Micron’s M600 solid-state drive reviewed

Flash memory is simultaneously shrinking and expanding. The actual memory cells are getting smaller thanks to advancements in fabrication technology, fueling an increase in density that enables higher-capacity chips. For the most part, that’s just great. Higher bit densities reduce the cost per gigabyte, while higher-capacity chips facilitate SSDs with more total storage.

Unfortunately, stuffing more gigabytes into each chip also presents a problem. SSDs now require fewer chips to hit the same capacity, and in some cases, they don’t have enough NAND components to fully exploit the parallelism in modern controllers. That’s why smaller SSDs typically have lower write performance ratings than larger ones.

Multiple SSDs use high-speed caches to make up for slower write performance, but none are quite as ambitious as Micron’s new M600. The memory giant’s latest contender can toggle any part of its NAND between SLC and MLC modes, allowing it to tap unused area as a dynamic SLC cache. Micron claims this cache delivers a nice performance boost at lower capacities, and we’ve tested the 256GB version against the 1TB flagship and dozens of other SSDs to see how well the scheme works.

We’ve also taken a mini M.2 variant for a spin to see how it stacks up. There’s much to discuss, so let’s get started.

A familiar cocktail with a twist

The M600’s hardware components are similar to those in the Crucial MX100, which is sold by Micron’s consumer brand. Both drives combine the company’s homegrown 16-nm flash with an off-the-shelf Marvell controller.

Although Crucial used to mirror Micron’s offerings, the MX100 marked the beginning of a divergence. Having identical lineups created too much confusion and pricing overlap, Micron told us, so the two brands will do their own thing moving forward. They’ll continue to pull from the same internal technology portfolio—and Crucial will still target consumers while Micron focuses on businesses, system builders, and PC makers—but their drives will be different.

Forgive the depth of field; I’m running out of ways to photograph SSDs.

Micron’s 16-nm NAND can switch between SLC and MLC modes on the fly. The MX100 doesn’t take advantage of this capability, but in the M600, custom firmware manages the transitions at the block level. The caching scheme, dubbed Dynamic Write Acceleration, houses incoming data in SLC blocks before moving the data to MLC storage during idle periods.

SLC NAND has higher writes speeds because the programming process is much simpler with only one bit per cell. However, SLC also has less storage capacity than two-bit MLC NAND, reducing the M600’s effective cache size to half its MLC capacity. This discrepancy poses a problem for sustained writes that leave no idle time to transfer data from SLC to MLC blocks. The M600’s solution is to switch to MLC writes as the NAND approaches saturation. If writes continue unabated, the M600 eventually starts transferring previously cached data to main storage while continuing to write fresh data from the host.

As one might expect, this juggling act has performance implications. The graph below shows how the M600 256GB’s sequential write speed changes as the drive fills up in an HD Tune test.

The M600 enjoys the higher write speed of SLC NAND up to 46% of its total capacity. Performance drops considerably in the stretch from 46-58%, when writes continue in MLC mode. After that, the drive has to perform SLC-to-MLC transfers while continuing to write incoming data. This multitasking plunges the sequential write speed below 100MB/s and into mechanical storage territory.

Obviously, Dynamic Write Acceleration isn’t ideal for sustained workloads. The scheme is meant to accelerate the short bursts of write activity found in typical client workloads.

There’s another potential fly in the ointment, though. Any cached data that makes its way to MLC storage will be written twice, eating into the NAND’s limited endurance. Fortunately, SLC writes are less damaging than MLC ones. Micron has also adjusted “proprietary NAND trims” to offset the endurance penalty associated with write caching. These top-secret settings are “configured during the NAND fabrication process,” and they can be “tuned for different performance and endurance characteristics.” Despite the name, they’re unrelated to the TRIM command used by the operating system to identify deleted data.

Tweaking NAND trims is apparently quite effective, because the M600 has the most impressive endurance ratings we’ve seen in a client-oriented SSD. Micron’s other drives in this category are specced for 72TB of total writes, and so are their Crucial cousins. But the M600 starts at 100TB for the 128GB version, and the endurance rating scales up with the capacity, taking advantage of the fact that larger drives have more flash to burn. The 1TB monster is good for 400TB of total writes, which works out to a staggering 374GB per day for the length of the three-year warranty.

Capacity Die config Max sequential (MB/s) Max 4KB random (IOps) Endurance

(TBW)

Read Write Read Write
128GB 8 x 16GB 560 400 90,000 88,000 100
256GB 16 x 16GB 560 510 100,000 88,000 200
512GB 32 x 16GB 560 510 100,000 88,000 300
1TB 64 x 16GB 560 510 100,000 88,000 400

Surprisingly, the 1TB and 512GB variants don’t have Dynamic Write Acceleration. Those drives are already fast enough for the controller, according to Micron, and the math works out. The Marvell chip can address up to four chips on each of its eight memory channels, making 32-die configurations ideal for peak performance. At 16GB per die, the cut-off point is 512GB.

Despite falling short of that threshold, the 256GB variant matches the performance specifications of its larger siblings. The 128GB version is a little slower, according to the official figures, but not nearly as much as one would expect given the die count.

The 1TB version (left) has four times the flash of the 256GB (right), whose back is completely bare.

In a moment, we’ll see how the M600 256GB and 1TB compare in a wide variety of tests. There are a few loose ends to tie up before that, including encryption support, which is particularly important to Micron’s corporate clients. The M600 supports all the right standards: TCG Opal 2.0, IEEE 1667, and Microsoft eDrive. Even if the 256-bit AES encryption doesn’t keep the NSA out of your business, it should at least prevent thieves and opportunists from snooping.

To protect against data loss due to flash failures, the M600 employs a RAID-like redundancy scheme called RAIN. It also has a measure of power-loss protection. These aren’t must-have features for a lot of enthusiasts, but they provide some peace of mind, and they should appeal to business customers. Noticing a theme yet?

The M600’s high endurance ratings are likely to attract folks who write a lot of data, making health monitoring especially important. Unfortunately, Micron doesn’t offer an easy-to-use Windows app with health information and other drive-related functionality. At least the SMART attributes are loaded with tickers for lifetime remaining, host writes, reserve blocks, reallocated sectors, and various error types. Anyone who wants to monitor drive health with third-party disk utilities can easily tap into those attributes.

And then there are the miniature M600s…

 

A moment with M.2

The M600 family combines a full line of 2.5″ drives with a collection of tiny mSATA and M.2 units. All the mSATA flavors have the same footprint, but the M.2s are available in single-sided 2280 and double-sided 2260 variants of the form factor. Those numbers refer to the width (22 mm) and the length (either 60 or 80 mm) of the “gumstick” circuit board. Here’s how the M.2 2280 version of the M600 256GB compares to its 2.5″ counterpart:

Awww, isn’t it cute?

The mini versions of the M600 are basically the same as the 2.5″ drives. However, they only scale up to 512GB, and they all have Dynamic Write Acceleration. According to Micron, SLC caching can convey power efficiency benefits by speeding the write process, allowing the drive to return to a low-power state more quickly. Since the mSATA and M.2 models are mostly meant for notebooks, they all have the feature enabled.

DevSleep support is a pre-requisite for SSDs targeting modern notebooks, so it’s no surprise that the M600 supports the ultra-low-power state. The drive also has a built-in throttling mechanism that dials back performance if thermals exceed acceptable limits. Pretty standard stuff.

Despite its fancy form factor, the M600 M.2 uses the same old Serial ATA interface as the rest of the family. It requires an M.2 slot with SATA connectivity, so it won’t work with the PCIe-only M.2 implementations on some motherboards. That SATA requirement also precludes the M.2 drive from working in PCIe adapter cards.

Our storage test systems are too old to have M.2 slots, so we couldn’t throw the mini M600 256GB into the ring with the rest of the SSDs. We did, however, run a few quick tests on the drive and its 2.5″ twin using a newer Gigabyte X99-UD4 motherboard. We also ran the same tests on a handful of comparable Crucial SSDs. All of those drives use slight variations of the same Marvell controller paired with different flash configurations. The M500 and MX100 both employ 16 x 16GB NAND dies, while the M500 has a 32 x 8GB array.

The M.2 and 2.5″ versions of M600 256GB shadow each other in these tests. Although the M.2 is technically faster in a few instances, it’s only ahead by a smidgen.

Impressively, the M600 keeps up with the M550, which has twice as many NAND dies. The M500 and MX100 lag behind, and they’re particularly slow in the sequential write speed test. Dynamic Write Acceleration is evidently quite effective in CrystalDiskMark. Let’s bust out our usual benchmarks to see how it fares in a broader range of tests—and against a larger collection of competitors.

 

CrystalDiskMark — transfer rates

TR regulars will notice that we’ve trimmed a few tests from our usual suite of storage results. The drives were all benchmarked in the same way, but we’ve excluded the results for tests that have grown problematic or less relevant over time. This abbreviated format should be a little easier to digest until our next-gen storage suite is ready.

First, we’ll tackle sequential performance with CrystalDiskMark. This test runs on partitioned drives with the benchmark’s default 1GB transfer size and randomized data.

We’ve color-coded the results to make the M600—and some comparable Crucial SSDs—easier to pick out of the fray.

In the read speed test, most of the SSDs bump into the limits of the 6Gbps SATA interface. Even the M600’s middle-of-the-pack performance puts it relatively close to the leaders.

The field spreads out quite a bit more in the write speed test, but the M600 doesn’t miss a step. The 256GB and 1TB versions are both in the upper tier, again just shy of the leaders.

HD Tune — random access times

Next, we’ll turn our attention to random access times. We used HD Tune to measure access times across multiple transfer sizes. SSDs have near-instantaneous seek times, so it’s hard to graph the results on the same scale as mechanical drives. The WD Black and Seagate SSHD will sit out this round to focus our attention on the SSDs.

The M600’s random access times are competitive for the most part. However, the 256GB variant stumbles in the 1MB random write speed test, causing it to tumble to the back of the field. The Crucial M500 240GB and MX100 256GB both have quicker access times in that test despite their lack of SLC caching.

 

TR FileBench — Real-world copy speeds

FileBench, which was concocted by TR’s resident developer Bruno “morphine” Ferreira, runs through a series of file copy operations using Windows 7’s xcopy command. Using xcopy produces nearly identical copy speeds to dragging and dropping files using the Windows GUI, so our results should be representative of typical real-world performance. We tested using the following five file sets—note the differences in average file sizes and their compressibility. We evaluated the compressibility of each file set by comparing its size before and after being run through 7-Zip’s “ultra” compression scheme.

  Number of files Average file size Total size Compressibility
Movie 6 701MB 4.1GB 0.5%
RAW 101 23.6MB 2.32GB 3.2%
MP3 549 6.48MB 3.47GB 0.5%
TR 26,767 64.6KB 1.7GB 53%
Mozilla 22,696 39.4KB 923MB 91%

The names of most of the file sets are self-explanatory. The Mozilla set is made up of all the files necessary to compile the browser, while the TR set includes years worth of the images, HTML files, and spreadsheets behind my reviews. Those two sets contain much larger numbers of smaller files than the other three. They’re also the most amenable to compression.

To get a sense of how aggressively each SSD reclaims flash pages tagged by the TRIM command, the SSDs are tested in a simulated used state after crunching IOMeter’s workstation access pattern for 30 minutes. The drives are also tested in a factory fresh state, right after a secure erase, to see if there is any discrepancy between the two states. There wasn’t much of one with the M600, so we’re only presenting the used-state scores.

Although the M600 isn’t the fastest SSD in our file copy tests, it fares reasonably well overall. The 256GB and 1TB versions are competitive across all five tests, and they’re very equally matched. The same can’t be said for the old M500, which delivers much faster copy speeds in its 960GB incarnation than in its 240GB one. Dynamic Write Acceleration clearly gives the M600 256GB a boost here.

 

TR DriveBench 2.0 — Disk-intensive multitasking

DriveBench 2.0 is a trace-based test comprised of nearly two weeks of typical desktop activity peppered with intense multitasking loads. More details on are available on this page of our last major SSD round-up.

We measure DriveBench performance by analyzing service times—the amount of time it takes drives to complete I/O requests. Those results are split into reads and writes.

While the M600 1TB has relatively speedy mean service times for both reads and writes, the 256GB version is way behind according to both metrics. The smaller drive has the slowest mean read service time of any of the SSDs, and its mean write service time actually puts it behind the mechanical drives. The Crucial SSDs follow a similar pattern, but their smaller variants don’t suffer as much as the M600 256GB.

All the SSDs execute the vast majority of DriveBench requests in one millisecond or less—too little time for end users to perceive. We can also sort out the number of service times longer than 100 milliseconds, which is far more interesting data. These extremely long service times make up only a fraction of the overall total, but they’re much more likely to be noticeable.

Dynamic Write Acceleration may not produce competitive mean service times, but it definitely cuts down on the number of extremely slow writes. The difference between the M600 256GB and M500 240GB is a full order of magnitude. Even the M550 256GB, which boasts twice the NAND-level parallelism of its M600 sibling, has five times more writes over 100 milliseconds.

With reads, the M600 256GB crosses the 100-ms threshold more often than some of its Crucial peers, including the MX100. Dynamic Write Acceleration definitely doesn’t improve read performance.

Despite its lack of SLC caching, the M600 1TB logs far fewer painfully slow writes than its M550 counterpart. The two are on more equal footing with reads, though the M600 has a definite edge there, as well.

 

IOMeter

Our IOMeter workload features a ramping number of concurrent I/O requests. Most desktop systems will only have a few requests in flight at any given time (87% of DriveBench 2.0 requests have a queue depth of four or less). We’ve extended our scaling up to 32 concurrent requests to reach the depth of the Native Command Queuing pipeline associated with the Serial ATA specification. Ramping up the number of requests also gives us a sense of how the drives might perform in more demanding enterprise environments.

We run our IOMeter test using the fully randomized data pattern, which presents a particular challenge for SandForce’s write compression scheme. We’d rather measure SSD performance in this worst-case scenario than using easily compressible data.

There’s too much data to show clearly on a single graph, so we’ve split the results. You can compare the M600’s performance to that of the competition by clicking the buttons below each graph.

Instead of presenting the results of multiple access patterns, we’re concentrating on IOMeter’s database test. This access pattern has a mix of read and write requests, and it’s similar to the file server and workstation tests. The results for these three access patterns are usually pretty similar. We also run IOMeter’s web server access pattern as part of our standard suite of tests, but it’s made up exclusively of read requests, so the results aren’t as applicable to real-world scenarios. Our own web servers log a fair amount of writes, for example.


The M600 follows the same basic trajectory as its Crucial counterparts in IOMeter. I/O throughput only scales up slightly as the load increases, unlike with some of the other SSDs, whose transaction rates ramp up dramatically with the number of concurrent requests. The first few data points are the most relevant for client systems, and the M600 is at least fairly competitive there.

Note that the M600 256GB is a step behind the terabyte model throughout. Our IOMeter tests hammer the drives pretty much continuously for six hours straight, so they’re probably not ripe for Dynamic Write Acceleration.

Boot duration

Before timing a couple of real-world applications, we first have to load the OS. We can measure how long that takes by checking the Windows 7 boot duration using the operating system’s performance-monitoring tools. This is actually the first test in which we’re booting Windows off each drive; up until this point, our testing has been hosted by an OS housed on a separate system drive.

Level load times

Modern games lack built-in timing tests to measure level loads, so we busted out a stopwatch with a couple of titles.

Write caching won’t make Windows or games load faster, but the M600 doesn’t need any help on those fronts. Both capacities are within a second of the leaders in our load time tests. So is pretty much every other SATA SSD.

We’re working on an updated batch of load-time tests for our next-gen storage suite. Shoot me an email if you have any suggestions. (And thanks to those who have already chimed in.)

Power consumption

We tested power consumption under load with IOMeter’s workstation access pattern chewing through 32 concurrent I/O requests. Idle power consumption was probed one minute after processing Windows 7’s idle tasks on an empty desktop.

The M600’s idle power consumption is reasonably low—and pretty much identical for the 256GB and 1TB variants. However, the larger drive pulls 1W more when servicing a barrage of IOMeter requests. The continuous nature of our IOMeter load doesn’t present an opportunity for Dynamic Write Acceleration to improve power efficiency. There’s simply no downtime during which the drive can assume a lower-power state.

That’s it for performance. If you’re curious about the other SSDs in this review or about how we conduct our testing, hit up the methods section on the next page. Otherwise, feel free to skip ahead to the conclusion.

 

Test notes and methods

Here’s a full rundown of the SSDs we tested, along with their essential characteristics.

  Flash controller NAND
Adata Premier SP610 512GB Silicon Motion SM2246EN 20-nm Micron sync MLC
Adata Premier Pro SP920 512GB Marvell 88SS9189 20-nm Micron sync MLC
Corsair Force Series GT 240GB SandForce SF-2281 25-nm Intel sync MLC
Corsair Neutron 240GB LAMD LM87800 25-nm Micron sync MLC
Corsair Neutron GTX 240GB LAMD LM87800 26-nm Toshiba Toggle MLC
Crucial M500 240GB Marvell 88SS9187 20-nm Micron sync MLC
Crucial M500 480GB Marvell 88SS9187 20-nm Micron sync MLC
Crucial M500 960GB Marvell 88SS9187 20-nm Micron sync MLC
Crucial M550 256GB Marvell 88SS9189 20-nm Micron sync MLC
Crucial M550 512GB Marvell 88SS9189 20-nm Micron sync MLC
Crucial M550 1TB Marvell 88SS9189 20-nm Micron sync MLC
Crucial MX100 256GB Marvell 88SS9189 16-nm Micron sync MLC
Crucial MX100 512GB Marvell 88SS9189 16-nm Micron sync MLC
Intel 335 Series 240GB SandForce SF-2281 20-nm Intel sync MLC
Intel 520 Series 240GB SandForce SF-2281 25-nm Intel sync MLC
Intel 730 Series 480GB Intel PC29AS21CA0 20-nm Intel sync MLC
Intel 730 Series 480GB Intel PC29AS21CA0 20-nm Intel sync MLC
OCZ Vertex 4 256GB Indilinx Everest 2 25-nm Micron sync MLC
OCZ Vertex 450 256GB Indilinx Barefoot 3 M10 20-nm Micron sync MLC
OCZ Vertex 460 240GB Indilinx Barefoot 3 M10 19-nm Toshiba Toggle MLC
OCZ ARC 240GB Indilinx Barefoot 3 M10 A19-nm Toshiba Toggle MLC
Micron M600 256GB Marvell 88SS9189 16-nm Micron sync MLC
Micron M600 1TB Marvell 88SS9189 16-nm Micron sync MLC
SanDisk Extreme II 240GB Marvell 88SS9187 19-nm SanDisk Toggle SLC/MLC
Samsung 840 Series 250GB Samsung MDX 21-nm Samsung Toggle TLC
Samsung 840 EVO 250GB Samsung MEX 19-nm Samsung Toggle TLC
Samsung 840 EVO 500GB Samsung MEX 19-nm Samsung Toggle TLC
Samsung 840 EVO 1TB Samsung MEX 19-nm Samsung Toggle TLC
Samsung 840 Pro 256GB Samsung MDX 21-nm Samsung Toggle MLC
Samsung 850 Pro 512GB Samsung MEX 32-layer Samsung V-NAND
Seagate 600 SSD 240GB LAMD LM87800 19-nm Toshiba Toggle MLC
Seagate Desktop SSHD 2TB NA 24-nm Toshiba Toggle SLC/MLC
WD Caviar Black 1TB NA NA

The solid-state crowd is augmented by a couple of mechanical drives. WD’s Caviar Black 1TB represents the old-school hard drive camp. Seagate’s Desktop SSHD 2TB is along for the ride, as well. The SSHD combines mechanical platters with 8GB of flash cache, but like the Caviar Black, it’s really not a direct competitor to the SSDs. The mechanical and hybrid drives are meant to provide additional context for our SSD results.

If you chose to read this page rather than skipping to the conclusion, you might be interested in a couple more naked circuit board shots. No celebrity iCloud accounts were compromised to obtain these pictures.

We used the following system configuration for testing:

Processor Intel Core i5-2500K 3.3GHz
CPU cooler Thermaltake Frio
Motherboard Asus P8P67 Deluxe
Bios revision 1850
Platform hub Intel P67 Express
Platform drivers INF update 9.2.0.1030

RST 10.6.0.1022

Memory size 8GB (2 DIMMs)
Memory type Corsair Vengeance DDR3 SDRAM at 1333MHz
Memory timings 9-9-9-24-1T
Audio Realtek ALC892 with 2.62 drivers
Graphics Asus EAH6670/DIS/1GD5 1GB with Catalyst 11.7 drivers
Hard drives Seagate Desktop SSHD 2TB with CC43 firmware

WD Caviar Black 1TB with 05.01D05 firmware

Adata Premier SP610 512GB with N0402C firmware

Adata Premier Pro SP920 512GB with MU01 firmware

Corsair Force Series GT 240GB with 1.3.2 firmware

Corsair Neutron 240GB with M206 firmware

Corsair Neutron GTX 240GB with M206 firmware

Crucial MX100 256GB with MU01 firmware

Crucial MX100 512GB with MU01 firmware

Crucial M500 240GB with MU03 firmware

Crucial M500 480GB with MU03 firmware

Crucial M500 960GB with MU03 firmware

Crucial M550 256GB with MU01 firmware

Crucial M550 1TB with MU01 firmware

Intel 335 Series 240GB with 335s firmware

Intel 520 Series 240GB with 400i firmware

Intel 730 Series 480GB with XXX firmware

OCZ Vector 150 256GB with 1.1 firmware

OCZ Vertex 450 256GB with 1.0 firmware

OCZ Vertex 460 240GB with 1.0 firmware

OCZ ARC 100 240GB with 1.0 firmware

Micron M600 256GB with E100 firmware

Micron M600 1TB with E100 firmware

SanDisk Extreme II 240GB with R1131

Samsung 830 Series 256GB with CXM03B1Q firmware

Samsung 840 Series 250GB with DXT07B0Q firmware

Samsung 840 EVO 250GB with EXT0AB0Q firmware

Samsung 840 EVO 500GB with EXT0AB0Q firmware

Samsung 840 EVO 1TB with EXT0AB0Q firmware

Samsung 840 Pro Series 256GB with DXM04B0Q firmware

Samsung 850 Pro 512GB with EXM01B6Q firmware

Seagate 600 SSD 240GB with B660 firmware

Power supply Corsair Professional Series Gold AX650W
OS Windows 7 Ultimate x64

Thanks to Asus for providing the systems’ motherboards and graphics cards, Intel for the CPUs, Corsair for the memory and PSUs, Thermaltake for the CPU coolers, and Western Digital for the Caviar Black 1TB system drives.

We used the following versions of our test applications:

Some further notes on our test methods:

  • To ensure consistent and repeatable results, the SSDs were secure-erased before almost every component of our test suite. Some of our tests then put the SSDs into a used state before the workload begins, which better exposes each drive’s long-term performance characteristics. In other tests, like DriveBench and FileBench, we induce a used state before testing. In all cases, the SSDs were in the same state before each test, ensuring an even playing field. The performance of mechanical hard drives is much more consistent between factory fresh and used states, so we skipped wiping the HDDs before each test—mechanical drives take forever to secure erase.

  • We run all our tests at least three times and report the median of the results. We’ve found IOMeter performance can fall off with SSDs after the first couple of runs, so we use five runs for solid-state drives and throw out the first two.

  • Steps have been taken to ensure that Sandy Bridge’s power-saving features don’t taint any of our results. All of the CPU’s low-power states have been disabled, effectively pegging the 2500K at 3.3GHz. Transitioning in and out of different power states can affect the performance of storage benchmarks, especially when dealing with short burst transfers.

The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at a 75Hz screen refresh rate. Most of the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.

 

Conclusions

We’d normally bust out our famous value scatters to summarize the M600’s performance, pricing, and overall value proposition. However, the drive won’t be sold through normal retail channels, and Micron isn’t publishing an official price list. All the company would tell us is that the 1TB version will cost around $450 in single-unit quantities, which is pretty affordable even in the context of aggressively discounted consumer drives.

Although we can’t factor in pricing, we can calculate the overall performance scores for the M600 256GB and 1TB. These scores are based on a subset of the performance data from our full suite, with CrystalDiskMark’s sequential transfer rates substituted for older HD Tune results. (More details about how we calculate overall performance are available here.)

The M600 1TB is slightly ahead of the equivalent M550—and it’s one of the fastest SSDs overall. The 256GB version doesn’t score as well as its larger sibling, but it does outpoint comparable counterparts in the MX100, M500, and M550 families. The fact that the 16-die M600 scores higher overall than the 32-die M550 is a testament to the benefits of Dynamic Write Acceleration. To be fair, though, there are lots of older 240-256GB SSDs with higher overall performance.

As we saw in our individual tests, this particular breed of SLC caching doesn’t excel in all situations. Dynamic Write Acceleration is much better suited to short, bursty transfers than prolonged workloads. It’s also more effective when there’s lots of free capacity to serve as SLC blocks. Those attributes are probably a decent fit for typical corporate clients, who are unlikely to fill their drives or hammer them with sustained I/O. That said, they’re less ideal for power users with overflowing Steam libraries and demanding usage patterns.

Even if it’s not a perfect solution to limited NAND-level parallelism, Dynamic Write Acceleration is clearly effective, and the underlying technology is very slick. It will be interesting to see how on-the-fly switching manifests in future Micron SSDs, including those sold under the Crucial banner. There’s considerable potential for further performance improvements in future drives that ditch the pokey SATA interface for faster PCI Express links.

In the meantime, SLC caching allows the M600 256GB to be more competitive than any 16-die SSD ought to be. The 1TB version is a great all-rounder even without the aid of write caching, and both have industry-leading endurance ratings. They also have everything else one might want in a business-oriented SSD, including robust data protection and encryption support.

All things considered, the M600 delivers an appealing package for not only Micron’s typical customers, but also PC enthusiasts and everyday consumers. It’s a shame the drive won’t be available through traditional e-tail channels. If you manage to find one at a decent price, the M600 is definitely worth considering—especially at higher capacities. Another way around the limitations of 16-die configs is simply to buy a bigger SSD.

Comments closed
    • moose17145
    • 5 years ago

    I really hope we get some full reviews on M.2 SSDs soon… as I am particularly interested in how these little guys compare to their 2.5″ counterparts

      • Rza79
      • 5 years ago

      Agree, I would like to see the Plextor M6e reviewed.

    • DarkMikaru
    • 5 years ago

    Great review yet again sir. Thank you for all your time and hard work to compile this information for us. After having read a number of comments from fellow readers they do bring up a good point. Though, I am careful to not say that your reviews are a waste of time… they certainly are not. I like seeing where my Samsung 840 stands compared to what the latest and greatest has to offer. If anything though, if you want to shake things up a bit.. why not a Budget SSD review?

    Recently I picked up a Silicon Power S60 120GB for 57.99, a Kingston V300 120GB for 59.99 and broke out an older Samsung 840 Evo 120 GB I picked up for 79 a while back. Then benchmarked them with AS SSD, CrystalDisk & ATTO and the read write speeds swing wildly between drives. IF anyone is interested I’ll post a link to my findings sometime.

    Lets face it.. you just can’t beat $60 bucks for a new SSD that will breathe new life into that old laptop / desktop system. I understand that as a tech forum we all wanna see what the fasted gear is. But it would be nice to review the more affordable offerings sometimes.

    • Anonymus_notthetroll
    • 5 years ago

    Interesting article. Particularly the implementation of this slc/mlc caching technique (dwa). While not earth shattering at face value, still a very clever and, apparently, effective trick. Whether or not it has a “future” is a different topic. Probably not, but you never know. Very cool nonetheless.

    While cost/Gb is often THE mitigating factor in a majority of contemporary ssd buyers, are not similar metrics (cost/x) appllied to quite literally every other aspect of computing. Or am i mistaken. Simply an observation. That part of this review as a matter of fact, coupled with underlying theme that the folks at micron know a thing or two about solid state storage, has led me to keep an eye for the crucial brand in the future. For what its worth.

    Off topic here (still ssd related), but any news on that whole Samsung 840 (non-pro) series issue. Where files that havent been used for X amount of time are afflicted with drastically lowered read speeds?? Just curious (ahhh found it, Sep. 26th aritcle–thx Geoff)

    • f0d
    • 5 years ago

    great review
    but…

    im so bored of looking at ssd reviews now, one is 3 or 5 ms faster than the rest at loading windows, another is 5 or 10 Mb faster transfer speeds than another its all just boring imo

    there really isnt any difference between any of them that you could tell in day to day usage so now i just buy whats cheapest and seems to be more reliable

    heck i cant tell the difference between my old vertex 2 60G and my 250G 840 evo when im just using it and not benchmarking it

    its starting to become like cpu’s where fast enough is fast enough and there isnt any need to upgrade

    • Chrispy_
    • 5 years ago

    So, having benchmarked so many SSDs Geoff, can you really tell any difference whatsoever between a premium “1042% faster than baseline” 850 Pro, and a “914% faster than baseline” MX100 from the bargain bin?

    I think the article is great, but SSD reviews are getting to the point where I skip straight to the conclusion and go “yep, it doesn’t suck – so what’s the cost/GB?” Those are the only two metrics that even matter anymore….

      • Milo Burke
      • 5 years ago

      Sorry Geoff, Chrispy is right. But we still love you. =]

      • dragontamer5788
      • 5 years ago

      I can’t.

      When cheaper PCIe SSDs are released next year, we’ll probably start “feeling” the difference. For now, everything seems to be running up against the 6Gbps limit of SATA3.

      (I know PCIe SSDs exist… but rumor is that Sandforce SSDs will have PCIe signals soon)

      • hansmuff
      • 5 years ago

      It’s too subjective, and tied to whatever use case you have. If you go by cost/GB you don’t need reviews at all. For myself, I do need reviews. I want to see all the hardware features and if they work and how. I want to know if the manufacturer includes software that helps. I want to know warranty. I mean, cost/GB? Go buy a spinning disk, that’s maximum cost/GB.

        • Meadows
        • 5 years ago

        Hear, hear, Geoff.

        Since almost every SSD is virtually like any other (performance-wise), why not include feature comparison tables in future reviews? Should provide visual guidance where benchmarks give none anymore. See how many boxes are ticked with regards to warranty, cache, included accessories, included software, what have you. Could be nice.

        Your “Test notes and methods” page does include a table but you only list 2 parameters almost nobody sincerely cares about, all the while ironically calling them “essential characteristics”.

        Edit: could include your performance baseline numbers and cost/GB numbers in there too, and you might want to let readers sort the list by whichever column they prefer, assuming that doesn’t require too much web wizardry.

      • Dissonance
      • 5 years ago

      Outside of corner cases, it’s hard to tell the difference between the upper-tier drives. I’ve been talking about performance being pretty comparable in that upper tier for a while now, and a lot of our recent recommendations have been heavily based on the cost per gig.

      Heck, even the SSDs with only average overall performance still score well in our load time tests.

      We have a few things planned for our next-gen storage suite that may highlight other differences between drives. SSDs with PCIe interfaces have the potential to deliver big gains, as well, though the jury’s still out on whether those improvements will be noticeable with typical desktop workloads. Hopefully we’ll be able to shed some light on that question in future reviews.

        • Chrispy_
        • 5 years ago

        Looking forward to it!
        The SSD endurance experiment is really interesting to keep tabs on. Perhaps you guys can work miracles with the increasingly boring topic of consumer SSD performance until we get native SATA-Express controllers to speed things along a bit.

        From what I can tell the difference between a cheap SSD and an expensive one is slightly more sequential write (barely of relevance to everyday tasks) and performance consistency under tortuous conditions that don’t exist outside of a rackmount database server. The only consumer who cares about that is the one who spends all day benchmarking his drive instead of actually using it 😉

        • Bauxite
        • 5 years ago

        Its ok though, the busy server world is going to love the new “mainstream” pci ssds. (yeah they’ve been around awhile, but frankly far overpriced and not that great)

        When you can put your entire database, or a terabyte+ of compressed logs on one drive that is approaching some metrics ram had 10 years ago, they really scream.

          • UberGerbil
          • 5 years ago

          One of the companies I use for hosting has been flogging an SSD tier (ie, you pay more and your server runs off SSDs). It’s more than a little dubious for a lot of web sites — unless you’re doing heavy db work, or have an enormous site that can’t get cached, the page load latency of the internet is going to eclipse the lag from secondary storage — but cheap SSDs clearly are already having an impact even on the lowest end of the busy server world.

    • Meadows
    • 5 years ago

    Why is it that out of all these very many SSDs, so few raise the bar and so rarely, too?

    I find myself just skipping most pages whenever a “new” SSD is reviewed. At least the last several times have been like that.

      • Milo Burke
      • 5 years ago

      I like the idea of using MLC space as SLC for speed until it is filled up. But honestly, who runs their SSD’s forever at under 40% capacity? Cool idea, but it won’t raise the bar for most of us.

      And the rest … you’re right, Meadows, few new ideas that really shake things up.

      • Firestarter
      • 5 years ago

      6 Gbit/s SATA

      I’m waiting for the first consumer NMVe drives to hit the market

    • UnfriendlyFire
    • 5 years ago

    “The 1TB monster is good for 400TB of total writes, which works out to a staggering 374GB per day for the length of the three-year warranty.’

    The only true way to find out is to put it into the SSD Endurance test.

      • hansmuff
      • 5 years ago

      .. which will also show another very important metric: does the drive go into read-only at the end of service life, or just die and be inaccessible?

      I have high hopes for the 840 Pro in that part of the test…

    • davidbowser
    • 5 years ago

    Geoff,

    Have you or any of the TR team attempted to get an explanation from Micron/Crucial on their >100ms lag issues? I would love to see an “inside the second” type of analysis on that.

    Crucial has been very good about leading on price but has not shown any tremendous advances since the C300 (I have several) was the first cheap SATA3 SSD. They have been delivering incremental updates to inexpensive, middle-of-the-pack-performance drives for the last few years and I think that the only thing holding them back from being the outright market leaders is the lag.

    Thoughts?

      • hansmuff
      • 5 years ago

      My thought is price. The drives are priced really low and something has got to give. Why would the manufacturer explain to you what their strategy is for minimizing RMA and other support cost while reviews are positive?

      • Dissonance
      • 5 years ago

      I’ve mentioned it to them on several occasions but haven’t been given any specifics as to the cause. The number of extremely long service times seems to be falling with each new generation, though, so progress is being made.

      We may be able to dig into DriveBench results a little deeper to see where those slowdowns occur. That’s on the table as a possibility for our next-gen storage suite, but it’s a little complicated, so I can’t make any promises. The data sets are much larger than a typical ITS benchmark run, which typically has only a few thousand frame times. A three-run DriveBench 2.0 trace totals ~40 million individual I/Os.

    • wmgriffith
    • 5 years ago

    I’m not sure about SLC writes being less damaging than MLC writes. I would guess that a given flash cell would be usable for 2-4 times longer as SLC than as MLC, but this is not saying the same thing.

    I’m curious how this drive would perform in the endurance testing that TR is known (famous?) for. In later life, this drive could re-configure itself for SLC-only operation at half its rated size. As we know from your previous testing, though, end-of-life operation of an SSD doesn’t exactly track with what would make engineering sense.

    It would be great if one could buy an SSD and be able to configure it as S-, M-, or even T-LC on the user side. Is it possible that this drive could more or less behave as an SLC drive if you just forced usage to a single partition at half (or 45%) size?

      • Waco
      • 5 years ago

      The last point is the interesting one here – if I could buy the 1 TB drive and run it in SLC mode at all times it would be totally worthwhile to have that option.

Pin It on Pinterest

Share This