This past summer, I put a set of fancy racing wheels on my road bike. They’re a study in aerodynamic lightness: deep rims and only as many spokes as is absolutely necessary, all spun from deliciously sexy carbon fiber. I race occasionally and ride quite a lot, so it wasn’t too difficult to justify the indulgence. Over the years, I’ve learned that I’m in tune enough with my bike to detect the sometimes subtle impact that high-end gear can have on my riding experience.
The wheels made a difference immediately. Within minutes, I noticed that the bike was accelerating quicker under power, especially when going uphill. On downhills, I found myself holding a little more speed. Rough pavement felt smoother thanks to the compliance built into the carbon spokes, too.
Put those wheels under a casual rider, and he might not notice. But I am not a casual rider… much like PC enthusiasts are not casual computer users. Enthusiasts spend multiple hours a day sitting in front of systems they’ve optimized for optimal performance. We’re a discerning bunch, and solid-state drives are our fancy racing wheels.
That’s right, I just defied this industry’s longstanding tradition of car analogies to drag out my 10-speed. Deal with it, because the shoe fits. SSDs speed system performance in ways that are often subtler than the impact of a faster CPU or graphics card. Thanks in part to near-instantaneous access times, solid-state offerings are much more responsive than mechanical hard drives. They seem to never get hung up, an attribute that effectively smoothes out any bumps in the road. At least as far as PC hardware goes, SSDs are also pretty sexy.
Among the solid-state drives available for purchase today, Intel’s 510 Series is the fresh new face. The drive comes more than a year and a half after Intel released its second-generation X25-M, and it’s just the beginning. Intel says the 510 Series is the first of a number of new SSD products due out this year. One of those drives will feature the long-awaited successor to the flash controller behind the X25-M line. The updated controller is being designed in-house by Intel, but it’s not finished yet.
Rather than using one of its own controllers, Intel has equipped the 510 Series with a Marvell 88SS9174 controller familiar from Crucial’s RealSSD C300. This is a major shift for a company that has made a habit of tying together its semiconductor products. Seeing a Marvell controller on the 510 Series’ circuit board is kind of like popping the hood of the latest Ferrari and finding an engine from Porsche. You didn’t expect me to avoid car analogies completely, did you?
The decision to use third-party controller silicon makes sense if you step back and realize that Intel is in the business of selling solid-state drives rather than flash controllers. If Intel has multiple new SSD products due out this year, it’s not unreasonable to expect some variety on the controller front. Besides, the chip giant says that modern controllers are much improved over the often-flaky initial offerings that the X25-M was designed to combat. Intel is confident enough in the Marvell chip to use it in a high-end SSD.
While Crucial’s RealSSD C300 uses a Marvell 88SS9174-BJ2P controller, the 510 Series is equipped with a slightly newer BKK2 revision of the same silicon. When asked how the BKK2 differs from the BJ2P, Marvell told us the newer revision is a “comparable device.” According to Intel, the hardware behind the 510 Series’ controller chip is identical to what’s available in competing products.
What makes the 510 Series’ implementation of this particular chip unique is the firmware that comes loaded onto the drive. Intel has optimized the firmware to improve sequential throughput and to make the most of the controller’s 6Gbps Serial ATA interface—you know, the one that nicely matches the high-speed SATA ports in new Sandy Bridge motherboards. We couldn’t get Intel to reveal any specifics on exactly how the firmware has been tweaked, probably because it doesn’t want competitors catching wind of its approach. The firmware tuning present in the 510 Series won’t be shared with other SSD makers using the Marvell chip.
|Flash controller||Marvell 88SS9174|
|Flash type||Intel 34-nm MLC NAND|
|Available capacities||120, 250GB|
|Sequential reads||450MB/s (120GB)
|Sequential writes||210MB/s (120GB)
|Random 4KB reads||20,000 IOps|
|Random 4KB writes||8,000 IOps|
|Read latency||65 µs|
|Write latency||80 µs|
|Idle power||100 mW|
|Active power||380 mW|
|Warranty length||Three years|
Based on the 510 Series’ performance specifications, it looks like Intel’s firmware fiddling paid huge dividends. The drive’s 500MB/s sequential read rate is quite impressive, as is the 315MB/s write speed. Those figures apply to the 250GB drive on a 6Gbps interface. 120GB versions of the 510 Series are a little slower, especially when it comes to sequential writes.
Regardless of drive capacity, the 510 Series promises to churn through 20,000 random 4KB reads and 8,000 random 4KB writes. Surprisingly, those specs are less impressive than the 35,000 random reads and 8,600 random writes offered by the latest X25-M. Crucial’s RealSSD C300 claims random 4KB read and write performance of 60,000 and 45,000 IOps, respectively, so it’s clear that Intel’s optimizations for sequential throughput have cost the 510 Series in other areas.
Intel wouldn’t confirm whether wear-leveling or garbage-collection components of the 510 Series’ firmware have been tweaked. However, it did concede that, with typical client workloads, the drive’s write amplification factor is “about the same” as that of the X25-M. The write amplification factor determines how much flash capacity is consumed by every host write, so it has a hand in dictating a drive’s useful life.
The other part of the lifespan equation depends on the flash chips used. Rather than jumping to 2x-nm NAND, the 510 Series uses 34-nano MLC flash chips built by Intel’s joint venture with Micron. 16 of these chips dot the 250GB drive we’ll be testing today—eight on the top of the circuit board and another eight on the back. The chips conform to the ONFI 2.0 spec, which allows for per-chip transfer rates up to 133MB/s.
Although the Marvell controller’s eight memory channels fall two short of the 10 channels provided by the X25-M, the old Intel controller is an ONFI 1.0 affair that caps per-chip transfers at 50MB/s. The 510 Series may have fewer channels, but each one should be quite a bit faster. With 128MB of DRAM memory onboard, the 510 Series also has four times the cache of its predecessor.
As you may have gleaned from the 510 Series’ available capacity points, the drives use a typical overprovisioning percentage of about 7%. Users can increase the amount of flash capacity allocated to “free area” (essentially a working space used by the drive’s controller) by leaving a chunk of the drive unpartitioned.
Like most new SSDs, the 510 Series’ retail box contains a handy mounting bracket that allows the drive to slide into the 3.5″ bays common in desktop enclosures. You also get a Serial ATA data cable and power adapter. Free drive-cloning software is offered on Intel’s website for folks who want to transition from an existing drive.
Our testing methods
Before dipping into our benchmark results, let’s take a quick look at the mix of rivals we’ve put together to face the 510 Series, and the methods we use to test storage devices here at TR. We include these details to help you better understand and replicate our results, but if you’re already familiar with our approach to storage testing, feel free to skip ahead to the benchmarks. We won’t be offended.
Today, the 510 Series will face off against a collection of SSDs based on a handful of different controllers. Note that it and the RealSSD C300 are the only ones with 6Gbps SATA interfaces. We’re using a Sandy Bridge motherboard with 6Gbps SATA connectivity, so those drives have a distinct advantage over the others. The Agility 2 also has somewhat of an edge thanks to a 28% overprovisioning percentage that’s four times what’s typical for consumer-grade SSDs. We’ve found that SandForce-based SSDs tend to run slower when they set aside a more traditional 7-8% of their flash capacity as spare area.
|Flash controller||Interface speed||Cache size||Total capacity|
|Corsair Nova V128||Indilinx Barefoot ECO||3Gbps||64MB||128GB|
|Crucial RealSSD C300||Marvell 88SS9174-BJP2||6Gbps||256MB||256GB|
|Intel X25-M G2||Intel PC29AS21BA0||3Gbps||32MB||160GB|
|Intel 510 Series||Marvell 88SS9174-BKK2||6Gbps||128MB||250GB|
|OCZ Agility 2||SandForce SF-1200||3Gbps||NA||100GB|
Yes, we know there’s no Vertex 3 in that list. OCZ’s first crack at SandForce’s next-gen controller is still waiting on final firmware. A beta drive with pre-production firmware is benching beside me as I type this, though. Look for more coverage of the drive soon.
Speaking of firmware, we’ve updated all the drives to the latest and greatest with the exception of the Nova. This Indilinx-based drive debuted well into the controller’s life, so the initial release should have all of the kinks ironed out. Corsair tells us there are no firmware updates for the Nova.
We’re in the midst of overhauling our storage test systems here at TR, a plan that was stalled briefly by Intel’s Sandy Bridge chipset bug. The new suite of tests is coming soon, and it should be worth the wait. In the interim, we’ve whipped up an abbreviated version with a handful of new and old tests that cover the basics.
The block-rewrite penalty inherent to flash memory, the TRIM command designed to offset it, and the last workload an SSD tackled can all impact drive performance, so we’ll provide a little more detail on exactly how we test SSDs. Before testing, each drive is returned to a factory-fresh state with a secure erase. Next, we fire up HD Tune and run a series of read and write tests covering transfer rates and random access times. HD Tune is designed to run on unpartitioned drives, so TRIM won’t be a factor. The command requires a file system to be in place.
After HD Tune, we partition the drives and fire up a series of IOMeter workloads using the latest version of that app. 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 delete that file before moving onto our used-state file copy tests, after which we tackle disk-intensive multitasking. Our multitasking benchmark requires an unpartitioned drive; like HD Tune, it shouldn’t be affected by TRIM.
With our multitasking tests completed, we secure-erase the drives once more and launch a final instance of our scripted file copy test. This procedure should ensure that each SSD is tested on an even playing field—and in best- and worst-case performance scenarios.
We run all our tests at least three times and report the median of the results. We’ve found that IOMeter performance can fall off after the first couple of runs, so we use five in total and throw out the first two. Each drive’s performance over the last three runs has been pretty consistent thus far. We’ve also seen remarkable consistency with our new FileBench copy test, which we’re currently running five times while we tune the scripting. We used the following system configuration for testing:
|Processor||Intel Core i7-2500K 3.3GHz|
|Motherboard||Asus P8P67 PRO|
|Platform hub||Intel P67 Express|
|Platform drivers||INF update 18.104.22.1685
|Memory size||8GB (2 DIMMs)|
Corsair Vengeance DDR3 SDRAM at 1333MHz
|Audio||Realtek ALC892 with 2.58 drivers|
Gigabyte Radeon HD 4850 1GB
with Catalyst 11.2 drivers
Corsair Nova V128 128GB with 1.0 firmware
Intel X25-M G2 160GB with 02M3 firmware
Intel 510 Series 250GB with PWG2 firmware
OCZ Agility 2 100GB with 1.29 firmware
Crucial RealSSD C300 256GB with 0006 firmware
|Power supply||OCZ Z-Series 550W|
|OS||Windows 7 Ultimate x64|
Thanks to Asus for providing the system’s motherboard, Gigabyte for the graphics card, Intel for the CPU, Corsair for the memory, OCZ for the PSU, and Western Digital for the Caviar Black 1TB system drive.
We used the following versions of our test applications:
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 — Transfer rates
HD Tune lets us look at transfer rates in a couple of different ways. We use the benchmark’s “full test” setting, which tracks performance across the entire drive and lets us create the fancy line graphs you see below. This test was run with its default 64KB block size.
Although the 510 Series’ sustained read speed doesn’t match the 500MB/s claimed on Intel’s spec sheet, the SSD manages nearly 400MB/s in HD Tune. That puts the drive comfortably ahead of its closest competitor, the RealSSD C300. Interestingly, the older drives occupy a pretty tight spread between 234 and 243MB/s.
Looking at the line graph reveals that the 510 Series’ transfer rates are a little bit smoother than those of the C300. The Crucial drive’s performance doesn’t oscillate enough to cause concern, but it doesn’t flat-line like the X25-M.
Switching to a write speed test makes the line graph a lot more interesting. Notice that the 510 Series’ transfer rate falls into two brief but distinct valleys that repeat across the drive’s full capacity. With the exception of the Nova, whose write speed follows a sawtooth pattern, the other SSDs have more consistent transfer rates than the 510 Series.
Even at its slowest, the 510 Series is still the fastest of the bunch. The drive’s average write speed is over 50MB/s quicker than its closet competitor, an Agility 2 whose SandForce controller has a secret sauce of DuraWrite compression techniques that speed write performance. If you didn’t think the X25-M was long overdue for a replacement, take note of the fact that the 510 Series’ average write speed is nearly three times that of its predecessor.
Next up, let’s probe the performance of each drive’s onboard cache with a couple of burst speed tests.
Yeah, this one’s not even close. The 510 Series has over 125MB/s on the C300 with both reads and writes. Total carnage.
HD Tune — Random access times
In addition to letting us test transfer rates, HD Tune can measure random access times. We’ve tested with four transfer sizes and presented all the results in a couple of line graphs. We’ve also busted out the 4KB and 1MB transfers sizes into bar graphs that should be easier to read.
Contrast these results with the access times of current mechanical hard drives, and it’s clear that all the SSDs are exceptionally quick. That said, 510 Series lags way behind the leaders with the 4KB transfer size. The drive may be less than two tenths of a millisecond short of first place, but that’s longer than you might think in a world where modern CPUs are flipping billions of bits per second on multiple cores.
Things improve for the 510 Series when we look at the much larger 1MB transfer size. The new Intel drive comes out on top, followed closely by the C300. Although the delta between those two is pretty tight, the others are 1-2 milliseconds behind.
The 510 Series may have been slow with 4KB random reads, but the drive redeems itself with writes. Kudos to the X25-M for stubbornly clinging to the lead, even if it completely tanks with the 1MB transfer size. With 1MB random writes, the 510 Series enjoys a comfortable lead over the C300 and everything else.
TR FileBench — Real-world copy speeds
Our resident developer, Bruno “morphine” Ferreira, has been hard at work on a new file copy benchmark for our storage reviews. FileBench is the result of his efforts, and it debuts today. This shining example of scripting awesomeness runs through a series of file copy operations using Windows 7’s xcopy command. Using xcopy produces nearly identical copy speeds to dragging and dropping files using the Windows GUI, so our results should be representative of real-world performance.
To reduce the number of external variables, FileBench runs entirely on the drive that’s being tested. Files are copied from source folders to temporary targets that aren’t deleted until all testing is complete. Copy speeds were tested with the SSDs fresh from a secure erase and in a tortured used state after more than half a day’s worth of IOMeter thrashing.
To gauge performance with different kinds of files, we tested with four sets. The movie set includes six video files of the sort one might download off BitTorrent. Total payload: 4.1GB. Our MP3 file set uses a chunk of my music archive, which is made up of high-bitrate MP3s and associated album art. This one has 549 files that add up to 3.47GB. The Mozilla file set includes the huge selection of files necessary to compile Firefox. All told, there are 22,696 files spread across only 923MB. Finally, we have the TR file set, which contains several years worth of the images, HTML files, and spreadsheets behind my reviews. This set has the largest number of files at 26,767, but it’s heftier than the Mozilla set with 1.7GB worth of data.
Intel’s focus on improving sequential throughput pays dividends for the 510 Series in FileBench. The drive has a huge lead with the movie and MP3 file sets, where it more than doubles the copy speeds of the old X25-M. Things get a little tighter with the Mozilla and TR file sets, although the 510 Series remains the quickest drive of the bunch when copying smaller files.
The 510 Series doesn’t skip a beat when it’s copying files in a used state versus a fresh one. Neither do most of the other SSDs. However, the Agility 2’s SandForce controller is known for slowly reclaiming trimmed flash pages, which leads to sluggish used-state copy performance with our movie and MP3 file sets. The Agility 2’s used-state copy speeds are more competitive with the Mozilla and TR file sets, suggesting that the additional overhead associated with transferring a large number of small files is enough to mask the lack of urgency with which the SandForce controller clears trimmed flash pages. Note that all the drives copy the larger files much faster than the smaller ones.
TR DriveBench — Disk-intensive multitasking
TR DriveBench 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 set of multitasking workloads that combine common desktop tasks with disk-intensive background operations like compiling code, copying files, downloading via BitTorrent, transcoding video, and scanning for viruses. You can read more about these workloads and desktop tasks on this page of our SSD value round-up.
A new version of DriveBench complete with updated traces is in the works. This old suite of workloads still has some life left in it, though.
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 with each multitasking workload.
Interesting. The 510 Series just eclipses the performance of the C300 when we look at an overall average of five workloads. Both of those drives have a commanding lead over the X25-M, which sits impressively in third place ahead of the much newer Agility 2.
Let’s break down the overall average into individual test results to see if the 510 Series held the lead throughout.
Nope. In fact, the Intel drive pulls up short of the C300 with the majority of our workloads. However, it never really gets dominated by the Crucial drive. The same can’t be said in reverse. With the file copy workload, the 510 Series lays down a beating on the C300 and the rest of the field. Intel’s margin of victory with that one workload is big enough to lift the drive’s overall average ahead of the C300’s.
As a control, we also recorded a trace of our foreground tasks, while nothing was going on in the background.
Without a file copy, this one goes to the C300.
DriveBench lets us start recording Windows sessions from the moment the storage driver loads during the boot process. We can use this capability to gauge boot performance, this time with TweetDeck, Pidgin, AVG, Word, Excel, Acrobat, and Photoshop loading from the Windows startup folder.
Don’t read too much into these results. The startup trace took only 9 seconds to run on the C300, 11 seconds on the Nova, and 10 seconds on the other drives. Real-world Windows 7 startup tests will return in our new storage suite.
Our IOMeter workloads are made up of randomized access patterns, making them perfect candidates to exploit the wicked-fast access times of solid-state storage. The app bombards drives with an escalating number of concurrent IO requests and should do a good job of simulating the demanding environments common in enterprise applications. We tested using the “pseudo random” data pattern, which is IOMeter’s old default and somewhat amenable to the compression mojo built into SandForce controllers. Additional testing with the “full random” data pattern revealed only a minor drop in the Agility 2’s performance, so we’re sticking with pseudo random for now.
Over the last few years, we’ve watched new storage controller drivers (including the Intel RST drivers used in this review) effectively cap IOMeter performance scaling beyond 32 outstanding I/O requests. The Serial ATA spec’s Native Command Queue is 32 slots deep, and more than one drive maker has told us that this queue is rarely full. As a result, we’re only testing up to 32 concurrent I/O requests.
If you were wondering when Intel’s decision to optimize the 510 Series’ firmware for sequential throughput would come back to bite the drive, that’d be right about now. With the file server, workstation, and database access patterns, which are made up of a mix of read and write requests, the 510 Series is nearly the slowest drive of the pack. The C300 and Agility 2 both offer substantially higher transaction rates that rise dramatically as the load increases. There’s much less of an improvement in the 510 Series’ performance when it’s hammered with heavier loads.
Of course, the 510 Series does fare much better with the web server test pattern. That one’s made up exclusively of read requests, with which the 510 is considerably more competitive. The drive may not be able to drop its X25-M shadow, but it easily exceeds the performance of the OCZ and Corsair drives and comes close to the C300’s peak transaction rate. At least in IOMeter, the 510 Series’ shortcomings on the random I/O front seem to be confined to writes.
A number of new solid-state drives are slated to arrive over the next little while, including models based on SandForce’s second-gen controller and others using the same Marvell chip as the 510 Series. Given what’s coming, and the fact that we’re working on an updated storage suite, I’m going to reserve final judgment on the 510 Series until everyone’s cards are on the table. We have, however, learned a few interesting things about the drive today.
Through firmware optimization, Intel has made the 510 Series particularly adept at sequential transfers. Not only did the drive perform well with HD Tune’s targeted read and write speed tests, the 510 Series was easily the fastest when copying actual files in the real world.
The 510 Series’ quick copy performance paid huge dividends with at least one of our multitasking workloads, and the drive was pretty competitive with the rest. However, it’s clear that Intel’s optimizations have cost the 510 Series in other areas. Random reads were a little slow with small transfer sizes, and the drive’s IOMeter transaction rates failed to impress with workloads that mix read and write operations. The 510 Series fared far better with IOMeter’s web server access pattern, which is made up exclusively of read operations.
IOMeter performance is important if you’re running a server, but its workloads are less indicative of what consumer-grade SSDs are likely to face, even in the hands of enthusiasts and power users. Our DriveBench multitasking workloads are probably more representative of what goes on inside desktop PCs. The 510 Series did just fine with those workloads, suggesting that consumers won’t necessarily miss the random I/O performance that Intel is giving up.
If you’re putting together a new system and can’t wait for the dust to settle as other new SSDs come to market, the 510 Series looks like a solid choice. Don’t expect a bargain, though. Newegg is selling the 250GB drive we tested for $615, while a 120GB model can be had for $290. At least that’s better than the cost of Corsair’s latest Performance 3 Series SSDs, which use the same Marvell controller. The 120GB Performance 3 costs $320, and the 256GB version is out of stock. 256GB flavors of the C300 will set you back $485, while 128GB variants cost $265.
As more next-gen drives become available, I expect we’ll see some shuffling on the pricing front. The performance landscape will change, as well—and probably sooner one might expect. Stay tuned.