Apple was an early adopter of PCIe SSDs, introducing them in 2013 when the NVMe specification was still in its infancy and before any NVMe hardware was available. Apple's earliest PCIe SSDs used the AHCI protocol for compatibility with existing operating systems, but hardware compatibility was a very different story. Apple's PCIe SSDs used a proprietary form factor rather than the M.2 standard that went on to become the standard for client PCIe SSDs. So even though Apple's machines from 2013 through at least 2016 (depending on the model) included the fastest storage that money could buy at the time, those systems have been left behind as the NVMe storage market has matured from an exotic high-end novelty into the technology that's rapidly displacing SATA for mainstream computing.

This is where Mac accessory and upgrade specialist Other World Computing (OWC) comes in. OWC has offered several aftermarket SSDs in Apple's custom not-quite-M.2 form factor, culminating in the recent release of the Aura Pro X2 SSD. This is a modern high-end SSD with 3D TLC NAND and the latest Silicon Motion SM2262EN controller, with the reference M.2 PCB layout adjusted to fit Apple's form factor. The Aura Pro X2 is sold either as a bare drive, or in an upgrade kit that includes an external USB enclosure for the Apple original SSDs it replaces.

The Aura Pro X2 isn't OWC's first attempt to offer an upgrade for Apple PCIe SSDs, but it's the first one that does the job well. Their first Aura SSD hit the market back when Mac OS X didn't include a standard NVMe driver, so the Aura had to present a standard AHCI interface. Rather than use an outdated AHCI PCIe SSD controller comparable to the ones in the early Apple PCIe SSDs, OWC put two SATA SSD controllers and a RAID controller onto one card. This allowed OWC to provide a functional drop-in replacement that could offer higher capacities, but it was a big step backward in performance (and probably power efficiency, but we didn't get the chance to test it).

Apple eventually added a standard NVMe driver to MacOS, albeit after retiring upgradable internal storage from almost all of their product line. OWC responded with the Aura Pro X SSD, based on Micron 32-layer 3D MLC NAND and the Silicon Motion SM2260 NVMe controller. Unfortunately, Micron's first-generation 3D NAND and Silicon Motion's first-generation NVMe controller were both disappointing performers, so the Aura Pro X was again not a clear upgrade over the Samsung-based Apple original SSD.

Micron and Silicon Motion have since fixed their performance issues and the current generation of SSDs with 64-layer Micron 3D TLC and Silicon Motion SM2262(EN) controllers are serious competitors at the high end of the consumer SSD market, and a big step up from anything that was available in the 2013-2015 time frame. With the Aura Pro X2, OWC can now offer performance and capacity far beyond what Apple's factory-installed SSDs could provide.

The downside to the OWC Aura Pro X2 is that as it's a niche product, retail pricing is well above commodity M.2 SSDs. The Aura Pro X2 currently starts at 26¢/GB, when M.2 SSDs with the same hardware are retailing for just over half that price. The price disparity is even worse at 2TB, which may be the most important capacity for the Aura Pro X2 since Apple never offered a 2TB SSD in this form factor.

HP EX950 and adapter compared to OWC Aura Pro X2

OWC doesn't have a complete lock on this upgrade market. The Apple PCIe SSD form factor is a bit longer than M.2, so it's possible to use a standard M.2 NVMe SSD and a dirt-cheap passive adapter. In Apple's laptops, these adapters are just a hair too thick, so closing the machine back up completely leaves the bottom panel bulging slightly and puts pressure on the adapter and SSD connector. These connectors weren't designed to bear the weight of the machine, so there's some risk of a mechanical failure leading to an unreliable connection. However, I've been using one in my personal 13" MacBook Pro for several months with no trouble so far other than a bit of creaking in the bottom panel when there's too much pressure near the SSD.

OWC Aura Pro X2 Specifications
Capacity		240 GB	480 GB	1 TB (960GB)	2 TB (1920GB)
Form Factor		Apple custom, double-sided
Interface		NVMe 1.3 PCIe 3.1 x4
Controller		Silicon Motion SM2262EN
NAND		IMFT 64-layer 3D TLC
Sequential Read		2989 MB/s	3282 MB/s	3194 MB/s	3194 MB/s
Sequential Write		1208 MB/s	2432 MB/s	2488 MB/s	2488 MB/s
Power	Active	5.7 W
	Idle	0.3 W
Endurance		150 TB 0.34 DWPD	225 TB 0.27 DWPD	450 TB 0.27 DWPD	900 TB 0.27 DWPD
Warranty		5 years
Retail Price (drive only)		$109.99 (46¢/GB)	$159.99 (33¢/GB)	$249.99 (26¢/GB)	$599.99 (31¢/GB)

Since macOS has supported standard NVMe drives for over a year and a half, there's no need for the OWC Aura Pro X2's hardware to differ from typical M.2 SSDs in any way other than the physical form factor. The specifications are basically what we expect from a typical high-end M.2 NVMe SSD in today's market, but OWC's sequential IO ratings are a bit lower than the most optimistic numbers we see for M.2 drives. The 5-year warranty and ~0.3 DWPD endurance rating are normal for high-end drives. Aside from the form factor and price, the only thing that really stands out in the spec table is that OWC is using more overprovisioning than the the other SM2262EN drives we've encountered: 960GB for our review sample rather than the 1000GB or 1024GB usable capacities we've previously tested.

The layout of the Aura Pro X2 is very similar to M.2 SSDs with the SM2262EN controller. Apple's form factor is the same 22mm width as M.2 SSDs, and the extra 9mm length doesn't provide enough space to move around the major components, though some of the smaller passives have been rearranged. We're still looking at a double-sided drive, with two NAND packages and one DRAM package on each side. The Apple original SSDs we have on hand are actually more crowded with four NAND packages on each side, but the 64Gb per-die capacity of the Samsung 19nm MLC Apple was using back then is a far cry from the 256Gb TLC dies that now dominate the market.

Aura Pro X2 SSD Compatibility
	Supported Models
MacBook Pro	Late 2013 to 2015 (MacBookPro11,x - 12,x)
MacBook Air	2013 to 2017 (MacBookAir6,x - 7,x)
Mac mini	2014 (Macmini7,1)
Mac Pro	2013 (MacPro6,1)

As a Mac-specific SSD, the OWC Aura Pro X2 demands a different testing procedure from our usual mix of Windows and Linux based test. We've still run it through our usual tests by putting the Aura Pro X2 in an adapter that lets it fit in M.2 slots on our normal desktop testbeds. We've also ported our Linux-based synthetic tests over to macOS with a few changes, and tested the Aura Pro X2 in two different MacBook Pro machines.

For those macOS tests, we're comparing against two Apple original SSDs and several current high-end M.2 NVMe SSDs used in an adapter. The Apple original SSDs are both Samsung designs, similar to their XP941 and SM951 OEM M.2 drives. The latter drive uses the same UBX controller as the Samsung 950 Pro, the first retail M.2 NVMe SSD, while the older Apple drive uses the UAX controller that only supports PCIe 2.0 speeds. Samsung is still using MLC NAND in their top of the line 970 PRO (more than two generations removed from the Apple SSDs), while the rest of the market has concluded that 3D TLC NAND is fast enough and much more affordable.

The newer SSDs included in this review all have a capacity advantage over the 128GB and 512GB Apple SSDs we are comparing against. In general, larger drives are faster because they have more NAND flash memory dies to use in parallel. However, since the older Apple SSDs use NAND with a much lower per-die capacity, they aren't as handicapped as their total capacity might suggest. Most of the performance improvements the newer SSDs provide come from controller improvements and from using NAND that is fundamentally faster on a per-die basis.

macOS Performance

Our AnandTech Storage Bench trace-based tests are Windows-only, and running them on a Mac in a Boot Camp configuration wouldn't give us much insight beyond running the same ATSB tests on our usual desktop testbed. However, our suite of synthetic benchmarks uses the cross-platform FIO tool, so these tests could be easily adapted to run under macOS.

Our usual configuration for these tests has the drive under test as a secondary drive, with nothing on the drive other than the data written by the test itself. This is the best way to measure a drive's raw performance capability without any interference and minimal overhead, but that's not usually a very realistic configuration. Without our ATSB tests to give a more real-world picture to counterbalance the idealized synthetic benchmark results, we chose to reconfigure the synthetic tests to use a somewhat more realistic configuration. Instead of running the OS off a separate drive, the drives under test were used as the only storage attached to the system. A clean install of MacOS 10.14 Mojave was used on each drive, and instead of having the tests access the drive as a raw block device, the tests were run against ordinary files residing on the same APFS filesystem as macOS and the testing software.

The amount of data read or written by each test is the same as for our usual Linux-based synthetic benchmark results. However, with the overhead of a copy-on-write filesystem and background traffic from the OS getting in the way, and with an entirely different host operating system, the numbers below are not at all comparable to our previous results on our standard desktop testbed.

These tests were run on two different MacBook Pro systems: a base model 13" Retina MacBook Pro from 2015 with a dual-core Broadwell processor, and a high-end 15" Retina MacBook Pro (Late 2013 model) with a quad-core Haswell processor. The only attempt to tune the machines for performance was changing two sysctls to allow more threads to be used to handle asynchronous IO—necessary for testing higher queue depths, but not needed for real-world low queue depth workloads.

Random Read Performance

Our first test of random read performance uses very short bursts of operations issued one at a time with no queuing. The drives are given enough idle time between bursts to yield an overall duty cycle of 20%, so thermal throttling is impossible. Each burst consists of a total of 32MB of 4kB random reads, from a 16GB test file. The total data read is 1GB.

Consistent with our usual Linux-based testing, the two drives with Silicon Motion SM2262EN controllers (OWC Aura Pro X2 and HP EX950) offer the best burst random read performance. The WD Black SN750 fares poorly, ending up slower than the more recent Apple OEM drive but outperforming the older PCIe 2.0 Apple OEM drive.

Our sustained random read test adds higher queue depths to the score: queue depths from 1 to 32 are tested, and the average performance across QD1, QD2 and QD4 is reported as the primary score. Each queue depth is tested for one minute or 32GB of data transferred, whichever is shorter. After each queue depth is tested, the drive is given up to one minute to cool off so that the higher queue depths are unlikely to be affected by accumulated heat build-up. The individual read operations are again 4kB, and are taken from a 64GB test file.

On the longer random read test, the Aura Pro X2 remains one of the fastest SSDs, and the Samsung 970 PRO is essentially tied. The HP EX950's performance was inconsistent between the two laptops.

OWC Aura Pro X2 960GBSamsung 970 PRO 1TBHP EX950 1TBApple SM0128GApple SM0512FWD Black SN750 1TB

At sufficiently high queue depths, the HP EX950 and Samsung 970 PRO develop small performance leads over the Aura Pro X2. Additionally, the 13" MacBook Pro is severely bottlenecked by its weaker CPU that is unable to keep the SSDs busy under these conditions.

Random Write Performance

Our test of random write burst performance is structured similarly to the random read burst test, but each burst is only 4MB and the total test length is 128MB. The 4kB random write operations are distributed over a 16GB test file, and the operations are issued one at a time with no queuing.

The modern NVMe drives all provide similar performance on this burst random write test, while the older Apple OEM drives are at a clear disadvantage but aren't unusably slow. The 15" rMBP has significantly better performance than the 13" with the modern drives, but both machines offer roughly similar performance with the original Apple SSDs.

As with the sustained random read test, our sustained 4kB random write test runs for up to one minute or 32GB per queue depth, covering a 64GB test file and giving the drive up to 1 minute of idle time between queue depths to allow for write caches to be flushed and for the drive to cool down.

The results for the longer random write test are pretty similar to the burst random write test above. The modern NVMe drives all perform about the same, and the higher CPU power of the 15" machine compared to the 13" makes a bigger difference for these fast drives than it did for the original Apple SSDs.

OWC Aura Pro X2 960GBSamsung 970 PRO 1TBHP EX950 1TBApple SM0128GApple SM0512FWD Black SN750 1TB

Random write doesn't improve with higher queue depths under these test conditions. Instead, it is either flat or declining as the OS has to juggle more threads to issue IO, and there may be some locking somewhere in the filesystem and storage stack that's preventing the kind of scalability we see on Linux.

Mixed Random Performance

Our test of mixed random reads and writes covers mixes varying from pure reads to pure writes at 10% increments. Each mix is tested for up to 1 minute or 32GB of data transferred. The test is conducted with a queue depth of 4, and is limited to a 64GB test file. In between each mix, the drive is given idle time of up to one minute so that the overall duty cycle is 50%.

The Aura Pro X2's performance on the mixed random I/O test is comparable to the other recent TLC drives but a bit slower than the Samsung 970 PRO. The modern NVMe drives are about twice as fast as the Apple originals on the 15" MacBook Pro, but on the more CPU-limited 13", they Aura Pro X2 is only about 50-60% faster.

OWC Aura Pro X2 960GBSamsung 970 PRO 1TBHP EX950 1TBApple SM0128GApple SM0512FWD Black SN750 1TB

The recent NVMe SSDs all generally show performance increasing as workload becomes more write-heavy, but the 13" MBP's CPU bottleneck cuts off that growth for most of the second half of the test. The Apple originals show relatively flat performance across most of the test, but the older 512GB drive has a bit of the uptick at the end that we typically expect.

Sequential Read Performance

Our first test of sequential read performance uses short bursts of 128MB, issued as 128kB operations with no queuing. The test averages performance across eight bursts for a total of 1GB of data transferred from a 16GB test file. Between each burst the drive is given enough idle time to keep the overall duty cycle at 20%.

The OWC Aura Pro X2 and other modern drives all provide a bit less than 1GB/s on the burst sequential read test, as does the more recent of the two Apple OEM SSDs, while the older Apple drive delivers worse performance than we now expect from a low-end SATA SSD.

Our test of sustained sequential reads uses queue depths from 1 to 32, with the performance scores computed as the average of QD1, QD2 and QD4. Each queue depth is tested for up to one minute or 32GB transferred, from a 64GB test file.

On the longer sequential read test with some higher queue depths, the Aura Pro X2 is again more or less tied for first place, and the WD Black SN750 is slightly slower than the other modern drives. The older Apple SSD is still markedly slow, but the gap is down to roughly a factor of two rather than three. Differences between the two laptops are for the most part much smaller than they were for the burst sequential read test.

OWC Aura Pro X2 960GBSamsung 970 PRO 1TBHP EX950 1TBApple SM0128GApple SM0512FWD Black SN750 1TB

Most of these SSDs offer full sequential read performance at QD2, but the more recent Apple SSD needs higher queue depths, especially on the 13" laptop. Several of the drives (including the OWC) show a bit of a decline in performance at the highest queue depths where the OS has many active threads processing IO requests.

Sequential Write Performance

Our test of sequential write burst performance is structured identically to the sequential read burst performance test save for the direction of the data transfer. Each burst writes 128MB as 128kB operations issued at QD1, for a total of 1GB of data written to a 16GB test file.

The modern NVMe drives all perform similarly on the burst sequential write test. The OWC Aura Pro X2 seems to be the slowest of the four, though not by a big enough margin to worry about. The older of the two Apple SSDs provides sub-SATA performance, while the more recent but smaller drive is a bit faster than SATA drives are capable of.

Our test of sustained sequential writes is structured identically to our sustained sequential read test, save for the direction of the data transfers. Queue depths range from 1 to 32 and each queue depth is tested for up to one minute or 32GB, followed by up to one minute of idle time for the drive to cool off and perform garbage collection. The test is confined to a 64GB test file.

On the longer sequential write test with some higher queue depths, the Aura Pro X2 ends up slightly faster than the other SM2262EN-based drive (the HP EX950), but the differences between the modern NVMe drives are still small compared to their lead over the Apple original drives.

OWC Aura Pro X2 960GBSamsung 970 PRO 1TBHP EX950 1TBApple SM0128GApple SM0512FWD Black SN750 1TB

Most of the drives hit full speed at QD2 and provide steady performance for the rest of the test. The HP EX950 is the least consistent, especially on the older Haswell MacBook Pro.

Mixed Sequential Performance

Our test of mixed sequential reads and writes differs from the mixed random I/O test by performing 128kB sequential accesses rather than 4kB accesses at random locations, and the sequential test is conducted at queue depth 1. The range of mixes tested is the same, and the timing and limits on data transfers are also the same as above.

The OWC Aura Pro X2 runs into serious and surprising trouble on the mixed sequential IO test, and it affects the test runs on both of the MacBook Pros used. The overall average performance ends up being slightly worse than the older Apple original SSD, and about a third of what the HP EX950 (slowest of the three modern M.2 drives) scores.

OWC Aura Pro X2 960GBSamsung 970 PRO 1TBHP EX950 1TBApple SM0128GApple SM0512FWD Black SN750 1TB

Looking closely at what happened to the Aura Pro X2 during the mixed sequential IO test, at either end of the test when the workload is pure reads or writes it is competitive with the other modern NVMe drives, but when reads and writes are interleaved it falls apart. Performance for 90% reads isn't too bad and still beats the Apple original SSDs, but for all the other mixes the Aura Pro X2 is a serious step backward.

Test Setup

The rest of the tests in this review were conducted on our regular desktop testbed, with the OWC Aura Pro X2 installed in an M.2 adapter. As with our usual SSD reviews, these tests run on Windows (ATSB tests) and Linux (synthetic benchmarks) rather than macOS. The older Apple SM0512F SSD is included because it presents a standard AHCI interface that is software-compatible with SATA controllers, but the more recent Apple SM0128G SSD uses a non-standard protocol and cannot be properly detected on non-Apple systems even with the adapter that works for the SM0512F and OWC Aura Pro X2.

Since the rest of these test results are directly comparable to our usual review results, we've thrown in older numbers for a few more SSDs, including two entry-level NVMe SSDs: the Phison E8-based Kingston A1000, and the Intel 660p QLC-based SSD.

AnandTech 2018 Consumer SSD Testbed
CPU	Intel Xeon E3 1240 v5
Motherboard	ASRock Fatal1ty E3V5 Performance Gaming/OC
Chipset	Intel C232
Memory	4x 8GB G.SKILL Ripjaws DDR4-2400 CL15
Graphics	AMD Radeon HD 5450, 1920x1200@60Hz
Software	Windows 10 x64, version 1709
	Linux kernel version 4.14, fio version 3.6
	Spectre/Meltdown microcode and OS patches current as of May 2018

• Thanks to Intel for the Xeon E3 1240 v5 CPU
• Thanks to ASRock for the E3V5 Performance Gaming/OC
• Thanks to G.SKILL for the Ripjaws DDR4-2400 RAM
• Thanks to Corsair for the RM750 power supply, Carbide 200R case, and Hydro H60 CPU cooler
• Thanks to Quarch for the HD Programmable Power Module and accessories
• Thanks to StarTech for providing a RK2236BKF 22U rack cabinet.

Whole-Drive Fill

This test starts with a freshly-erased drive and fills it with 128kB sequential writes at queue depth 32, recording the write speed for each 1GB segment. This test is not representative of any ordinary client/consumer usage pattern, but it does allow us to observe transitions in the drive's behavior as it fills up. This can allow us to estimate the size of any SLC write cache, and get a sense for how much performance remains on the rare occasions where real-world usage keeps writing data after filling the cache.

OWC Aura Pro X2 960GBHP EX950 1TBIntel SSD 660p 1TBSamsung 970 EVO Plus 1TBSamsung 970 PRO 1TBSilicon Power P34A80 1TBWD Black SN750 1TBApple SM0512F (500GB)

The 960GB OWC Aura Pro X2's SLC cache is plenty fast, and lasts for about 147GB of writes before performance starts to drop. Initially, it goes down to about 850 MB/s, but just before the 600GB mark it drops again to be only slightly faster than SATA. Performance recovers a bit through this last phase, and ends up almost back up to the respectable second phase speeds. Overall, this behavior is similar to the HP EX950 that uses the same controller, but the EX950 tends to be a bit faster overall.

Average Throughput for last 16 GB	Overall Average Throughput

The overall average write speed puts the Aura Pro X2 as only slightly faster than the Apple 500GB drive, and half the speed of the fastest modern TLC drive. But this obscures the fact that the Apple drive doesn't have an SLC cache and never gets much above 800 MB/s during the fill, while the Aura Pro X2 writes at nearly 2.5GB/s for any ordinary real-world duration.

AnandTech Storage Bench - The Destroyer

The Destroyer is an extremely long test replicating the access patterns of very IO-intensive desktop usage. A detailed breakdown can be found in this article. Like real-world usage, the drives do get the occasional break that allows for some background garbage collection and flushing caches, but those idle times are limited to 25ms so that it doesn't take all week to run the test. These AnandTech Storage Bench (ATSB) tests do not involve running the actual applications that generated the workloads, so the scores are relatively insensitive to changes in CPU performance and RAM from our new testbed, but the jump to a newer version of Windows and the newer storage drivers can have an impact.

We quantify performance on this test by reporting the drive's average data throughput, the average latency of the I/O operations, and the total energy used by the drive over the course of the test.

The OWC Aura Pro X2 performs about the same on The Destroyer as the other SM2262EN-based drive, the HP EX950. These are both fairly slow compared to other current high-end NVMe SSDs, but almost twice as fast as the early Apple PCIe SSD.

The average latency for the Aura Pro X2 on The Destroyer is in line with expectations, but the 99th percentile latency is far higher than the HP EX950 and the older Apple SSD.

The OWC Aura Pro X2 shows more differences from the other SM2262EN drive when the average latency is broken down by reads and writes. For reads, the OWC drive is significantly faster than the HP EX950 and is comparable to the Phison E12-based Silicon Power drive. For writes, the OWC is slower than the EX950 but still well ahead of the Apple SSD and the current entry-level NVMe drives.

The 99th percentile read latency of the Aura Pro X2 on The Destroyer is competitive with other current high-end NVMe drives, but the 99th percentile write latency is a problem: it's a bit worse than the MLC-based Apple SSD, and several times higher than the best current TLC drives.

The OWC Aura Pro X2 is more power efficient than expected, using less energy to complete The Destroyer than most other drives in this batch, while the Apple SSD and the HP EX950 are some of the most power-hungry under load.

AnandTech Storage Bench - Heavy

Our Heavy storage benchmark is proportionally more write-heavy than The Destroyer, but much shorter overall. The total writes in the Heavy test aren't enough to fill the drive, so performance never drops down to steady state. This test is far more representative of a power user's day to day usage, and is heavily influenced by the drive's peak performance. The Heavy workload test details can be found here. This test is run twice, once on a freshly erased drive and once after filling the drive with sequential writes.

The average data rates for the OWC Aura Pro X2 on the Heavy test are not competitive with other current high-end NVMe drives, but at least it avoids the horrible full-drive performance we've seen from other SM2262EN drives. And it is still substantially faster than the older Apple SSD, for both full and empty drive test runs.

The 99th percentile latency problems with the Aura Pro X2 show up again on the Heavy test, but these would still be reasonable scores for a SATA SSD; it doesn't suffer like a full Intel 660p. Average latency is sub-par for what should be a high-end NVMe SSD, but is still an improvement over the older Apple drive and the current entry-level NVMe drives.

The average read and write latencies for the Aura Pro X2 are both a clear improvement over the Apple SSD but are nothing special compared to high-end M.2 NVMe SSDs.

The OWC Aura Pro X2 has competitive QoS for read operations when the Heavy test is run on an empty drive, but when full the 99th percentile read latencies degrade to entry-level NVMe performance. The 99th percentile write latencies are poor for both test runs.

The Aura Pro X2 again ends up with pretty good power efficiency, coming close to the WD Black SN750 that sets the standard to beat for high-end NVMe drives. The Apple SSD stands out with much higher energy consumption than even the most power-hungry of the modern high-end M.2 drives, and to complete the Heavy test it requires more than twice the energy that the OWC drive uses.

AnandTech Storage Bench - Light

Our Light storage test has relatively more sequential accesses and lower queue depths than The Destroyer or the Heavy test, and it's by far the shortest test overall. It's based largely on applications that aren't highly dependent on storage performance, so this is a test more of application launch times and file load times. This test can be seen as the sum of all the little delays in daily usage, but with the idle times trimmed to 25ms it takes less than half an hour to run. Details of the Light test can be found here. As with the ATSB Heavy test, this test is run with the drive both freshly erased and empty, and after filling the drive with sequential writes.

On the Light test, the OWC Aura Pro X2 again performs more like an entry-level NVMe SSD, but unlike the Heavy test there isn't a big gap between tiers of NVMe SSDs. The Aura Pro X2 clearly outperforms the older Apple SSD, even in the worst-case scenario of a full drive.

The average latency of the Aura Pro X2 during the Light test is a bit worse than most high-end NVMe SSDs, but isn't high enough to worry about. The 99th percentile latency is rather high for the worst-case test run on a full drive, but this 2ms is only marginally slower than the old Apple drive and is not even the worst we've seen from a SM2262EN drive.

The average read latencies for the Aura Pro X2 on the Light test are competitive with high-end M.2 NVMe drives, though the latency for the full-drive test run is a bit high. The average write latency is clearly higher than typical for high-end NVMe drives, whether the test is run on a full or empty drive.

The high-end NVMe drives almost all have extremely low 99th percentile write latencies on the Light test, and the Aura Pro X2 can't match that performance even when the test is run on an empty drive. For reads, the 99th percentile latency is competitive when the test is run on an empty drive, and is on par with the old Apple SSD when full.

The OWC Aura Pro X2 saves a lot of power relative to the old Apple SSD or the fastest current NVMe SSDs. It doesn't quite match the WD Black SN750's efficiency, but it's better than we expected from a drive with the SM2262EN controller.

Power Management Features

Real-world client storage workloads leave SSDs idle most of the time, so the active power measurements presented earlier in this review only account for a small part of what determines a drive's suitability for battery-powered use. Especially under light use, the power efficiency of a SSD is determined mostly be how well it can save power when idle.

For many NVMe SSDs, the closely related matter of thermal management can also be important. M.2 SSDs can concentrate a lot of power in a very small space. They may also be used in locations with high ambient temperatures and poor cooling, such as tucked under a GPU on a desktop motherboard, or in a poorly-ventilated notebook.

OWC Aura Pro X2 NVMe Power and Thermal Management Features
Controller	Silicon Motion SM2262EN
Firmware	S0121C
NVMe Version	Feature	Status
1.0	Number of operational (active) power states	3
1.1	Number of non-operational (idle) power states	2
	Autonomous Power State Transition (APST)	Supported
1.2	Warning Temperature	75°C
	Critical Temperature	80°C
1.3	Host Controlled Thermal Management	Supported
	Non-Operational Power State Permissive Mode	Not Supported

The OWC Aura Pro X2 declares support for all the usual power management features expected on a modern M.2 NVMe SSD, with two idle states that  balance power savings against transition latency. The drive provides fairly conservative estimates for maximum power in its active power states – in practice, our synthetic tests didn't push it much beyond 4W.

OWC Aura Pro X2 NVMe Power States
Controller	Silicon Motion SM2262EN
Firmware	S0121C
Power State	Maximum Power	Active/Idle	Entry Latency	Exit Latency
PS 0	9.0 W	Active	-	-
PS 1	4.6 W	Active	-	-
PS 2	3.8 W	Active	-	-
PS 3	45 mW	Idle	2 ms	2 ms
PS 4	4 mW	Idle	15 ms	15 ms

Note that the above tables reflect only the information provided by the drive to the OS. The power and latency numbers are often very conservative estimates, but they are what the OS uses to determine which idle states to use and how long to wait before dropping to a deeper idle state.

Idle Power Measurement

SATA SSDs are tested with SATA link power management disabled to measure their active idle power draw, and with it enabled for the deeper idle power consumption score and the idle wake-up latency test. Our testbed, like any ordinary desktop system, cannot trigger the deepest DevSleep idle state.

Idle power management for NVMe SSDs is far more complicated than for SATA SSDs. NVMe SSDs can support several different idle power states, and through the Autonomous Power State Transition (APST) feature the operating system can set a drive's policy for when to drop down to a lower power state. There is typically a tradeoff in that lower-power states take longer to enter and wake up from, so the choice about what power states to use may differ for desktop and notebooks, and depending on which NVMe driver is in use. Additionally, there are multiple degrees of PCIe link power savings possible through Active State Power Management (APSM).

We report three idle power measurements. Active idle is representative of a typical desktop, where none of the advanced PCIe link or NVMe power saving features are enabled and the drive is immediately ready to process new commands. Our Desktop Idle number represents what can usually be expected from a desktop system that is configured to enable SATA link power management, PCIe ASPM and NVMe APST, but where the lowest PCIe L1.2 link power states are not available. The Laptop Idle number represents the maximum power savings possible with all the NVMe and PCIe power management features in use—usually the default for a battery-powered system but rarely achievable on a desktop even after changing BIOS and OS settings. Since we don't have a way to enable SATA DevSleep on any of our testbeds, SATA drives are omitted from the Laptop Idle charts.

Note: We recently upgraded our power measurement equipment and switched to measuring idle power on our Coffee Lake desktop, our first SSD testbed to have fully-functional PCIe power management. The below measurements are all new this month, and are not a perfect match for the older measurements in our previous reviews and the Bench database.

The OWC Aura Pro X2 is the first drive we've subjected to our updated idle power measurement test that seems to have a compatibility problem. The active idle and desktop idle numbers are in line with expectations and are comparable to other SM2262EN drives. When the lowest power PCIe ASPM features are enabled, the Aura Pro X2 is no longer able to stay at a low power level and instead jumps up to almost half its active idle.  This might be a side effect of the adapter we're using to get the drive working with standard M.2 slots.

The Apple SSD uses AHCI instead of NVMe, and none of the usual settings for manipulating power levels for SATA or PCIe drives seem to be of any use. Its active idle power draw is far higher than NVMe drives with modern controllers, and turning PCIe ASPM on makes it draw even more power. Apple almost certainly has non-standard ways to put this drive into a properly low-power state, but we aren't able to achieve this on our desktop testbed that is equipped to measure idle power.

The desktop idle state that works properly for the Aura Pro X2 shows moderately higher wake-up latency than other Silicon Motion drives, which are already some of the slowest NVMe drives when it comes to coming out of sleep states. When attempting to use the deepest PCIe ASPM idle settings, the Aura Pro X2 doesn't go to sleep and consequently has minimal wake-up latency, as does the Apple SSD for which we were unable to trigger any sleep states.

Conclusion

The OWC Aura Pro X2 is based on much newer technology than the Apple original SSDs it is intended to replace. In principle, this allows for not only higher capacities at lower prices, but also better performance and power efficiency. The older Macs that the Aura Pro X2 is designed for impose some performance limitations that modern machines don't experience, so in most real-world use cases the Aura Pro X2 isn't able to show off the full capabilities of its newer hardware.

Our macOS-based testing showed that the performance differences between modern NVMe drives are largely erased by bottlenecks elsewhere: filesystem overhead and the general inefficiency of performing asynchronous IO using kernel thread pools on low-power mobile CPUs with low core counts. In spite of these limitations, the Aura Pro X2 is consistently able to deliver better performance than the Apple original SSDs, especially for random IO. The differences in benchmark scores aren't always large enough to have a dramatic impact on real-world use, but the Aura Pro X2 is definitely faster overall. That's something that could not be said for OWC's earlier attempts to provide an upgrade in this form factor.

(from top: HP EX950 1TB, OWC Aura Pro X2, Apple SM0512F)

Putting the Aura Pro X2 in an adapter and testing it on our usual desktop testbed allowed us to dig into its power efficiency and explore its performance potential with fewer limitations from the host system, which may be more relevant to Mac Pro users than MacBook Pro users. We found that the Aura Pro X2 was generally slower than current high-end M.2 NVMe SSDs, though it typically still outperforms entry-level NVMe drives. Surprisingly, this lower performance enabled much better power efficiency than we've seen from other drives using the Silicon Motion SM2262EN controller, though the Aura Pro X2 isn't quite as efficient as the Western Digital WD Black SN750. High-end drives tend to sacrifice efficiency in an attempt to set benchmark records. That is pointless for the Aura Pro X2 that is intended for systems where the host CPU and OS will be the more significant bottleneck, so OWC made the right tradeoffs with this drive.

The only truly disappointing performance result was on the mixed sequential IO test under macOS, where the Aura Pro X2 was pathologically slow except with very read-heavy mixes and the pure read or write phases at the beginning and end of the test. In spite of this, the average across all the mixes we test was only slightly slower than the older Apple SSD. (This behavior was not evident when testing the Aura Pro X2 on our desktop testbed under Linux, so it seems this was due to a poor interaction between the drive and macOS/APFS.)

For users who have Apple's later PCIe SSD based on the Samsung UBX controller (also seen in the Samsung 950 PRO), upgrading to a newer drive like the OWC Aura Pro X2 won't bring any huge performance increases, but the improvements to power efficiency in newer SSD controllers and flash memory may help offset the battery life degradation in an aging notebook. The earlier Apple PCIe SSDs based on the Samsung UAX controller are distinctly slower than NVMe SSDs, but still outperform SATA drives and are fast enough for most use cases. Thus, the main selling point of the Aura Pro X2 is that it allows for a big capacity upgrade: Apple never offered a 2TB option in this form factor, and for some machines even 1TB wasn't an option when they were new. And Apple's build-to-order SSD upgrades have always been expensive even compared to the ridiculous prices most other OEMs charge.

For Mac mini and 2013 Mac Pro users, the obvious solution for a storage upgrade is to buy an adapter and use a much cheaper standard M.2 NVMe SSD. These machines are much smaller than typical desktops, but they still have room to spare for the extra height of an adapter. For the notebooks, an adapter can work, but it prevents the bottom panel of the case from being fully closed without bulging and putting pressure on the adapter itself. Which probably increases the odds of one of the connectors or solder joints breaking—these weren't designed to be load-bearing. For most users, this is probably an acceptable tradeoff for getting access to the much broader market for standard M.2 SSDs.

NVMe SSD Price Comparison (June 5, 2019)
	240-280GB	480-512GB	960GB-1TB	2TB
OWC Aura Pro X2	$109.99 (46¢/GB)	$159.99 (33¢/GB)	$249.99 (26¢/GB)	$599.99 (31¢/GB)
Silicon Power P34A80	$37.99 (15¢/GB)	$59.99 (12¢/GB)	$109.99 (11¢/GB)	$264.99 (13¢/GB)
ADATA XPG SX8200 Pro		$74.99 (15¢/GB)	$149.99 (15¢/GB)
HP EX950		$86.99 (17¢/GB)	$152.99 (15¢/GB)	$305.99 (15¢/GB)
Intel 660p		$61.99 (12¢/GB)	$99.99 (10¢/GB)	$194.99 (10¢/GB)
Samsung 970 EVO Plus	$69.99 (28¢/GB)	$117.99 (24¢/GB)	227.99 (23¢/GB)	$499.99 (25¢/GB)
Samsung 970 PRO		$159.99 (31¢/GB)	$332.99 (33¢/GB)
Western Digital WD Black SN750	$69.99 (28¢/GB)	$107.99 (22¢/GB)	$227.99 (23¢/GB)

The OWC Aura Pro X2 does not have any true direct competitors on the retail market. They also have a lot of leeway to charge a premium for these upgrade parts while still staying far below what Apple charges for build-to-order storage upgrades. But the availability of cheap adapters and even some SSDs bundled with an adapter means that the Aura Pro X2 is in competition with the broader M.2 NVMe SSD market.

Almost every M.2 NVMe SSD still in production beats the Aura Pro X2 in price; even the Samsung 970 PRO manages to just barely undercut OWC at 512GB for the same price as OWC's 480GB. The cheapest TLC-based high end drives such as the Phison E12-based Silicon Power P34A80 are less than half the price per GB of the OWC Aura Pro X2.

Even adding in $15-20 for the necessary adapter does nothing to change the story. The Aura Pro X2 is simply way too expensive. If OWC was providing their Envoy Pro USB enclosure for the Apple original SSDs bundled at these prices, then they would be closer to sanity, but the bundles are $60-80 more expensive than the bare drive prices shown above.

OWC has also recently introduced the Aura N, based on the entry-level Phison E8 controller platform. This is probably still plenty fast for use in older Macs and also tends to be more efficient than high-end NVMe SSDs. However, their pricing on the Aura N is so far only $20 cheaper than the Aura Pro X2 at best, so it really isn't at all competitive over M.2+adapter solutions either.