free hardware monitoring tools.
To optimize CrystalDiskMark for testing, run the tool as administrator, set the number of tests to 9and change the test size to 64 GB. The high test size will overwhelm the data controller and force it to work very hard, while the high test count will ensure that the test runs long enough to be meaningful (around 10-20 minutes).
Then click Settings and click on the Settings in the submenu, then set the Interval time (sec) has 0 and click ALL RIGHT. Don’t skip this step; By default, CrystalDiskMark waits 5 seconds between each test to allow the NVMe to complete its internal processes and cool down slightly, which is the opposite of what we want.
To configure HWiNFO, launch the tool and select Sensors onlythen click To start. Scroll to CLEVERfollowed by the model name of your SSD and the drive letter as defined in Windows (for example, C:). If you don’t test your primary disk, just find the CLEVER for the one you are testing.
You’ll typically see three disk temperatures here, and Drive 2 temperature Or 3 is the controller: the temperature you should focus on. You will know what it is once the test starts because it will be much higher than the temperature of the NAND flash memory. On my device it’s Drive Temperature 3.
You need to configure HWiNFO to record your temperatures by right-clicking SMART and clicking Enable logging. After that you can click Start of logging (the paper with the plus sign) in the bottom right corner, name your test and choose where to save it. When you have completed the tests, you will click on the same button, which will now be called Stopping recording to complete the test.
This will create a CSV report, complete with all the sensor data which you can then use to create a line graph. Logging your temperatures is optional, as you can also just keep an eye on your current and maximum NVMe temperatures while the test is running.
When you are finished setting up HWiNFO, return to CrystalDiskMark and click “All” to begin testing while keeping an eye on the temperatures in HWiNFO.
The data is clear: even the stock motherboard’s NVMe heatsinks are pretty awesome
If you have it, use it
As you can see from my test, the data is clear: my NVMe controller never got close to the thermal throttling temperatures. The maximum temperature recorded during my test was just 68°C, which is lower than the 70°C figure that Kingston says is the maximum before the NVMe reduces its speed to reduce temperature buildup.
Granted, my NVMe SSD is a fairly inexpensive model that you might find in a typical gaming PC, but then again, that’s the majority of NVMe SSD buyers: how many gamers are wasting money on super-fast PCIe Gen 5 NVMe drives anyway?
Stop Buying PCIe 5.0 SSDs for Your Gaming PC
Doubling what you spend probably won’t double your performance.
Additionally, this was an extreme stress test designed to push the NVMe to its maximum. Temperatures don’t get close to maximum when playing.
So if you’re worried about your expensive NVMe SSD overheating while you’re gaming, you really don’t have to, as long as you have some sort of heatsink.
It’s a cheap and easy upgrade to make
While gamers like me don’t have to worry about NVMe temperatures, some people have legitimate use cases for super-fast NVMe PCIe Gen 5 due to the super-fast file transfer speeds. If you’re one of these people, you should consider upgrading to a more robust NVMe heatsink like the Thermalright HR-09 – and you definitely shouldn’t run your NVMe bare.
Stop Running Your NVMe “Bare”: Why Modern SSDs Need a Heatsink
Your SSD is hotter than it should be, and it’s only going to get worse
