Round four of
MacNN's SSD testing is complete -- and this round will likely wrap up the testing, for reasons we will discuss within. In this test, we again took four SSDs requested by users to task, including an OWC model, to see how well they all did in a time-compressed environment with Apple's OS X 10.10.4 and 10.10.5 Trim implementation.
History
We covered what the Trim function is in the
previous article which launched this series. Up until the OS X 10.10.4 update, non-Apple drives had to rely on a sometimes dangerous hack to implement the cleanup feature. With 10.10.4 Apple has implemented a way to enable Trim for third-party SSDs, without resorting to the hacks. However, prior to executing the terminal command to turn the feature on,
Apple issues a stern warning, making it perfectly clear that the risk of data loss falls squarely in the hands of the user.
There are two types of Trim -- queued, and sequential. We're testing sequential Trim, as implemented by the drive's firmware. Our
first round of testing found no issues with the aforementioned sequential trim as implemented by Apple in the Samsung 840 and 850 evo and pro series.
Testing Gear
We tortured a Samsung 850 evo and pro m.2 SSD on the rig, in conjunction with an OWC Extreme Pro 240GB SSD. The testing hardware is a 2012 i7 Mac mini that was running OS X 10.10.4 and updated to OS X 10.10.5 during the testing process. Connected to the Mac mini to support the drives are a pair of RocketStor 5212 Thunderbolt SATA docks. All of the AC adapters for the hardware are connected to an uninterruptible power supply, to prevent power outages or brownouts from effecting the testing process or inducing any SSD failure.
Limitations of the test
We're aware that the test isn't perfect. It is an accelerated process, where we're artificially loading and unloading the SSDs to implement the idle Trim feature faster than normal use. Our use pattern on the drives is far heavier than normal OS drive usage, and less sporadic -- a SSD with an installed OS is "tickled" a lot, for virtual memory, OS file access, app loading, and the like. What we're doing is more akin to constant use of the drive for a Photoshop scratch disk.
What this test is not is a comprehensive test of Trim's reliability under OS X and with all SSDs ever manufactured. Results for this round specifically apply only to the Samsung m.2 SSDs and the OWC SSD that we tested in this round. This test could have been performed under Windows as well, as the real issue here is the drive series' firmware routines, and less about any implementation by Apple.
Protocol
Transferred files vary in size from 64KB through 6.3GB, adding up to 150GB per transfer, lessened this round because of the smaller size of the drives on the test rig. The drive copies were set up "round robin," progressing through drives one through three, one copy at a time. When a copy option off the drive was completed, the data was deleted through a normal Finder move to trash and empty trash routine.
Drives sat completely idle between transfer rounds other than OS indexing, giving each drive as much time idle as both reading and writing, and more than enough time for Trim collection to begin between operations. At the conclusion of the test, each drive has seen over 1PB of data copied to the drive, and just as much read.
Conclusions from round four
We didn't see any data corruption in the process, related to cell death or Trim data loss at all. We have seen some cell failures, but as with previous testing, all of the cell failures are attributable to "age" which we've artificially sped up for this test. Importantly, we've put several years of use on these drives in a very short period of time. We're very pleased with the OWC drive, which is reporting significantly less in the way of cell failures than we've seen with any of the other drives we've tested, with 200GB more data copied to it than all the others.
The future of Trim and MacNN SSD testing
El Capitan is upon us. We've looked at it in the still beta (but golden master) OS, and the warning about potential data loss while implementing Trim from the terminal is gone. We're reasonably certain at this point that Apple engineers looked at warnings about queued Trim in data centers under Linux, and responded with the warning in Yosemite out of an abundance of caution. So, this round of Trim testing concludes our investigation into the issue that wasn't.
However, we're putting all these SSDs freshly loaded up with our test suite of data into our heat chamber,
as discussed in August. If you're not familiar, we've started real-life testing of unpowered, depleted SSDs, exposed to high temperatures for extended periods of time to check on both data retention, as well as drive survivability.