A six pack of solid-state drives compared

For years, solid-state drives have been great in theory, but a little problematic in practice. The rationale for replacing mechanical media with solid-state silicon certainly seems sound. Without moving parts, SSDs can be far more durable and shock-tolerant than drives that have to contend with platters spinning at thousands of revolutions per minute. The lack of mechanical latency allows SSDs to seek several orders of magnitude faster than traditional hard drives, as well.

The first flash-based solid-state drives delivered on those fronts, and they even added an attractive element of low power consumption to the mix. However, these drives were saddled with a few rather serious problems. While even the first flash drives boasted near-instantaneous seeks, their sustained transfer rates were quite slow. Initial capacities were relatively small, as well, and then there was the associated cost, which can only be described as exorbitant.

In practical terms, then, early SSDs were pretty much a failure.

But the nice thing about solid-state drives is that, in theory, Moore’s Law should eventually solve all of the above problems. Indeed, semiconductor progress has effectively dealt with most of them already. Thanks to smaller and denser memory chips, drives with capacities up to 256GB are easy to come by today. Such capacities may not match the multiple terabytes one can get with mechanical desktop drives, but they are more than adequate for most folks. They may even be sufficient to handle even a PC enthusiast’s collection of OS, application, and game files.

Meanwhile, smarter drive controllers have improved SSD throughput by addressing more memory channels in parallel. Intel’s X25-M has already shed the solid-state legacy of poor sustained transfer rates.

That’s two for two so far, which leaves us only with price. And wouldn’t you know, the picture doesn’t look all that dire. Take the X25-M, for example. Intel’s first SSD debuted just eight months ago in an 80GB flavor for just under $600. Thanks to a steady diet of price cuts, you’ll pay closer to $325 for the very same drive today. Plenty of cheaper options are available, too, with some lower-capacity SSDs dropping below $100.

Solid-state drives still have much progress to make on the cost-per-gigabyte front, but depending on your priorities and budget, one could argue that, for SSDs, pratice has finally caught up with theory, resulting in some good and practical storage solutions. Thus, now is a perfect time to check in on the market to see what’s available. We’ve wrangled half a dozen SSDs from Corsair, OCZ, Samsung, Super Talent, and Transcend to see how they stack up against each other—and against Intel’s X25-M standard.

Lining ’em up
We’re focusing our attention today on solid-state drives that use multi-level cell (MLC) flash memory. MLC memory has slower write speeds and a lower write-erase cycle tolerance than single-level cell (SLC) flash, but it’s considerably cheaper, which is why MLC chips have become the de facto standard for consumer-level solid-state drives. SLC flash is predominantly confined to more expensive enterprise-class drives like Intel’s X25-E Extreme.

There’s much more to SSDs than their choice of flash memory types, though. As we saw with the X25-M, a good drive controller design can have a major impact on performance. Intel has so far kept its SSD controller in the family, sharing it only with partners like Kingston, who are selling re-badged X25-series drives. There are other drive controller players in the market, though.

Among the rest, Samsung is easily the most established. Samungs mainstream 256GB SSD, the PB22-J, uses a disk controller of the company’s own design, and you’ll find the same chip in Corsair’s new P256. In both cases, the Samsung controller is paired with 128MB of DRAM cache.

Korean firm Indilinx has only recently entered the SSD market, and its latest controller can be found in OCZ’s Vertex and in Super Talent’s UltraDrive ME. Super Talent silkscreens its own name onto the UltraDrive’s controller chip, but we can confirm that it is indeed Indilinx silicon inside. Both Indilinx implementations employ DRAM cache memory of 64MB. That’s still double the cache available even in most high-end desktop drives.

In addition to its high-end Vertex drive, OCZ also sent over an SSD from its mid-range Apex series. Rather than using a single controller, the Apex combines a trio of JMicron chips in an internal RAID 0 configuration; a pair of JMF602B controllers manage the flash memory, while a JMB309 picks up striping duties. The JMF602B also makes an appearance in Transcend’s awkwardly-named TS32GSSD25S-M, although only in a single-chip configuration without internal RAID. Since the JMicron controller doesn’t support separate cache memory, neither the Apex nor the Transcend drives have DRAM chips onboard.

Controllers and cache configurations largely determine SSD performance, but before we put today’s contenders to the test with an extensive suite of benchmarks, it’s worth perusing each drive’s purported maximum read and write speeds. While not necessarily indicators of a drive’s sustained throughput in the real world, these theoretical peaks do provide fodder for forming initial expectations.

The OCZ drives claim the highest peak transfer rates, with the Vertex apparently capable of reading at 250MB/s and the Apex equaling that speed with writes. Unlike the others, the Apex’s maximum write speed is higher than its peak read rate—likely an artifact of the drive’s funky internal RAID 0 config. There’s only a 20MB/s gap between the Apex’s max read and write rates, too; that difference grows to 70MB/s with the Vertex, which can only write at up to 180MB/s.

With peak read and write rates of 220 and 200MB/s, respectively, the Corsair and Samsung drives appear to be more balanced than the Vertex. Super Talent’s UltraDrive ME, on the other hand, looks to be a little slower overall. The UltraDrive’s read and write speed maximums are much higher than those of the Transcend drive, though. To be fair, the Transcend is by far the cheapest drive of the bunch. The Transcend also has the lowest storage capacity of the lot—and more importantly, half as many flash chips as the other drives in this round-up.

Of course, Transcend does make higher-capacity SSDs. Corsair, OCZ, Samsung, and Super Talent also offer more affordable drives with lower capacities. All seem to agree on covering their drives with two-year warranties. That’s a year’s less coverage than one typically gets with standard desktop or notebook drives, and high-end drives like WD’s Black Series offer up to five years of coverage.

Samsung’s PB22-J is the only exception on the warranty front because it’s not actually sold as a bare drive. The PB22-J is only offered in complete systems from the likes of Dell, HP, and others, and those vendors determine the drive’s warranty coverage and effective price. Worry not, though. The PB22-J is all but identical to Corsair’s P256, which is available for sale as a bare drive.

A note on SSD performance degradation

Not long ago, the folks at PC Perspective uncovered an interesting problem with Intel’s X25-M SSD: its performance dropped over time. AnandTech also dug into the issue, illustrating a similar performance drop with solid-state drives from other manufacturers. But why do SSDs slow down over time? To answer that question, we have to explore the nature of flash memory and how drives interact with modern operating systems.

Flash memory stores data in cells—two bits per cell, for MLC flash. These cells can’t be addressed individually. Instead, they’re organized into pages that are typically 4KB in size. These pages are a part of larger blocks, which have 128 pages each, for a total of 512KB.

While data can be read from individual pages, it can only be written directly to empty ones. If a drive needs to write to a page that already holds data, it has to rewrite the entire block. During a block rewrite, the contents of a block must first be read into a drive’s cache. The pages to be rewritten are then modified, and the entire block is rewritten. Adding these read and modify steps to the write process predictably causes a performance hit.

Eventually, though, an SSD is going to run out of fresh pages. That can actually happen quicker than one might expect, because those old pages getting marked as deleted can chew up space quickly. An SSD can have plenty of “free” storage capacity and yet no empty pages available for writing, bringing the block rewrite penalty into play for every subsequent write operation and slowing performance accordingly.

So how much of a performance penalty does a block rewrite incur? To find out, we used a handy app called HDDerase to wipe the contents of each drive, returning it to a factory-fresh state with empty pages. (The Apex drive’s funky RAID config didn’t get along with HDDerase, so we emptied the drive by flashing its firmware instead). Next, we put each drive through a 4k random writes test with IOMeter and recorded the average response time.

The JMicron-based Apex and Transcend drives have a problem with IOMeter’s default configuration, which uses a starting sector setting of 0 and a maximum disk size value of 0. This config yielded much lower performance than we expected from the JMicron drives, so we retested them using a starting sector of 512 and a 2000000 maximum disk size value, as suggested by OCZ. These settings yielded results more in line with our expectations, so we’ve included them in the graphs below and in the IOMeter section later in the review. In both cases, we’ve marked the custom IOMeter configs with a (2) to separate them from the results obtained with the default settings. Incidentially, we also tried using starting sectors of 1, 20, 50, and 100, but ran into the same abysmal performance as with a zeroed starting sector.

With a clean slate, the Indilinx-based Vertex and UltraDrive have the lowest response times, followed by the X25-M and the Samsung-based PB22-J and P256. Even with their custom IOMeter config, the JMicron-based drives still lag behind the others. They’re significantly slower with IOMeter’s default config, though.

So how do the drives respond after they’ve been used throughly? We ran the same random writes test on each one after it’d been beaten and battered by our storage benchmark suite, and obtained the following results:

The Indilinx-based drives are still in the lead, but their response times have increased by more than an order of magnitude. The X25-M is also notably slower, but only by about 3.5 times. Interestingly, the Samsung-based drives suffer the most here. Their used response times grow to 15 times that of their factory-fresh scores. The random write performance of the Apex and Transcend SSDs changes very little when the drives are used, regardless of the IOMeter config.

Wiping drives is easy enough with HDDerase, but that’s hardly a practical solution to maintaining SSD performance levels. Most folks are going to end up using, er, used drives. Fortunately, our hard drive testing methodology puts drives in the pratical equivalent of a used state right off the bat. HD Tach is the first benchmark we run, and its write speed test writes the full length of the disk, which should eliminating any empty pages. We run HD Tach three times, too, leaving little chance that any pages will remain untouched.

Just to be sure, I checked a fresh X25-M’s random write performance after three runs through HD Tach. The drive’s 2.3-2.5 ms response times were a little off those we observed after the X25-M completed our full test suite, but they’re close enough to give us confidence that HD Tach effectively rids SSDs of empty pages. Keep in mind that the test results on the following pages are not indicative of how the drives perform in a pristine, empty state.

Hopefully, we won’t need to track fresh versus used SSD performance for too long. The storage industry is working on a proposed TRIM command to alleviate SSD performance degradation. Rather than simply marking pages as available when files are deleted, TRIM would require that a page’s contents be emptied. This provision wouldn’t avoid the block rewrite penalty, but it would shift the performance hit to the time of deletion, which makes more sense than hampering writes.

TRIM requires a compatible operating system, and it looks like Windows 7 will support it. OCZ has also produced a TRIM application for its Vertex drive that clears any occupied pages marked as available. However, this app is very much in beta form, and it can only be run manually.

Although not related to the TRIM command, Intel recently updated the firmware of its X25-M series to improve the drive’s long-term performance. We’ve flashed our X25-M with this latest 8820 release and retested the drive to see how it fares against the previous firmware revision.

Our testing methods
Today we’re throwing our six pack of SSDs into the ring against a collection of competitors, including Intel’s X25-M, which we’ve tested with its original firmware and the latest 8820 release. We’ve also included performance results from Intel’s SLC-based X25-E Extreme to illustrate how the fastest single-level-cell drive compares. And lest you think we’ve forgotten about traditional hard drives, you’ll find results for Western Digital’s 7,200-RPM Scorpio Black notebook drive and its 10k-RPM VelociRaptor 300GB. If you’re curious to see how these SSDs stack up against common desktop drives, you can compare the scores on the following pages with those from our recent review of the Caviar Green 2TB. We used the exact same test system and applications in the Caviar Green review, which includes benchmark results for more than a dozen 3.5″ desktop drives.

To make our graphs easier to read, drives are colored by manufacturer. We used two shades of blue for the X25-Ms to differentiate between the original firmware and the latest 8820 revision.

All tests were run three times, and their results were averaged, using the following test system.

Our test system was powered by an OCZ PowerStream power supply unit.

We used the following versions of our test applications:

• WorldBench 5.0
• Intel IOMeter v2004.07.30
• Xbit Labs File Copy Test v1.0 beta 13
• HD Tach v3.01
• Far Cry v1.3
• DOOM 3
• Intel iPEAK Storage Performance Toolkit 3.0

The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at an 85Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests.

All 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.

WorldBench
WorldBench uses scripting to step through a series of tasks in common Windows applications. It then produces an overall score. WorldBench also spits out individual results for its component application tests, allowing us to compare performance in each. We’ll look at the overall score, and then we’ll show individual application results.

A total of 11 points separates the fastest MLC-based SSD from the slowest in WorldBench. The X25-M still reigns supreme here, but the Samsung and Corsair drives aren’t far off the pace. Curiously, OCZ’s mid-range Apex drive proves faster than the high-end Vertex. The Vertex and UltraDrive are only a hair quicker than the Transcend drive.

Multimedia editing and encoding

MusicMatch Jukebox

Windows Media Encoder

Adobe Premiere

VideoWave Movie Creator

Scores are close through most of WorldBench’s multimedia editing and encoding tests. Only Premiere really stresses the storage subsystem, and again, the X25-M leads the MLC field. The Corsair, Samsung, and OCZ Apex are neck-and-neck with the Intel drive, though. They’re certainly faster than the OCZ Vertex, Transcend, and Super Talent.

Image processing

Adobe Photoshop

ACDSee PowerPack

WorldBench’s Photoshop results aren’t particularly interesting, but the field is really spread out in the ACDSee test. Once more, the X25-M is king of the MLC hill with close to a minute’s lead on Corsair and Samsung. The OCZ and Super Talent drives lag even further behind as the Transcend brings up the rear.

Note how SSD performance is already segmented along drive controller lines. The all-but-identical Samsung and Corsair drives clocked similar completion times in the ACDSee test, as did the Indilinx-powered Vertex and UltraDrive. There’s a bigger gap between the Transcend drive and the Apex, but keep in mind that the latter is running its JMicron controllers in a RAID 0.

Multitasking and office applications

Microsoft Office

Mozilla

Mozilla and Windows Media Encoder

Although there are small differences in performance from one drive to the next in WorldBench’s office and multitasking tests, the results are largely a wash.

Other applications

WinZip

Nero

The results of the WinZip and Nero tests are far more interesting. In WinZip, the X25-M is still the fastest MLC-based drive on the block, followed closely by contenders from Corsair and Samsung. The UltraDrive and Vertex look evenly matched in this test, and they’re both faster than the Apex and Transcend drives.

Nero really spreads out the field, but things don’t change much at front of the pack. The X25-M is still the quickest of the lot, with the Corsair and Samsung twins nipping at its heels. OCZ’s Apex is close to two minutes slower than the leaders, and it’s substantially faster than the Transcend, Super Talent, and Vertex drives.

Again, SSD performance is largely defined by the drives’ underlying storage controllers. The completion times for the Samsung-powered drives are very close, as are those for the Indilinx-equipped models.

Boot and load times
To test system boot and game level load times, we busted out our trusty stopwatch.

Only a few seconds separate the quickest-booting SSD from the slowest on our test system. Our mechanical hard drives are actually quicker overall, with nearly a two-second lead on Intel’s quickest SSDs.

Our SSD level load times are all within a few seconds of each other, but the solid-state drives have a clear advantage over their mechanical counterparts. Among the SSDs, the Intel, Corsair, and Samsung drives emerge at the front of the field.

File Copy Test
File Copy Test is a pseudo-real-world benchmark that times how long it takes to create, read, and copy files in various test patterns. File copying is tested twice: once with the source and target on the same partition and once with the target on a separate partition. Scores are presented in MB/s.

To make things easier to read, we’ve separated our FC-Test results into individual graphs for each test pattern. We’ll tackle file creation performance first.

Obviously, the faster write speeds offered by the X25-E’s SLC memory give it a huge edge in file creation tests. Check out the significant improvement in write performance with the X25-M’s latest firmware, though. The firmware update is the only thing keeping the X25-M ahead of its challengers from Corsair and Samsung in four of the five test patterns.

Although not as consistent a performer as the Intel, Corsair, and Samsung drives, OCZ’s Apex does reasonably well with the MP3 and install test patterns. The other drives don’t handle any of the test patterns particularly well.

As we turn our attention to read performance, the X25-M jumps out ahead of the Corsair and Samsung drives. The new Intel firmware doesn’t offer much of a performance advantage here, and in some cases, it’s slightly slower than the original.

The OCZ and Super Talent drives are in the middle of the field. The relative position of these three changes with each test pattern, but they’re pretty close in performance overall. Again, the Transcend drive brings up the rear.

FC-Test – continued

Copy tests combine read and write operations, and the X25-M’s new firmware again provides a healthy performance boost. Although the Apex does well with the Install test pattern, there’s a clear separation between the top three drives—the X25-M, P256, and PB22-J—and their rivals from OCZ, Super Talent, and Transcend.

The results of our partition copy tests play out much like the first round of copy tests. Transcend gets a brief reprieve from the back of the field with the ISO test pattern, but it’s still the slowest contender overall.

The trend of storage controllers defining SSD performance can be seen throughout our FC-Test results. Sure, the P256 and PB22-J are essentially identical drives, but the Vertex and UltraDrive are not, save for their use of the same Indilinx controller.

iPEAK multitasking
We’ve developed a series of disk-intensive multitasking tests to highlight the impact of seek times and command queuing on hard drive performance. You can get the low-down on these iPEAK-based tests here. The mean service time of each drive is reported in milliseconds, with lower values representing better performance.

The X25-M gets knocked off its pedestal a little in iPEAK, where even with the latest firmware, the drive is slightly slower overall than the offerings from Corsair and Samsung. The P256 and PB22-J prove faster than the X25-M with workloads that include file copy operations, while the Intel drive comes out ahead when the secondary task is a Virtualdub import.

The OCZ, Super Talent, and Transcend drives are also more competitive with that Virtualdub import as a secondary task, but they’re much slower overall than the top three. If we average the mean service time of each drive across all nine workloads, that group of four scores in the 1.55-1.95 millisecond range. The top three, by comparison, are between 0.73 and 0.78 milliseconds.

If you’re keeping score, iPEAK is yet another example of how storage controllers define SSD performance. Here, OCZ’s Vertex implementation of the Indilinx controller appears to be slightly faster than Super Talent’s take on the chip.

IOMeter
IOMeter presents a good test case for both seek times and command queuing. As with the random writes test, we used two different IOMeter configs for the JMicron-based Apex and Transcend drives. Results obtained with the custom config are denoted with a (2) in the graphs below.

Yeah, so the write speed advantage inherent to the X25-E’s SLC-based flash memory gives it huge wins with three of four test patterns. Only with the web server test pattern, which is made up exclusively of read operations, is the Extreme not way out ahead of the rest of the pack.

Since we’re focusing on MLC-based drives today, let’s take another look at IOMeter transaction rates with the X25-E omitted.

Although the X25-M is still the class of the field with IOMeter’s read-dominated web server test pattern, it loses the crown to the Indilinx-powered Vertex and UltraDrive in the database and workstation test patterns. The Corsair and Samsung drives offer identical performance here, and their transaction rates are quite low across the file server, database, and workstation test patterns. That’s not entirely unexpected given the drives’ weak random write performance in a used state. We’ve also yet to see Samsung drives—whether they’re SSDs or mechanical offerings—perform well in IOMeter.

Interpreting the performance of the Apex and Transcend drives is tricky because we had to re-test them with different settings than the others. With a custom config that avoids the obvious problems we saw at IOMeter’s defaults—marked as (2) in the graph legends above—the drives look very competitive in the database and workstation test patterns. However, they’re also abysmally slow when running with the default IOMeter settings.

IOMeter’s CPU utilization roughly tracks with transaction rates. The more I/O requests are processed, the more CPU cycles are used.

HD Tach
We tested HD Tach with the benchmark’s full variable zone size setting.

Despite relatively poor showings in our real-world file creation, read, and copy tests, the Vertex and UltraDrive both do quite well in HD Tach’s synthetic read and write speed tests. Both are considerably faster than the Corsair and Samsung drives in the read speed test, and when it comes to writes, they’re way ahead of the X25-M. The Apex and Transcend drives are considerably less competitive here, especially in the write speed test.

I wouldn’t read too much into HD Tach’s sustained read and write speed drag races, if only because the results here clash with what we observed in more real-world file creation, read, and copy tests. For example, according to HD Tach, the Vertex writes 2.5 times faster than the X25-M. However, in our real-world file creation tests, the Intel drive clocked in at 1.8-2.4 times faster than the Vertex, depending on the test pattern.

The burst rates of the Intel, Corsair, and Samsung drives are pretty close. OCZ’s Apex isn’t far off the leaders, but the Vertex, UltraDrive, and Transcend don’t even reach 200MB/s.

HD Tach’s random access time test nicely illustrates just how much faster solid-state drives are than their mechanical counterparts when it comes time to seek. Most of the SSDs turn in a 0.1-millisecond access time, but the JMicron-powered Apex and Transcend drives clock in at 0.2 ms.

Our results from HD Tach’s CPU utilization test fall within the +/- 2% margin of error for this test.

Noise levels
Noise levels were measured with an Extech 407727 Digital Sound Level meter 1″ from the side of the drives at idle and under an HD Tach seek load. Drives were run with the PCB facing up.

Solid-state drives are completely silent. The 42.6 decibels registered by all the SSDs here represents the noise made by the rest of the system. With this setup, moving from mechanical to solid-state storage trims about five decibels.

Power consumption
For our power consumption tests, we measured the voltage drop across a 0.1-ohm resistor placed in line with the 5V and 12V lines connected to each drive. We were able to calculate the power draw from each voltage rail and add them together for the total power draw of the drive.

The Corsair and Samsung drives have the lowest power consumption of the bunch, despite packing 256GB of storage each. At idle, they barely consume a quarter of a watt, which is less than half the power demanded by an X25-M. The UltraDrive’s power draw is also quite low, while the Apex drive’s is surprisingly high. Indeed, the Apex draws more power than the Scorpio Black—a 7,200RPM mechanical hard drive.

Conclusions
The results of our extensive performance testing confirm that drive controllers play the biggest role in determining SSD performance. Most of these drives use the same flash memory chips, so the controller and cache are really the only meaningful differences between them. There’s no difference between the controllers and cache used on the Corsair P256 and the Samsung PB22-J, and those drives were neck-and-neck throughout. The OCZ Vertex and Super Talent UltraDrive ME use the same Indilinx controller and both have 64MB caches, and they offered similar performance across the board. Of course, the performance of the JMicron-powered OCZ Apex and the Transcend TS32GSSD25S-M varied quite a bit due to the former’s RAID config. However, both exhibited the same IOMeter quirks and comparatively slow random write performance after a fresh erase.

Then there’s the Intel controller in the X25-M. Our results confirm the X25-M is still the fastest MLC-based SSD overall. Sure, the Intel drive lost ground in a few tests, falling behind the Samsung-powered SSDs in iPEAK and the Indilinx-powered drives in a couple of IOMeter workloads, not to mention turning in comparably slow sustained write speeds in HD Tach. But across the breadth of our tests, and most importantly, in those that best simulate real-world tasks, the X25-M remains the quickest and most consistent performer.

Among the drives we assembled today, the X25-M’s closest competition comes from the Samsung PB22-J and its Corsair-branded P256 counterpart. These drives offer solid real-world performance, speedy multitasking, and remarkably low power consumption given their 256GB capacities. However, I wouldn’t throw either into an I/O-heavy enterprise environment—relatively low transaction rates suggest that these drives are ill-equipped to handle demanding multi-user loads. I’m also not keen on the Samsung controller’s sluggish random write performance with well-used drives.

The Samsung-powered drives are the most expensive of the lot, with the Corsair model ringing in at $699 online. They’re 256GB drives, so the cost per gigabyte works out to only $2.73—expensive by mechanical standards, but markedly cheaper than the more than $4 per gigabyte you’ll pay for 80 and 160GB versions of Intel’s X25-M.

At a cost per gigabyte around $3.20, the Indilinx-powered OCZ and Super Talent drives squeeze between the Samsung and Intel models. However, the Vertex and UltraDrive are slower overall, and they trail by substantial margins when file copy operations are involved. To Indilinx’s credit, besting the X25-M’s random write performance and IOMeter transaction rates (in a couple of test patterns, anyway) is no easy task. It’s just difficult to recommend either the Vertex or the UltraDrive given how short they pull up elsewhere, particularly since neither is priced low enough to offer compelling value.

This brings us to the JMicron camp, where we have wildly different entries from OCZ and Transcend. The JMicron controller is clearly the worst of the bunch, exhibiting slow random writes even with a factory-fresh drive and uninspired performance overall. The Apex did outrun its high-end Vertex cousin in many tests, but next to the Samsung drives, which do have a higher cost per gigabyte, the Apex is still slower in most tests—while consuming quite a bit more power.

As for the Transcend, well, it’s cheap. There’s something to be said for being able to pick up a silent, power-efficient, and shock-tolerant 32GB SSD for around $100. At that price, though, the cost per gigabyte isn’t that great, and neither is the Transcend’s performance.

In the end, only three of the SSDs we looked at today are strong enough to earn TR Recommended distinction. Intel’s X25-M remains the all-around performance leader, and even with a higher cost per gigabyte than its rivals, the 80GB model’s $325 asking price is relatively affordable. The Samsung PB22-J and its Corsair twin aren’t quite as quick as the X25-M overall, and with these 256GB models running about $700, they’re not as affordable. However, a low cost per gigabyte combined with frugal power consumption and solid performance makes these drives easy to recommend.