Intel’s X25-V solid-state drive on its own and in RAID

The rise of solid-state drives has been one of the most exciting developments in PC hardware over the last few years. With essentially instantaneous seek times, SSDs can access data an order of magnitude quicker than even the most exotic high-RPM hard drives. With no moving parts, SSDs are impervious to mechanical failure, able to withstand shocks that would destroy traditional hard drives. Their silicon roots also make SSDs completely silent and quite power-efficient compared to their mechanized brethren. And the prospects for SSDs look even brighter when one considers the performance potential of what are essentially parallel arrays of flash memory chips.

Really, it’s no wonder many of us are eyeing SSDs for our desktops and notebooks. In a high-performance PC, an SSD is perfect for an OS and applications drive, augmented by a larger pool of mechanical storage. Such a configuration offers excellent performance for the data you want to access quickly and loads of capacity for the rest. Those who opt for a low-power mechanical hard drive to serve as secondary storage will be treated to a virtually silent setup, as well.

Few notebooks can accommodate multiple hard drives, but rugged shock tolerance and the ability to prolong battery life makes SSDs particularly tempting for portables. Solid-state drives have an even greater performance advantage over the 2.5″ mechanical drives typically found in notebooks, which are considerably slower than full-grown 3.5″ desktop models.

Of course, all this awesomeness comes at a cost—quite a high one, in fact. SSD prices have plunged dramatically in recent years, but drives are still far from cheap, especially when one considers the cost per gigabyte.

The relatively high price of solid-state storage is certainly not lost on Intel. At the Consumer Electronics Show back in January, the company said its SSD plans for 2010 would focus more on driving down costs and lowering prices than on increasing performance. Indeed, Intel kicked off the year with its very first value-oriented SSD, the X25-V 40GB. Low-cost, low-capacity SSDs have existed for some time, of course, but the X25-V marks Intel’s first foray into a market that’s sure to receive an increasing amount of attention. The drive launched at $130 and has already dropped to just $115 online.

Compared to a mechanical drive, 40GB for $115 may not seem like much of a value. But this 40GB isn’t saddled with the baggage associated with spinning platters; it’s spread across 34-nm flash memory chips hooked up to the same controller Intel uses in its standard-setting X25-M SSDs. Controller architectures tend to dictate SSD performance, and Intel’s design has already proven itself to be more robust than other solutions on the market.

The X25-V’s relatively low asking price opens the door to folks who might otherwise have been unable to afford a higher-capacity SSD. Given the fact that the X25-V’s cost per gigabyte is competitive with most other SSDs, enterprising users can also combine multiple drives in a RAID 0 array without exceeding the cost of single drives that offer similar capacities. With zero chance of mechanical failure, striping is a lot safer with SSDs than it is with mechanical hard drives.

Naturally, we had to test the X25-V for ourselves. And then Intel sent two, so we’ve run the X25-V through our new gauntlet of storage tests both on its own and in a striped RAID 0 array. Read on to see how Intel’s value SSD fared.

X25-M/2?
As you may have gathered, the X25-V looks to be a stripped down version of the X25-M, with the same storage controller and the same 34-nm flash memory chips. Because the X25-V only has five flash chips onboard, it can only exploit half of the controller’s ten memory channels. That translates to lower performance ratings, as summarized in the chart below.

The X25-V’s maximum sustained read speed is only 32% slower than the X25-M’s, but Intel’s value SSD pulls up quite a bit shorter with writes. The X25-V’s 35MB/s sustained write speed rating is half what Intel quotes for the 80GB X25-M and a little more than a third the speed of a 160GB drive. Plus, 35MB/s just sounds, well, sluggish.

In addition to slower sustained transfer rates, the X25-V carries lower performance ratings for random 4KB reads and writes. Again, the drive is much further behind the X25-M with writes than it is with reads. The X25-V also has a longer write latency than the X25-M, although read latencies are equal between the two.

I’m going to go out on a limb here and bet that the power consumption figures Intel has published for the X25-V aren’t entirely accurate. With fewer flash chips than the X25-M, I’d expect the V to pull less power both at idle and under load. Intel’s datasheets suggest differently, but we’ll test power consumption for ourselves in a moment.

The X25-V may not measure up to the X25-M’s performance levels, but it still inherits quite a few perks from the elder members of the G2 family. Like Intel’s other second-gen drives, the X25-V has built-in garbage-collection and wear-leveling algorithms, native support for the TRIM command built into Windows 7, and an SSD Optimizer application capable of clearing erased flash pages in Windows XP and Vista. Some combination of these measures is necessary to keep an SSD operating in top shape. The Optimizer can be set to run on a schedule, which is much more convenient than having to invoke the process manually. However, Intel notes that system use should be minimized while the Optimizer is running. The application’s support docs suggest an Optimizer pass should only take a few minutes, which should be easy enough for most users to sit through or schedule around.

Intel packs the X25-V into the same 2.5″, 9.5-mm mobile hard drive form factor being used by just about every other SSD maker. The drive’s metal casing actually measures closer to 6.5 mm thick, but a black spacer on the top of the drive brings it up to 9.5 mm. For those looking to install the X25-V in a desktop enclosure, a 3.5″ bay adapter is included in the retail box. 2.5″ drive bays have slowly started creeping into enthusiast-oriented cases, but they’re few and far enough between to make the adapter a nice addition to the overall package.

Three years of warranty coverage is the de facto standard for consumer-grade hard drives, and the X25-V doesn’t budge from that mark. The X25-M is also covered for three years, as are most of the SSDs we’ve reviewed recently.

Our testing methods
If you’re unfamiliar with The Twins, our new duo of storage test platforms, I recommend checking out this page from our recent VelociRaptor VR200M review. These systems pack potent hardware and have been furiously testing hard drives and SSDs for weeks now. Unfortunately, Intel still hasn’t resolved the performance scaling issue we found in its latest storage controller drivers for the P55 chipset. As a result, The Twins are still running the Microsoft AHCI driver built into Windows 7.

Before dipping into pages of benchmark graphs, let’s set the stage with a quick look at the players we’ve assembled to take on the X25-V. Below is a chart highlighting some of the key attributes that can affect drive performance.

The X25-V will do battle with a slew of SSDs based on a range of different controllers from Indilinx, Intel, JMicron, Marvell, and Toshiba. All of those drives cost about three times what you’ll pay for the X25-V—an important consideration to keep in mind when looking at the test results on the following pages.

What about newer SSDs like Crucial’s RealSSD C300 and all that SandForce-based hotness that was on display at CES? Well, I’ve actually been busy testing a trio of SandForce-based drives. We should have a review of those published very soon. As for the C300, it’s been sitting on the shelf awaiting a firmware update to address performance issues. That firmware update has finally been released, and by the time you read this, the RealSSD should be mid-way through our test suite.

For some additional perspective, we’ve included performance data from a trio of mechanical hard drives. Western Digital’s 10k-RPM VelociRaptor VR200M is the fastest mechanical hard drive that plugs into a Serial ATA interface, so it has perhaps the best chance of challenging SSDs on the performance front. We’ve also added the original VelociRaptor VR150M to the mix, since it still has quicker access times than most desktop drives. Finally, we’ve thrown in a two-terabyte Caviar Black to represent the best performance 7,200-RPM mechanical drives have to offer.

Since the X25-V is relatively inexpensive, we couldn’t resist combining two of them in a RAID 0 array for some additional testing. Besides, it would have been cruel to pit the budget X25-V against a stack of more expensive competitors without throwing the value drive a bone. We crafted our array using the Serial ATA RAID controller built into Intel’s P55 chipset. The array was configured with a 128KB stripe size—the default in Intel’s RAID BIOS—and we used the company’s 9.6.0.1014 Rapid Storage Technology drivers. Unfortunately, although the Intel drivers support TRIM on SSDs acting as single drives, they can’t pass the command to solid-state drives that are members of a RAID array. We have yet to encounter a RAID controller that claims to support TRIM for SSDs running in RAID.

The lack of TRIM support could prove problematic for our RAID config given how much the block-rewrite penalty inherent to flash memory can curtail long-term SSD performance. This penalty and the TRIM command designed to offset it both complicate our testing somewhat, so I should explain our SSD testing methods in greater detail. Before testing the drives, each was returned to a factory-fresh state with a secure erase, which empties all the flash pages on a drive. Next, we fired up HD Tune and ran full-disk read and write speed tests. The TRIM command requires that drives have a file system in place, but since HD Tune requires an unpartitioned drive, TRIM won’t be a factor in those tests.

After HD Tune, we partitioned the drives and kicked off our usual IOMeter scripts, which are now aligned to 4KB sectors. When running on a partitioned drive, IOMeter first fills it with a single file, firmly putting SSDs into a used state in which all of their flash pages have been occupied. We deleted that file before moving onto our file copy tests, after which we restored an image to each drive for some application testing. Incidentally, creating and deleting IOMeter’s full-disk file and the associated partition didn’t affect HD Tune transfer rates or access times.

Our methods should ensure that each SSD is tested on an even, used-state playing field. However, differences in how eagerly an SSD elects to erase trimmed flash pages could affect performance in our tests and in the real world.

With few exceptions, all tests were run at least three times, and we reported the median of the scores produced. We used the following system configuration for testing:

We used the following versions of our test applications:

• WorldBench 6
• Intel IOMeter 2006.07.27
• Xbit Labs File Copy Test 0.3
• HD Tune 4.01
• Visual Studio 2008 with 03-23-2010 Firefox source
• Call of Duty: Modern Warfare 2
• Crysis Warhead

The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at a 75Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests.

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.

HD Tune
We’ll kick things off with HD Tune, which replaces HD Tach as our synthetic benchmark of choice. Although not necessarily representative of real-world workloads, HD Tune’s targeted tests give us a glimpse of each drive’s raw capabilities. From there, we can explore which drives live up to their potential.

Intel may say that the X25-V can only read at 170MB/s, but the drive does a little better than that here. The X25-V averages out to an impressive 185MB/s, which puts it right behind the SiliconEdge Blue and the SSDNow V+.

Of course, that’s just running as a single drive. Combine two X25-Vs in RAID, and you’re looking at sustained read speeds of well over 300MB/s. The RAID config’s read rate does oscillate quite a bit, between 325 and 375MB/s, but that’s a very good place to be considering the X25-M only manages 216MB/s in this test.

The X25-V exceeds Intel’s conservative performance ratings in HD Tune’s write speed test, as well. Still, with an average write speed of only 39MB/s, Intel’s value SSD is easily the slowest of the lot.

Adding a second drive in RAID does improve write speeds by nearly a factor of two, but that’s still not enough to catch the X25-M, let alone the rest of the field. As you can see by the X25-M’s place in the standings, Intel SSDs have comparatively weak write performance, at least in synthetic tests such as this one.

Next up: some burst-rate tests that should test the cache speed of each drive.

This test should measure cache speed, at least—except for when Intel’s RAID driver secures a slice of system memory for use as cache. Obviously, the RAID results aren’t directly comparable to the others here. On its own, the X25-V’s burst speed isn’t terribly impressive. The drive only manages to hit 115MB/s, which is 11MB/s shy of the X25-M, but still slower than everything else.

Our HD Tune tests conclude with a look at random access times, which the app separates into 512-byte, 4KB, 64KB, and 1MB transfer sizes.

All of the SSDs have freakishly low read access times with smaller transfer sizes, and the X25-V is among the best of the bunch. However, Intel’s value SSD starts to falter as we hit the 64KB and 1MB transfer sizes. RAID helps much more with the latter than it does with the former.

The X25-V finds itself at the front of the pack through our first two random-write transfer sizes. Again, though, the X25-V starts to have problems when we step up to larger transfer sizes. The drive stumbles first with the 64KB transfer size before being completely tripped up by the 1MB test, in which its writes times are even slower than the 7,200-RPM mechanical drive.

Our RAID config cuts write access times by roughly half, at least for transfer sizes that take more than a tenth of a millisecond. Striping the X25-V also leads to quicker access times with smaller transfer sizes, but the differences are much smaller.

File Copy Test
Since we’ve tested theoretical transfer rates, it’s only fitting that we follow up with a look at how each drive handles a more realistic set of sequential transfers. 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. We’ve converted those completion times to MB/s to make the results easier to interpret.

Windows 7’s intelligent caching schemes make obtaining consistent and repeatable performance results rather difficult with FC-Test. To get reliable results, we had to drop back to an older 0.3 revision of the application and create our own custom test patterns. During our initial testing, we noticed that larger test patterns tended to generate more consistent file creation, read, and copy times. That makes sense, because with 4GB of system memory, our test rig has plenty of free RAM available to be filled by Windows 7’s caching and pre-fetching mojo.

For our tests, we created custom MP3, video, and program files test patterns weighing in at roughly 10GB each. The MP3 test pattern was created from a chunk of my own archive of ultra-high-quality MP3s, while the video test pattern was built from a mix of video files ranging from 360MB to 1.4GB in size. The program files test pattern was derived from, you guessed it, the contents of our test system’s Program Files directory.

Even with these changes, we noticed obviously erroneous results pop up every so often. Additional test runs were performed to replace those scores.

We’ve already seen the X25-V struggle in HD Tune’s write-speed drag race, so it’s no surprise to see the drive at the back of the field in our file creation tests. The X25-V still manages to eclipse Intel’s 35MB/s write-speed rating, but that’s only good enough to come close to the Plextor PX-128M1S SSD, which lacks TRIM support.

Even without TRIM working its magic for our RAID config, the striped X25-Vs manage to push file creation speeds into the 70MB/s range. Performance doesn’t scale up as nicely from one drive to two as it did in HD Tune, suggesting that the lack of TRIM support is costing the X25-V array here. It would take a lot more than a few MB/s to bring the X25-V’s file creation speed into contention with even a 7,200-RPM mechanical hard drive, though.

Like the X25-M, the V stumbles a bit when asked to read our MP3 file set. However, Intel’s value SSD is much more competitive with our collections of program and video files.

The big winner in this test is the X25-V RAID config, which proves to be much faster than everything else we’ve tested. Not even the pesky MP3 file set can hold our pair of X25-Vs out of first place.

The X25-V’s relatively weak write performance hurts the drive in our copy tests, which combine read and write operations. Even the TRIM-less Plextor SSD manages higher copy speeds with all three file sets, although the X25-Vs at least get some revenge in RAID.

Application performance
We’ve long used WorldBench to test performance across a range of common desktop applications. The problem is that few of those tests are bound by storage subsystem performance—a faster hard drive isn’t going to improve your web browsing or 3ds max rendering speeds. A few of WorldBench’s component tests have shown favor to faster hard drives in the past, though, so we’ve included them here.

The good news is that whatever seems to hold back the X25-M’s performance in WorldBench’s Photoshop test isn’t a problem for the X25-V. The bad news is that the drive is still a little slow in both Photoshop and Nero. Teaming a couple of X25-Vs in RAID does improve performance some, but not by the same sorts of margins we saw with sequential transfers.

Although source-code compiling isn’t a part of the WorldBench suite, we’ve often been asked to add a compile test to our storage reviews. And so we have. For this test, we built a dump of the Firefox source code from March 23, 2010 using Visual Studio 2008. This process writes over 22,000 files totaling about 840MB, so there’s plenty of disk activity. However, we had to restrict compiling to a single thread because using multiple threads in Windows 7 proved to be unstable.

Our compiling test doesn’t tease out any differences between the various storage configs, but a compiling job that makes effective use of multiple processor cores might. If you have any suggestions for a Windows 7 compiling test that won’t be bound by our CPU and preferably uses open-source code available to the general public, please shoot me an email.

Boot and load times
Our trusty stopwatch makes a return for some hand-timed boot and load tests. When looking at the boot time results, keep in mind that our system must initialize multiple storage controllers, each of which looks for connected devices, before Windows starts to load. You’ll want to focus on the differences between boot times rather than the absolute values.

This boot test starts the moment the power button is hit and stops when the mouse cursor turns into a pointer on the Windows 7 desktop. For what it’s worth, I experimented with some boot tests that included launching multiple applications from the startup folder, but those apps wouldn’t load reliably in the same order, making precise timing difficult. We’ll take a look at this scenario from a slightly different angle in a moment.

The X25-V proves to be among the fastest boot drives we’ve ever tested. However, the additional time needed to initialize our RAID array slows the boot process by four seconds.

A faster hard drive is not going to improve frame rates in your favorite game (not if you’re running a reasonable amount of memory, anyway), but can it get you into the game quicker?

The X25-V’s Modern Warfare 2 level load times are within a second of the leaders. Interestingly, the X25-V is much more consistent in this test than the X25-M, whose load times were all over the map between 14 and 40 seconds.

RAID doesn’t help much in Modern Warfare 2, but in Crysis, striping shaves more than seven seconds off the X25-V’s load time. Even more interesting is the fact that the X25-V loads our Crysis level more than a second faster than the X25-M. However, a single X25-V only saves a couple of seconds over our high-performance 7,200-RPM drive.

Disk-intensive multitasking
TR DriveBench is a new addition to our test suite that allows us to record the individual IO requests associated with a Windows session and then play those results back on different drives. We’ve used this app to create a new set of multitasking workloads that should be representative of the sort of disk-intensive scenarios folks face on a regular basis.

Each workload is made up of two components: a disk-intensive background task and a series of foreground tasks. The background task is different for each workload, but we performed the same foreground tasks each time.

In the foreground, we started by loading up multiple pages in Firefox. Next, we opened, saved, and closed small and large documents in Word, spreadsheets in Excel, PDFs in Acrobat, and images in Photoshop. We then fired up Modern Warfare 2 and loaded two special-ops missions, playing each one for three minutes. TweetDeck, the Pidgin instant-messaging app, and AVG Anti-Virus were running throughout.

For background tasks, we used our Firefox compiling test; a file copy made up of a mix of movies, MP3s, and program files; a BitTorrent download pulling seven Linux ISOs from 800 connections at a combined 1.2MB/s; a video transcode converting a high-def 720p over-the-air recording from my home-theater PC to WMV format; and a full-disk AVG virus scan.

DriveBench produces a trace file for each workload that includes all IOs that made up the session. We can then measure performance by using DriveBench to play back each trace file. During playback, any idle time recorded in the original session is ignored—IOs are fed to the disk as fast as it can process them. This approach doesn’t give us a perfect indicator of real-world behavior, but it does illustrate how each drive might perform if it were attached to an infinitely fast system. We know the number of IOs in each workload, and armed with a completion time for each trace playback, we can score drives in IOs per second.

Below, you’ll find an overall average followed by scores for each of our individual workloads. The overall score is an average of the mean performance score in each multitasking workload.

DriveBench doesn’t produce reliable results with Microsoft’s ACHI driver, forcing us to obtain the following performance results with Intel’s 9.6.0.1014 RST drivers. We couldn’t get DriveBench to play nicely with our the X25-V RAID config, either, which is why it’s not listed in the graphs below. DriveBench will only run on unpartitioned drives, so we tested drives after they’d completed the rest of the suite.

The X25-V sits almost exactly between the SiliconEdge Blue and the Plextor PX-128M1S overall. That’s not a bad place to be; the mechanical drives have much lower throughput here. Let’s break down DriveBench’s individual test to see how the X25-V fared with each workload.

No surprise here: the file copy workload proves to be the most difficult for the X25-V. However, Intel’s value SSD turns in stronger performances with the other workloads, which write much less data.

Curious to see whether removing the multitasking element of these tests would have any bearing on the standings, I recorded a control trace without a background task.

Although the pack shuffles a bit in this control test, the X25-V’s relative position is unchanged.

DriveBench also lets us start recording Windows sessions from the moment the storage driver loads during the boot process. We can use this capability to take another look at boot times, again assuming our infinitely fast system. For this boot test, I configured Windows to launch TweetDeck, Pidgin, AVG, Word, Excel, Acrobat, and Photoshop on startup.

Again, the X25-V finds itself in fifth place between the Plextor SSD and the rest of our silicon-based drives. The value SSD still churns out about four times the IOps of a the fastest mechanical drive.

IOMeter
Our IOMeter workloads are made up of randomized access patterns, presenting a good test case for both seek times and command queuing. The app’s ability to bombard drives with an escalating number of concurrent IO requests also does a nice job of simulating the sort of demanding multi-user environments that are common in enterprise applications.

When you consider how much more the other drives cost, the lone X25-V’s IOMeter performance really is quite impressive. The X25-V’s transaction rates are substantially higher than those of the Kingston, Plextor, and Western Digital SSDs, not to mention all the mechanical hard drives.

IOMeter transaction rates scale even higher with our RAID config, which makes quick work of the X25-M on its way to a dominating overall performance. If you judge SSDs in terms of IOps per dollar, the X25-V definitely lives up to its value branding.

Those worried that our software RAID config might chew through an inordinate number of CPU cycles need not fret. The X25-V RAID array’s IOMeter CPU utilization is quite reasonable, giving us a perfect opportunity to bust out some efficiency graphs that depict IOps per percent CPU utilization.

A single X25-V looks to be on par with the X25-M, but our RAID array clearly crunches more transactions for every CPU cycle it uses. Impressive.

Noise levels
Our acoustic and power consumption tests were conducted with the drives connected to the motherboard’s P55 storage controller. Noise levels were measured with a TES-52 Digital Sound Level meter 1″ from the side of the drives at idle and under an HD Tune seek load. Drives were run with the PCB facing up.

I’ve consolidated the solid-state drives here because they’re all completely silent. The SSD noise level depicted above is a reflection of the noise generated by the rest of the test system, which has a passively-cooled graphics card, a very quiet PSU, and a nearly silent CPU cooler.

As you can see, SSDs can lower noise levels when quite a bit when compared to performance-oriented mechanical hard drives. This is true particularly under seek loads, which is where mechanical drives are at their loudest.

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. Drives were tested while idling and under an IOMeter load consisting of 256 outstanding I/O requests using the workstation access pattern.

Remember that limb I went out on? Yeah, it just snapped. Despite having half the memory chips of the X25-M, the V pulls about the same amount of power at idle and only a fraction of a watt less under our IOMeter load. The X25-V’s power draw is still impressively low, but it’s not as power-efficient as the SSDNow or Nova drives, both of which pack considerably more capacity.

Capacity per dollar
After spending pages rifling through a stack of performance graphs, it might seem odd to have just a single one set aside for capacity. After all, the amount of data that can be stored on a hard drive is no less important than how fast that data can be accessed. Yet one graph is really all we need to express how these drives stack up in terms of their capacity, and more specifically, how many bytes each of your hard-earned dollars might actually buy.

We took drive prices from Newegg to establish an even playing field for all the contenders. Mail-in rebates weren’t included in our calculations. Rather than relying on manufacturer-claimed capacities, we gauged each drive’s capacity by creating an actual Windows 7 partition and recording the total number of bytes reported by the OS. Having little interest in the GB/GiB debate, I simply took that byte total, divided by a Giga (10⁹), and then by the price. The result is capacity per dollar that, at least literally, is reflected in gigabytes.

On a pure cost-per-gigabyte basis, SSDs remain almost offensively expensive when compared with their mechanical counterparts. The SSDs cluster between around 320 and 400 megabytes per dollar, with the X25-Vs sitting close to the middle of that range.

Since we’ve limited our RAID testing to a chipset implementation that comes free with every P55-based motherboard, the only additional cost is a second SSD. The RAID config offers almost exactly double the capacity of a single drive.