M.2 Drives Have An Overheating Problem

[SteelCrysis]

https://www.pugetsystems.com/labs/articl...lysis-776/

In an ideal situation, every component in the system is idle and producing a minimal amount of heat. In this situation, we are showing what should be the longest the Samsung 950 Pro can go without throttling due to drive temperature.

There is a lot of data crammed into the charts above, but the main thing to look at is how quickly the different slots start to throttle and what speeds they throttle to. The charts above can largely be summarized into the following:

It is clear that the "Upright M.2 Slot" takes the longest to start throttling and runs at consistently higher speeds than the other slots. In general, this slot took between ten and thirty seconds longer to start throttling and after throttling was up to 500 MB/s faster than some of the other slots! One thing we will point out is that the Random Read performance of this slot looks poor, but that is actually due to the X99 chipset and not due to the drive throttling.

For the other locations, the top and bottom slots were pretty similar. The bottom slot was better with only a single GPU but the top slot was actually slightly better with dual GPUs. However, both locations were better than the underside slot which ended up throttling to about half the performance the Samsung 950 Pro is capable of.

Worst case (heavy system load) results

While the previous section is a good indication of how soon a Samsung 950 M.2 drive might throttle in ideal situations, it is often the case that the system will be under a load when you need to access the drive. To simulate these situations, we loaded the GPU(s) with Furmark to increase the ambient temperature within the chassis. Note that we did not put a heavy load on the CPU because we wanted to avoid creating a CPU bottleneck that might cause lower results from our benchmark.

Compared to the ideal situation in the last section, loading up the video card(s) to increase the ambient temperature caused the M.2 drive to throttle much sooner. In the worst case, the drive was throttling after only 6 seconds and ended up slower than a standard SATA SSD! Once again, if we primarily look at the sequential read and write performance, the charts above can largely be summarized into the following:

Once again, the upright M.2 slot was by far the best slot position. This slot took significantly longer to start throttling than the other slots and especially with two GPUs it throttled much less aggressively. The performance of the Samsung 950 Pro was still cut roughly in half, but that is actually much better than what we saw with the other slot locations.

For the other locations, the bottom M.2 slot was the next best performer - taking roughly twice as long to start throttling than the top slot. It ended up at about the same speed as the top slot with a single GPU, but with two GPUs it ended up noticeably faster.

Conclusion

A lot of our test results really surprised us. We knew that M.2 drives would throttle if they got too hot, but we didn't know that we could cause it to happen so quickly or quite to the extent we measured. There are really three main conclusions that we came to:

1. The temperature reported by utilities such as AIDA 64 is not useful to gauge whether the drive is close to throttling or not. As we discussed in the test setup section, this is due to the fact that the Samsung 950 Pro actually has two thermal sensors, but only the sensor near the storage chips (not the control sensor) is being read. As far as we could find, it is not possible to read from the sensor that controls the drive throttling - but if anyone knows of a program that will read this sensor data, we are very interested to hear about it.
   
2. The ambient temperature within the system has a massive effect on when the drive throttles. It makes sense that a drive will throttle sooner in a hotter environment, but in the worst case we were seeing throttling in as soon as six or seven seconds!
   
3. The more exposed the M.2 drive, the longer it takes to throttle. Once again, this seems like a common sense point, but it ended up making a bigger difference than we thought. Especially when the system was put under a load, the upright M.2 slot (with the most open space around the drive) resulted in significantly less throttling than any of the other slots.
   

...
One thing we want to point out is that it is pretty uncommon to fully utilize a drive this fast to the same extent we did in this article. Very few programs will actually be able to read from a Samsung 950 Pro at full speed for more than a very short period of time - often the drive will have to regularly wait on the CPU or some other component to actually process the data. In truth, the most common situation you might run throttling is going to be if you have two M.2 drives and you are copying large amounts of data from one drive to the other.

Also, keep in mind that if you absolutely need a super fast drive to perform at full performance for longer periods of time, there are a number of alternatives you might consider. The first would be to use a U.2 drive instead like the Intel 750 series. As we showed in the control section, the Intel 750 did not throttle at all yet is very close in performance to the Samsung 950 Pro. Not many motherboards have U.2 support quite yet and the Intel 750 is a bit more expensive, but it should give you maximum performance all the time. Another solution would be to attempt to cool a M.2 drive with additional airflow or a heatsink in order to increase the time it takes to begin throttling. If you are interested in that option we performed further testing in our M.2 Drive Additional Cooling Testing article investigating a range of cooling methods.

We corresponded with Geoff Gasior from Asus throughout our testing and they were even kind enough to provide the motherboards we used for our testing. If you would like to view their take on the results, head over to their article:
M.2 SSD throttling tests vindicate ASUS motherboard designs

https://www.pugetsystems.com/labs/articl...sting-795/

In a recent article, we investigated how long it takes for a Samsung 950 Pro M.2 drive to throttle in a number of different M.2 slot locations and some of the results surprised us. In fact, in one instance we measured a 75% drop in performance after only 7 seconds!

If you are purchasing a M.2 drive because you need high transfer speeds, seeing our results probably threw up a big red flag for you. Having a high speed drive is great, but not so much if the speed only lasts a couple of seconds. Luckily, in most situations the Samsung 950 Pro drive actually lasts a good amount of time under full load before it begins to throttle. For those that need a M.2 drive to run at full speed for longer periods of time, however, we decided to run some quick tests to compare a number of different M.2 drive cooling methods.

If you prefer to skip all our testing and simply view our conclusions, feel free to jump ahead to the conclusion section.
...
Compared to just a stock Samsung 950 Pro M.2 drive without any additional cooling, every single method we tested did extremely well. Even in the worst case with the simple aluminum bar heatsink, the drive took 2.5 times longer before it started to throttle compared to the stock M.2 drive with no additional cooling. If we were to rate the different methods from most to least effective, they would be:

1. Tie between PCI-E Adapter w/ Heatsink, 120mm 12V Quiet Side Fan, 92mm 12V High Flow Side Fan. All three of these methods completely prevented the Samsung 950 Pro drive from throttling during our testing. If you fully load the drive for longer than we did (which would mean you need to read more than 875GB worth of data from a 512GB drive), the high flow side fan should perform better than the quiet side fan, but in a practical sense all three of these methods should effectively be able to prevent a Samsung 950 Pro drive from ever throttling
   
2. 120mm 5V Quiet Side Fan. While this cooling method was not able to completely prevent the drive from throttling when the system was under load, it allowed the drive to take 3-5 times longer to throttle and after throttling was 50% faster than a stock Samsung 950 Pro. For such a small amount of airflow, this is a much bigger difference than we expected and means you could read 455GB of data (nearly the entire drive) or write 172GB of data continuously before you saw any drop in performance.
   
3. Aluminum Bar Heatsink. Technically, this was the worst cooling method we tested but it was still a massive improvement over the stock drive without any cooling. The main downside to a simple heatsink like this is that the hotter the system gets, the less effective a heatsink can be. We only tested with a single GPU, but if you had two or more video cards under full load, it is very possible that a heatsink may be no better than a bare drive or in some situations may cause the drive to throttle even sooner.
   

The different methods we tested really boil down to two types of cooling: passive cooling with a heatsink and active cooling with a fan. Both can make a big difference, but one thing that was clear in our testing is that even a small amount of airflow over the drive can be extremely beneficial. While we did not specifically test it, even better would be to combine the two methods by having a heatsink on the drive along with a fan providing some airflow over the heatsink.

Keep in mind that in the real world, it is very uncommon to fully utilize a drive this fast to the same extent we did in our testing. Very few programs will actually be able to read from a Samsung 950 Pro at full speed for more than a very short period of time, but if you do have a situation where you need a M.2 drive to perform at full speed for longer periods of time this should give you an idea of what you may need to do to achieve this.

http://edgeup.asus.com/2016/06/m-2-ssd-t...d-designs/

Puget Systems did most of its testing on our Z170-WS, which has M.2 slots below the primary PCIe x16 mid-way up the board and below the secondary slot toward the bottom. They also tested the vertical M.2 slot on our X99 Deluxe, a unique configuration mirrored in the new X99-Deluxe II, and the underside slot on our H170I-Plus D3, a typical location for Mini-ITX motherboards with insufficient space topside.

The big news is that 950 Pro was unable to sustain its peak sequential speed in all four locations, regardless of whether the system was idling or occupied with a heavy graphics load. Throttling kicked in after as little as seven seconds of sustained testing or as much as 70. The declines were significant, too: from 40-70% with a single graphics card and 55-76% with two. Matt Bach, who conducted the testing and has previously demonstrated throttling on the 950 Pro and other Samsung M.2 SSDs, was particularly surprised by how quickly the new drive throttled in thermally challenging environments.
...
The underside slot has less airflow than any of the other positions. Predictably, the 950 Pro exhibited the most severe throttling in this location—and the biggest performance hit with a single graphics card under load. This is especially problematic because Mini-ITX motherboards typically don’t have room to put their M.2 slots anywhere else. The only other option is the PCI Express x16 slot, which is usually reserved for a graphics card

When designing our Maximus VIII Impact, we had to choose between an underside M.2 slot and a native U.2 port. Offering both would have required extra layers, inflating not only the cost of the board, but also the development time required to bring it to market. Since the Impact is all about bringing the best performance to Mini-ITX, we went with U.2 to avoid throttling issues that had been documented with other M.2 drives at the time. We caught some flak from people who wanted to build Mini-ITX rigs with M.2 drives, but Puget’s results support our decision. The 750 Series U.2 SSD showed no signs of temperature-induced throttling in any of the tests.

Frankly, we shouldn’t be surprised. The U.2 interface has four PCIe Gen3 lanes, just like current M.2 slots, but it uses a cabled connection that lets drives reside in standard cages, where they can benefit from natural chassis airflow. The larger form factor also leaves plenty of room for a heatsink to further aid cooling, and the 750 Series takes full advantage by integrating one into its shell.

We’ve long been aware of the benefits of U.2 storage, which is why native ports are integrated on several of our high-end motherboards, including the Maximus VIII Hero Alpha, Formula, and Extreme. The Z170-WS used in Puget’s testing has U.2 onboard, too, as do our latest X99-A II, X99-Deluxe II, Strix X99 Gaming, and Rampage V Edition 10. With the exception of the Maximus VIII Impact, all these boards also have M.2 slots.

If you still want to run an M.2 drive, the Puget’s companion post shows that a dedicated cooling fan or full-sized adapter card with heatsink can minimize throttling on the 950 Pro, if not eliminate it entirely. The company even sells the adapter, but Bach told us it tends to favor the 750 Series in desktop builds instead. The 950 Pro is used in notebooks, where there’s usually more ventilation—and no hot components in the immediate vicinity.

Serial ATA SSDs still outsell their PCIe counterparts by a huge margin—about 12:1 at Puget Systems—so it’s important to keep things in perspective. The fact is that SATA SSDs are fast enough for the vast majority of desktop applications. The one Puget System tested for throttling didn’t exhibit any, either.

Speaking of perspective, we should also reiterate that throttling was only evident under sustained loads. The 950 Pro takes some time to heat up, and outside of targeted benchmarks, most everyday storage loads are too short to trigger throttling. But we’d be critical of a CPU or GPU that behaved the same way, so it’s hard to give this story a positive spin. The best we can offer is a range of motherboards that situate their M.2 slots clear of potential hot spots, plus plenty of options with U.2 ports that let you avoid the problem entirely.