In the world of mechanical hard drives, the new 6Gbps Serial ATA specification means very little. This third-generation standard’s claim to fame is a faster host interface capable of shifting bits at 600MB/stwice the speed of the old 3Gbps spec. However, that 300MB/s “SATA II” interface was hardly a bottleneck for traditional hard drives. Even Western Digital’s latest VelociRaptor, which is the fastest mechanical drive that plugs into a Serial ATA port, can only sustain transfer rates up to 152MB/s. The ‘raptor’s cache didn’t push much more than 236MB/s in our burst speed tests, either, giving the drive little chance of living up to its 6Gbps interface speed.
If you want to wring more than 300MB/s from a mechanical hard drive, you’re going to have to combine several of them in RAID. Solid-state drive makers are actually faced with the same challenge. Individual flash chips don’t necessarily offer superior sequential throughput to traditional hard drives, which means that SSDs seeking to maximize performance must distribute the load across numerous chips tied to multiple memory channels, effectively creating a multi-channel array within the confines of a single drive.
With the average 2.5″ SSD capable of accommodating at least 16 flash memory chips (and even more with double-stacking), there’s plenty of parallelized performance potential just waiting to be tapped. That said, we’ve yet to see an SSD exceed the limits of the 3Gbps SATA spec. Solid-state drive makers have even shown admirable restraint in not endowing their drives with unnecessary 6Gbps SATA connectivity. All but one of them, that is. Crucial’s RealSSD C300 claims not just to support the new Serial ATA spec, but also to exploit it.
A long-time player in the memory industry, Crucial is better known for DIMMs than solid-state drives. The company’s latest SSD is the first on the market with a 6Gbps SATA interface and the only one whose performance ratings exceed second-gen SATA’s 300MB/s ceiling. Naturally, we had to test one for ourselves.
Marvell controller, take two
The RealSSD inherits its 6Gbps Serial ATA support from Marvell’s 88SS9174 flash controller. This new design succeeds the 88SS8014 controller we tested in Plextor’s PX-128M1S back in April. Marvell’s first controller was limited to a 3GBps SATA interface, and more importantly, it didn’t support the TRIM command built into Windows 7. The Plextor drive’s performance suffered mightily as a result, but we’re pleased to report that TRIM has made the cut for the new model. TRIM works in conjunction with Marvell’s garbage collection routine, which runs in the background to reclaim flash pages marked as available by the command. The frequency with which garbage collection is performed depends on how the drive is being used and how much free capacity it has available.
With eight memory channels, the Marvell controller has twice the number available in Indilinx’s popular Barefoot design but is two short of the ten channels Intel squeezed into its X25-series SSDs. Crucial claims the RealSSD can sustain a sequential read rate of 355MB/s when connected to a 6Gbps SATA interface. The drive’s sequential read performance purportedly drops to 265MB/s when using a 3Gbps link.
Connecting the RealSSD to a 3Gbps SATA controller won’t impede the drive’s write performance, though. Crucial lists a sequential write speed of 215MB/s for 256GB flavors of the C300 and 140MB/s for 128GB models. There’s no difference in sequential read speeds between the two capacities, but they do differ when it comes to random-4KB performance ratings. The larger of the two can purportedly push 60k IOps with 4KB random reads and 45k IOps with random writes, while the 128GB drive is limited to 50k IOps with reads and 30k for writes.
The lower-capacity model is the slower of the two because it has less internal parallelism than the 256GB drive. Both feature 16 flash memory chips split evenly between the two sides of the circuit board. However, the 128GB model uses 8GB flash chips that have two dies each, while the 256GB drive is populated with 16GB chips that have four NAND dies apiece. For those who are keeping score at home, that’s 32 dies for the 128GB drive and 64 for the 256GB one.
Even with only eight memory channels in the controller, it’s possible for all of the drive’s NAND dies to be active at the same time. According to Crucial, having access to the additional dies is especially helpful when writing to the drive because it takes “much longer” to program a NAND die than it does to issue the necessary commands and load the NAND cache registers with the data that’s set to be written.
Crucial is the consumer face of memory mogul Micron Technology, so it’s no surprise the RealSSD’s flash memory bears the Micron name. Like those found in most other recent SSDs, the MLC memory chips inside our 256GB RealSSD are fabricated using 34-nm process technology. The 34-nm flash lurking in modern solid-state drives usually conforms to the Open NAND Flash Interface (ONFI) 1.0 standard, but the RealSSD’s flash chips are hip to the much more recent ONFI 2.1 spec.
Among other things, the ONFI standard governs the speed of the NAND interface. Version 1.0 was limited to 50MB/s, while its 2.0 successor more than doubled the interface bandwidth to 133MB/s. ONFI 2.1 adds support for speeds of 166 and 200MB/s. The 2.1 spec also aims to improve read throughput with deeper pipelining and interleaving support.
Unfortunately, beyond listing support for ONFI 2.1, Micron’s website doesn’t provide any information on just how fast the RealSSD’s NAND components, marked MT29F128G08CKCBBH2, might be. Odds are they can do better than 50MB/s, but I’d be surprised if they can push more than 133MB/s given the fact that Micron VP of Memory System Development Dean Klein only mentioned second-gen ONFI support when the C300 was first introduced.
Flipping the C300’s circuit board reveals a DDR3 memory chip that serves as the drive’s cache. The chip weighs in at a hefty 256MB, which is several times the capacity used in most SSDs. Kingston’s Toshiba-based SSDNow V+ was the previous cache king with 128MB. Indilinx-based drives typically have 64MB caches, while Intel’s X25 series makes do with a measly 32MB. Crucial uses the same 256MB cache on 128 and 256GB versions of the RealSSD.
Like most solid-state drives on the market, the C300 is covered by a three-year warranty. Three-year warranties are common for consumer-grade hard drives, too, but premium and enterprise-oriented models usually get five years of coverage. Given their premium pricing, solid-state drives surely qualify as premium products. It would be nice if they got some additional warranty coverage to match.
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 months 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 for comparison. We’ve called up a wide range of competitors, including a selection of desktop hard drives, traditional notebook drives, Seagate’s Momentus XT hybrid, and a cubic assload of pure solid-state goodness. Below is a chart highlighting some of the key attributes of the contenders we’ve lined up.
|Flash controller||Interface speed||Spindle speed||Cache size||Platter capacity||Total capacity|
|Agility 2|| SandForce
|Caviar Black 2TB||NA||3Gbps||7,200 RPM||64MB||500GB||2TB|
|Force F100|| SandForce
|Momentus 7200.4||NA||3Gbps||7,200 RPM||16MB||250GB||500GB|
|Momentus XT||NA||3Gbps||7,200 RPM||32MB||250GB||500GB|
|RealSSD C300||Marvell 88SS9174||6Gbps||NA||256MB||NA||256GB|
|Scorpio Black||NA||3Gbps||7,200 RPM||16MB||160GB||320GB|
|Scorpio Blue||NA||3Gbps||5,400 RPM||8MB||375GB||750GB|
|SiliconEdge Blue||JMicron JMF612||3Gbps||NA||64MB||NA||256GB|
|SSDNow V+||Toshiba T6UG1XBG||3Gbps||NA||128MB||NA||128GB|
|VelociRaptor VR150M||NA||3Gbps||10,000 RPM||16MB||150GB||300GB|
|VelociRaptor VR200M||NA||6Gbps||10,000 RPM||32MB||200GB||600GB|
|Vertex 2|| SandForce
|X25-M G2||Intel PC29AS21BA0||3Gbps||NA||32MB||NA||160GB|
On the SSD front, we’ve collected all the relevant players, including drives based on Indilinx, Intel, JMicron, Marvell, SandForce, and Toshiba controllers. Although it might not seem like a fair fight, we’ve also thrown in results for a striped RAID 0 array built using a pair of Intel’s X25-V SSDs. The X25-V only runs a little more than $100 online, making multi-drive RAID arrays affordable enough to be tempting for desktop users. Our X25-V array was configured using Intel’s P55 storage controller, the default 128KB stripe size, and the company’s latest 126.96.36.1994 Rapid Storage Technology drivers.
Although our test system’s P55 chipset is limited to 3Gbps Serial ATA, the motherboard has an auxiliary Marvell 9128 storage controller with two 6Gbps SATA ports. In addition to testing the RealSSD connected to the P55’s storage controller, we’ve benched the drive with it attached to the Marvell chip running the latest 188.8.131.526 drivers.
I should note that we’ve actually had a C300 in our possession for several months now. The drive’s initial firmware release exhibited serious performance issues, so we elected to wait for an update that Crucial promised would address the problem. That update came in the form of a 0002 firmware release made available on Crucial’s website last month. We’ve used this latest 0002 firmware for all our testing.
The block-rewrite penalty inherent to SSDs 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. Testing drives in a used state may put the TRIM-less Plextor SSD at a disadvantage, but I’m not inclined to indulge the drive just because it’s using a dated controller chip.
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.
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 a drive’s raw capabilities. From there, we can explore whether the drives live up to their potential.
So, yeah, the C300’s pretty fast. Although it doesn’t eclipse 300MB/s even when connected to the 6Gbps Marvell controller, the RealSSD trumps all the other single-drive configurations on the 3Gbps Intel controller. The 6Gbps interface is good for an additional 50MB/s, leaving only the X25-V array with a higher sustained read speed.
The RealSSD doesn’t dominate in writes, but it still turns in an impressive performance, finishing just behind a trio of SandForce-based drives. Interestingly, our motherboard’s Marvell controller proves to be a liability here. The C300’s write speeds are about 30MB/s slower when connected to this 6Gbps controller.
Next up: some burst-rate tests that should test the cache speed of each drive.
Like Intel’s RAID drivers, Marvell’s 6Gbps SATA drivers are clearly using a slice of system memory as a read cache. That doesn’t seem to be the case for writes, though. When running on the 3Gbps Intel controller, the RealSSD sits in the middle of the pack.
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.
The RealSSD’s read access times are very competitive. It’s not the quickest of the bunch at the smallest transfer sizes, but then all the SSDs are amazingly fast up to 64KB. At the 1MB transfer size, the C300 is matched only by the X25-M and X25-V RAID array.
Writes prove to be a little more problematic for the C300, which finds itself lagging behind the faster SSDs by notable margins with the 512-byte and 4KB transfer sizes. The drive fares better with larger transfer sizes and ends up leading the way in the 1MB test.
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.
The RealSSD is wicked-fast in our file creation tests, holding a commanding lead over its closest competition with all three file sets. What’s more, the drive is roughly two times faster than the X25-M across the board.
What’s even more impressive is the fact that the C300 runs the table without the aid of 6Gbps SATA. The Marvell controller’s poor write speeds clearly hamper the RealSSD’s performance here.
The tables turn rather abruptly when we switch to reads, which present no problems for the 6Gbps Marvell controller. In fact, the 6Gbps C300 combo is the fastest single-drive config through all three file sets.
However, the RealSSD had some serious issues with the MP3 and program file sets when connected to the Intel controller. With the programs file set, performance alternated between around 75MB/s and just over 250MB/s. I’ve gone with the lower of the two results in the graph above because it appears you’re stuck with that speed at least half the time. The C300 was much more consistent with the MP3 file setconsistently slow, that is. As you can see, that particular file set has proven to be quite challenging for a number of high-profile SSDs, including the Indilinx-based Nova V128 and Intel’s X25-M.
Whatever afflicts the RealSSD in the read tests doesn’t dampen the drive’s copy performance. The Crucial SSD goes three for three, topping the field with each test pattern. Copy speeds are higher with the Intel controller, but the Marvell’s no slouch here; it’s the second-fastest configuration of the lot.
File copy speed
Although FC-Test does a good job of highlighting how quickly drives read, write, and copy different types of files, the app is antiquated enough to completely ignore the command queuing logic built into modern hard drives and SSDs. FC-Test only uses a queue depth of one, while Native Command Queuing can stack up to 32 I/O requests when asked. To get a better sense of how these drives react when moving files around in Windows 7, we performed a set of hand-timed copy tests with 7GB worth of documents, digital pictures, MP3s, movies, and program files. These files were copied from the drive to itself to eliminate any other bottlenecks.
Before conducting our first wave of tests, I secure-erased each of the SSDs to put them into a pristine, fresh state. Curious to see how the SSDs would handle the same test when in a used state, I ran our IOMeter workstation access pattern with 256 concurrent I/O requests for 30 minutes before launching into a second batch of copy tests.
IOMeter creates a massive test file that spans the entirety of a drive’s capacity, and deleting it to make room for our second salvo of copy tests should gives us a glimpse at each SSD’s TRIM recovery strategy. What we’ve essentially done here is filled all of an SSD’s flash pages, subjected the drive to a punishing workload with a highly-randomized access pattern, and then marked all of the flash pages as available to be reclaimed by garbage-collection or wear-leveling routines.
Mechanical hard drives aren’t subject to the block-rewrite penalty that causes SSD performance degradation as flash pages become occupied, so you won’t see any difference between their fresh- and used-state performances below. We tested the mechanical drives in both states just to be sure, though.
This is a recent addition to our test suite, and since we had to return our PX-128M1S sample to Plextor after reviewing the drive, we were unable to include it here. You’re probably not missing out, though. The Plextor SSD doesn’t support TRIM, so it wouldn’t have fared well.
Most of the SSDs, including the C300, offer near-identical copy performance between their fresh and used states. However, the SandForce-based drives and the SiliconEdge Blue are both slow to reclaim trimmed flash pages, resulting in slower used-state copy speeds. Our torture test also reduces the X25-V array’s copy speed to one fourth its fresh-state level, no doubt because Intel’s RAID drivers are incapable of passing TRIM commands to drives running in RAID.
Oddly, the Marvell controller seems to be hampering the C300’s used-state performance. It’s unclear whether TRIM commands are being passed improperly or whether our torture test somehow confused the controller or its drivers, but the problem doesn’t seem to be related to the Crucial drive itself. The RealSSD is comfortably ahead of the competition when running in 3Gbps mode, regardless of whether it’s in a used or fresh state.
We’ve long used WorldBench to test performance across a broad range of common desktop applications. The problem is that few of those tests are bound by storage subsystem performancea 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 RealSSD has a bit of a split personality in the Photoshop test, preferring the 6Gbps controller to the P55. At least it leads with one of the two configs. Both setups end up at the top of the heap in the Nero test, and by a healthy margin. We don’t see much action in the WinZip test, though.
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. Mozilla recommends that Firefox be compiled with a single thread.
Crucial finds itself within striking distance of first place in our compile test. We’re clearly CPU-bound here, but the C300 does yield quicker compile times when connected to the 6Gbps controller. When hanging off our motherboard’s P55 chipset, the RealSSD takes nearly a minute longer to complete the Firefox build.
We’re currently looking into alternatives for this test, so if you have a suggestions for a multithreaded compiling test that will run in Windows 7, won’t be bound by our CPU, preferably uses open-source code available to the general public, and isn’t OpenOffice (which we’re exploring already), 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 difference between the RealSSD’s boot time with the 3Gbps and 6Gbps controllers is more likely a factor of how long it takes each controller to initialize the drive rather than the speed of the interface. When compared to its peers on the Intel controller, the C300 is only about a second off the pace set by Intel’s X25 series, which puts it right in the mix with Indilinx- and SandForce-based SSDs.
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?
Notch two more wins for the RealSSD. Regardless of which controller we used, the C300 breezed through our level load tests faster than any other drive we’ve tested.
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 ignoredIOs 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 AHCI driver, forcing us to obtain the following performance results with Intel’s 184.108.40.2064 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. The app will only run on unpartitioned drives, so we tested drives after they’d completed the rest of the suite.
You’ve read this sentence before. The RealSSD leads the pack once more. Indeed, this is becoming a bit of a bore.
Sorry, that’s what happens when you write entirely too many storage-related articles in a row. Really, though, I’m not even close to tired of watching the RealSSD put its stamp on the SSD scene. The C300 turns in another impressive performance in DriveBench, not only beating the previous champion, but doing so in convincing fashion. The X25-M already enjoyed a healthy margin over its closest competition, and the RealSSD is even further ahead of rivals from Corsair, Kingston, and OCZ.
Let’s break down the overall average into individual test results to see if anything stands out.
Well, one thing does: the C300’s dominance. When connected to the Intel controller, it crunches more IOps than any other drive with each of our five workloads. The 6Gbps controller proves to be the quicker interface with the transcoding workload, but that config is actually quite a bit slower when a file copy is added to the multitasking mix.
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.
With multitasking out of the equation, the RealSSD still comes out on top.
DriveBench 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.
I wouldn’t make too much of the gap between the RealSSD setups; the startup trace is very short, and there was only a one-second difference in run times between the 3Gbps and 6Gbps controllers. That said, the C300 still ends up at the head of the class. When connected to the Intel controller, it shares the lead with a trio of SandForce-based drives and the X25-M.
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.
Like franchise players that put up big numbers during the regular season but choke under pressure in the playoffs, we’ve seen plenty of SSDs offer excellent performance in our sequential transfer rate tests and even in our multitasking workloads, only to be overwhelmed when presented with a suite of demanding IOMeter test patterns. Admittedly, these workloads batter drives with more I/O requests than they’re likely to see in a single-user desktop environment. They’re still an excellent measure of how drives react when pushed to their limits, though.
Fortunately for Crucial, the RealSSD fares quite well when brutalized by enterprise-class workloads. With the web server access pattern, which is made up exclusively of read requests, the C300 offers slightly better throughput than the X25-M, making it the fastest single-drive config.
The other three workloads include a mix of read and write operations, and the latter seem to be particularly amenable to the DuraWrite compression scheme offered by the SandForce-based Agility, Force, and Vertex SSDs. The RealSSD slots in between those SandForce drives and the rest of the field, which is still a rather nice place to be. There’s no shame in losing to SandForce’s funky write-compression mojo, especially when the RealSSD still manages much higher transaction rates than the other SSDs.
Our IOMeter CPU utilization results highlight some small differences between how the Intel and Marvell controllers consume CPU resources while handling the RealSSD, but we’re still looking at very low utilization overall. Keep in mind that the system’s CPU is a quad-core Core i5-750 with plenty of horsepower.
CPU utilization matters less than overall efficiency in IOMeter, so we’ve graphed the number of transactions per percent CPU utilization.
The RealSSD can’t touch the X25-V array’s production per CPU cycle, but that’s true of all of the single-drive setups. Because of its slightly lower CPU utilization, the 6Gbps C300 pairing is the more efficient of the two, at least under lighter loads. However, as the number of concurrent I/O requests ramps up, the Marvell controller’s CPU utilization increases, dropping that configuration’s efficiency.
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.
Our noise level and power consumption tests were conducted with the drives connected to the motherboard’s P55 storage controller.
I’ve consolidated the solid-state drives here because they’re all completely silent. The SSD noise level depicted below 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.
Solid-state drives have no impact on system noise levels. If you’re starting off with a quiet rig, the C300 isn’t going to make it any louder. A mechanical hard drive will, especially when it’s seeking.
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.
The RealSSD’s power consumption is a little higher than the competition’s, particularly when being hammered by IOMeter. I wouldn’t worry too much, though. The C300 only draws 1.6W under load and less than a watt at idle.
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.
With a $610 street price, the 256GB RealSSD is one of the more expensive SSDs on the market. Yet because it has such a generous capacity, the C300 offers compelling value, at least when judged against other solid-state drives. Mechanical hard drives and Seagate’s hybrid Momentus still give you much more capacity per dollar.
This one’s going to be easy because the results really do speak for themselves. Simply put, the RealSSD C300 turned in one of the finest all-around performances we’ve witnessed from a solid-state drive. We’re used to seeing SSDs post impressive scores in some tests but deliver sub-par or even horrendous results in others. After going over page after page of benchmark data, though, I’m hard pressed to find a definitive weakness in the C300. The drive’s random-write performance with smaller transfer sizes isn’t particularly inspiring, and the RealSSD falls victim to a couple of the same FC-Test file sets that create problems for other SSDs, but that’s really it. Otherwise, the C300 is a monster, often leading its rivals and rarely far behind them when it’s not.
I would, however, avoid using Marvell’s 6Gbps 9128 controller. The C300 is plenty fast on the 3Gbps SATA controller in Intel’s P55 chipset, so you don’t really need the next-gen interface. Plus, the Marvell controller gives up a lot of write performance. I have doubts about whether it’s treating the TRIM command properly. The RealSSD may fare better when connected to the 6Gbps SATA controller built into AMD’s SB850 south bridge.
All the fuss over faster interfaces takes away from what might be the most important ingredient to the RealSSD’s success: support for second-gen ONFI flash. I expect we’ll see similar flash technology in drives based on the next generation of controllers from the likes of Indilinx, Intel, SandForce, and others. For now, Crucial appears to be the only game in town, and the RealSSD is making the most of that advantage.
Now I will concede that the 256GB model’s $610 street price is a rather onerous pill to swallow, even if the associated cost per gigabyte compares quite favorably to other SSDs. Budget-minded folks will probably want to consider the 128GB variant, which sells for a less imposing $374, albeit with a slightly higher cost per gigabyte. That 128GB drive will be slower with writes and random reads, but I’d expect it to remain competitive overall. It should be just as fast as the 256GB model with sequential reads.
I could try to extrapolate the 128GB model’s performance based on Crucial’s lower performance ratings for the drive, but I’d be hesitant to make a recommendation without seeing how it fares in the real world. The 256GB model has certainly done enough to earn itself our Editor’s Choice award, though. The C300 offers the best all-around performance we’ve seen from any consumer-grade SSD, and it does so at a competitive cost per gigabyte. If that’s not Editor’s Choice material, I don’t know what is.