Intel burst onto the storage scene back in September with its X25-M solid-state drive. This 80GB SSD combined Intel’s formidable chip-making prowess with its years of experience in high-performance storage controller design, and the results were impressive to say the least. Indeed, the X25-M was arguably the fastest solid-state drive we’d ever tested, let down only by the comparatively slow write performance of its multi-level cell (MLC) flash memory chips.
There isn’t much Intel can do about the slow write speeds inherent to MLC memory. However, the company has crafted a new solid-state drive based on single-level cell (SLC) memory chips that aren’t plagued by poor write rates. This latest X25-E Extreme boasts the same 250MB/s sustained read speed as the X25-M, but write speeds have been boosted from a paltry 70MB/s to a much more impressive 170MB/s. Oh my.
Obviously, the X25-E Extreme is going to be faster than the X25-M. Read on to see where the X25-E’s faster write speeds help the most, and in some cases, where they improve performance more than you might expect.
Extreme for enterprise
Solid-state drives use either single-level or multi-level cell flash memory. The former stores one bit per memory cell (a value of 0 or 1) while the latter is capable of storing two bits per cell (with possible values of 00, 01, 10, and 11). Obviously, MLC flash has a significant advantage on the storage density front. However, that advantage comes at the cost of write speeds, which are typically much slower than reads. Intel’s MLC-based X25-M, for example, is capable of reading at up to 250MB/s, but its sustained write speed tops out at only 70MB/s. Single-level cell memory doesn’t suffer such a great disparity between read and write speeds, as evidenced by the X25-E Extreme, which reads at up to 250MB/s and writes at up to 170MB/s.
Of course, the more balanced transfer rates offered by SLC memory don’t come cheap. The X25-M 80GB is currently selling for $621 online, which works out to a seemingly exorbitant $7.76 per gigabyte. But that’s nothing compared to the cost of the X25-E Extreme 32GB, which at $719 online, rings in at an even steeper $22.47 per gigabyte. Solid-state storage isn’t cheap, and single-level cell implementations are about as expensive as SSDs get.
The prospect of shelling out three times as much per gigabyte for the X25-E is certainly daunting, but the drive does offer other perks to justify the premium. For example, its 75-microsecond read latency is 10 microseconds quicker than that of the X25-M. That’s not a huge margin, but within the confines of a modern PC, where bits flip at billions of times per second on multiple processor cores, it’s a notable improvement. To put things in perspective, it’s also worth noting that the access time of a VelociRaptor, which has faster seek times than any other Serial ATA hard drive, is two orders of magnitude slower at 7400 microseconds.
While the X25-E’s faster write speed and quicker access latency are great, it’s on the longevity front that the drive offers the biggest step up over the X25-M. Multi-level memory cells are limited to 10,000 write-erase cycles before they burn out. Single-level memory cells, on the other hand, are good for 100,000 write-erase cycles—a difference of one order of magnitude.
SLC-based flash drives should last much longer than those that use MLC memory chips. Exactly how much longer depends on several factors, including the write-erase content of the workloads involved, the size of the drive, its write amplification factor, and the efficiency of its wear-leveling algorithms. Solid-state drive makers tend not to discuss those last two factors, but according to Intel, they can have a profound impact on a drive’s actual lifespan.
Write amplification refers to the amount of data that must actually be written to a drive to complete a given write request. Say you have a 4KB write request and a drive with a 128KB erase block size. You can’t just erase and re-write 4KB of that 128KB erase block—you have to clear and rewrite the whole thing. The write amplification factor is the actual write size divided by the request size, which in this case is 32. Intel claims that the X25-E Extreme’s write amplification factor is less than 1.1, and that “traditional” SSDs have a write amplification factor of closer to 20.
Wear-leveling is also an important component of flash endurance, as drives spread the love in an attempt to distribute write-erase cycles evenly across available cells. This requires some bit shuffling, and drives must take care to ensure that their wear-leveling algorithms don’t burn through too many write-erase cycles in the process. According to Intel, most SSDs have a wear leveling efficiency factor of three. The X25-E Extreme’s wear leveling efficiency factor is quoted as less than 1.1.
If we combine all the factors that Intel says affect solid-state drive longevity, we come up with the following formula for cycling:
Cycles = (Host writes) * (Write amplification factor) * (Wear leveling factor) / (Drive capacity)
With write amplification and wear leveling efficiency factors of 1.1, and 20GB of write-erase requests per day for five years, we should only burn through 1380 cycles on the X25-E Extreme. The same workload on what Intel defines as a “traditional” SSD, with a write amplification factor of 20 and a wear-leveling efficiency of three, consumes more than 68,000 cycles. We don’t want to rely too much on Intel’s likely pessimistic assessment of the wear leveling efficiency and write amplification factors of other solid-state drives, but other SSD makers haven’t been able to give us equivalent numbers of their own.
The X25-E Extreme’s expected lifespan will, of course, depend on how many gigabytes of write-erase operations are thrown at it. Even with 100GB of write-erase per day, it’ll take more than 72 years to burn through the drive. Couple that with the Extreme’s two-million-hour Mean Time Between Failures (MTBF) rating, and one can probably expect the drive to last.
Like its MLC-based cousin, the X25-E uses a 10-channel storage controller backed by 16MB of cache. Amusingly, the cache is provided by Samsung—one of the biggest players in the SSD market—via a K4S281632I-UC60 SDRAM memory chip. The storage controller is an Intel design that’s particularly crafty, supporting not only SMART monitoring, but also Native Command Queuing (NCQ). NCQ was originally designed to compensate for the rotational latency inherent to mechanical hard drives, but here it’s being used in reverse, because Intel says its SSDs are so fast that they actually encounter latency in the host system. It takes a little time (time is of course relative when you’re talking about an SSD whose access latency is measured in microseconds) between when a system completes a request and the next one is issued. NCQ is used to queue up to 32 requests to keep the X25-E busy during any downtime between requests.
Even with its protective shroud removed, the Extreme looks not unlike the X25-M that preceded it. Both drives use a single circuit board populated with 10 memory chips on each side. Intel makes these chips itself using a 50nm fabrication process. With the X25-E, however, the connection points are covered with what appear to be drippings from World of Goo. No doubt this protective coating has been used to prevent enterprising pirates from, er, installing mod chips, or something.
Thus far, we’ve only touched on the performance benefits that solid-state hard drives can provde, but there are other advantages to moving out of the mechanical world. With no moving parts, SSDs are much more resistant to physical shock. They’re absolutely quiet, too, and typically consume much less power than traditional hard drives. Intel rates the X25-E’s idle power consumption at just 0.06W, and when active, that figure only jumps to 2.4W.
The X25-E’s paltry power consumption will be particularly attractive for the enterprise applications at which the drive is targeted. Indeed, this may be the first enterprise-class product to bring Extreme branding into rack servers. Dude, that’s totally where the X25-E’s power savings will add up, as stacks of drives are combined in RAID arrays where every watt saved will also lower cooling costs for the rack.
We’ll be comparing the performance of the X25-E Extreme with that of a slew of competitors, including Samsung’s SLC-based FlashSSD and Intel’s own X25-M. We’ve also thrown in a collection of the latest and greatest Serial ATA drives from Hitachi, Samsung, Seagate, and Western Digital. These drives differ when it comes to external transfer rates, spindle speeds, cache sizes, platter densities, and capacity, all of which can have an impact on performance. Keep in mind the following differences as we move through our benchmarks:
Barracuda 7200.11 1.5TB
Caviar SE16 (640GB)
Performance data from such a daunting collection of drives can make our bar graphs a little hard to read, so we’ve colored the bars by manufacturer, with the X25-E appearing in a lighter blue than Intel’s X25-M.
Our testing methods
All tests were run three times, and their results were averaged, using the following test system.
|Processor||Pentium 4 Extreme Edition 3.4GHz|
|System bus||800MHz (200MHz quad-pumped)|
|Motherboard||Asus P5WD2 Premium|
|North bridge||Intel 955X MCH|
|South bridge||Intel ICH7R|
|Chipset drivers||Chipset 188.8.131.523
|Memory size||1GB (2 DIMMs)|
|Memory type||Micron DDR2 SDRAM at 533MHz|
|CAS latency (CL)||3|
|RAS to CAS delay (tRCD)||3|
|RAS precharge (tRP)||3|
|Cycle time (tRAS)||8|
|Graphics||Radeon X700 Pro 256MB with CATALYST 5.7 drivers|
Seagate Barracuda 7200.11 1TB
Seagate Barracuda ES.2 1TB
Samsung SpinPoint F1 1TB
Hitachi Deskstar 7K1000 1TB
Western Digital RE2-GP 1TB
Western Digital VelociRaptor 300GB
Western Digital Raptor WD1500ADFD 150GB
Western Digital Caviar Black 1TB
Western Digital Caviar SE16 640GB
Western Digital RE3 1TB
Seagate Barracuda 7200.11 1.5TB
Western Digital Caviar Green 1TB
Samsung FlashSSD 64GB
Intel X25-M 80GB
Intel X25-E Extreme 32GB
|OS||Windows XP Professional|
|OS updates||Service Pack 2|
Thanks to NCIX for getting us the Deskstar 7K1000 and SpinPoint F1.
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 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.
The X25-E scores five points higher than its MLC-based cousin in WorldBench. That makes it the fastest drive we’ve ever tested here, and not by an insignificant margin.
Multimedia editing and encoding
Windows Media Encoder
VideoWave Movie Creator
Although it’s not often much faster than the competition, the Extreme still turns in the quickest times through WorldBench’s multimedia editing and encoding tests. Only in the Premiere test does the X25-E put some real distance between itself and the rest of the field, besting the VelociRaptor by 18 seconds.
The X25-E takes top honors in WorldBench’s ACDSee test, where it’s 34 seconds faster than the X25-M, likely thanks to the faster write speeds allowed by single-level cell memory.
Multitasking and office applications
Mozilla and Windows Media Encoder
Scores are close throughout WorldBench’s office and multitasking tests.
The X25-E bounces back strongly in WorldBench’s WinZip and Nero tests, though. In the former, it’s only a little quicker than the X25-M. However, in the Nero test the Extreme really shines, leading the other solid-state drives we tested by more than a minute.
To test system boot and game level load times, we busted out our trusty stopwatch.
Despite dominating WorldBench, the X25-E is only a middle-of-the-pack performer in our boot time test. It’s essentially tied with the X25-M here, and that puts it more than six seconds slower than Samsung’s FlashSSD.
SSDs dominate our level load tests, with the Extreme curiously turning in a slower Far Cry load time than the X25-M. Even then, it’s still a couple of seconds faster than the VelociRaptor.
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.
FC-Test’s scripted file creation sequences really stress a drive’s write performance, and it’s here that the benefits of SLC memory become readily apparent. The X25-E is more than three times faster than the X25-M virtually across the board, and it’s much quicker than any of the other drives we’ve assembled, as well.
As one might expect, the Extreme isn’t much quicker than the X25-M when we switch to FC-Test’s read benchmarks. The drives offer roughly equivalent read speeds here, although the X25-E is a little quicker with each workload, perhaps due to its slightly lower read latency.
Next, the true file-copy test combines read and write tasks.
With no write performance weakness to speak of, the X25-E dominates our file copy tests in dramatic fashion.
Copying from one partition to another doesn’t slow down the Extreme, either. It’s still the class of the field, by huge margins with each test pattern.
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.
Our iPEAK workloads were recorded using a 40GB partition, so they’re a little big for the 32GB X25-E. The app had no problems running, but it warned us that I/O requests that referenced areas beyond the drive’s 32GB capacity would be wrapped around to the beginning of the drive. Since there should be no performance difference between the beginning and end of an SSD, the results should be valid.
The X25-E continues to make short work of the competition, slicing through our first wave of multi-tasking workloads with apparent ease. Only the X25-M comes close, and even then, only with the one workload that doesn’t stress write performance.
Our second batch of multitasking workloads proves just as fertile ground for the X25-E, which again mops the floor with its rivals. The X25-M comes close to matching the Extreme’s performance with some workloads, but it’s well behind with those that include a file copy operation as the secondary task.
IOMeter presents a good test case for both seek times and command queuing.
Can you guess which IOMeter test pattern is exclusively made up of read operations? I thought so. Only with the web server test pattern does the X25-M manage to hang with Intel’s latest SSD, as both drives roar out to a huge lead over the competition. The M is quite a bit faster than the rest of the field when faced with IOMeter’s other test patterns, too, but it’s not nearly as quick as the Extreme. Based on these results, it’s easy to see why Intel is targeting the X25-E at enterprise environments. No other drive we’ve tested even comes close to matching its performance, and that includes Samsung’s SLC-based FlashSSD.
The X25-E’s IOMeter CPU utilization is much higher than the rest of the field, but that’s to be expected given the fact that the drive is doing significantly more work.
We tested HD Tach with the benchmark’s full variable zone size setting.
HD Tach’s pedal-to-the-floor sustained throughput drag races nicely put the X25-E’s transfer rates into perspective. The Extreme is shadowed by the X25-M in the read speed test, but it’s all alone in the lead when we look at writes. What’s particularly impressive here is that the X25-E actually exceeds its supposed maximum write speed of 170MB/s by nearly 38MB/s. Perhaps even more striking is the fact that the VelociRaptor is at least 100MB/s behind in both tests.
The Extreme retains the lead through HD Tach’s burst speed test, although it doesn’t have much of an advantage over the competition here.
All three of our solid-state drives share the lead here, flexing their huge access latency advantage over mechanical hard drives. This test only tracks random access times down to a tenth of a millisecond, so it doesn’t expose any differences in performance between the SSDs.
While the Extreme’s CPU utilization is the highest of the lot, keep in mind that HD Tach’s margin of error in this test is +/- 2%.
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.
These scores represent the noise generated by our test system as a whole, which is why the SSDs are all pegged at 42 decibels—that’s the ambient noise level of the rest of the system. On their own, flash drives don’t so much as make a sound, regardless of whether they’re idling or under load.
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. Through the magic of Ohm’s Law, we were able to calculate the power draw from each voltage rail and add them together for the total power draw of the drive.
Behold, the power efficiency advantage of SSDs. Mechanical hard drives don’t even come close.
Interestingly, Samsung’s FlashSSD consumes less power than both of Intel’s solid-state drives. We’re only talking about fractions of a watt, though.
I had high expectations for the X25-E Extreme, but even those were exceeded by the drive’s stunning performance. This is without a doubt the fastest solid-state drive we’ve ever tested, and based on just how much quicker it is than Samsung’s FlashSSD, I wonder if it might be the fastest flash drive we see for a while. Intel’s decades of chipset design experience likely deserve much of the credit here, as the X25-E’s storage controller seems to do an excellent job of exploiting all the performance potential available in single-level cell flash memory.
Obviously, the Extreme is at its best when subjected to demanding workloads that combine a highly-randomized mix of read and write requests. The drive’s jaw-dropping IOMeter performance is a testament to just how well it handles heavy multi-user loads, and that’s precisely why the X25-E is perfect for the enterprise environments at which it’s targeted. If all you’re running is a web server that’s going to be slammed with read requests all day, you might want to consider the X25-M. However, that drive is going to run out of write-erase cycles much quicker than the X25-E, which should last ten times longer thanks to the higher endurance of SLC memory.
Lest you think it’s only appropriate for servers, the X25-E’s dominating multitasking performance suggests that it’s a good candidate for workstations, too. Based on the drive’s crushing display in FC-Test, I’d say it’s an attractive option for any system that needs to shuffle files around quickly and often. And let’s not forget that the Extreme operates in absolute silence while barely sipping power.
Of course, the X25-E does have its limitations. For example, the drive won’t dramatically speed up typical desktop applications that aren’t terribly bound by storage subsystem performance. The X25-E’s modest 32GB storage capacity may prove too small for some applications, as well, although Intel is slated to release a 64GB version of the drive next year.
Perhaps the greatest stumbling block for the X25-E Extreme is one that affects all solid-state drives: price. The Extreme is currently available for as little as $719, which works out to a coronary-inducing $22 per gigabyte. But it would be unwise only to consider the drive in terms of cost per gigabyte. Given its target market, it’s far more appropriate to examine cost as it relates to performance. If we look at the Extreme’s cost per IOMeter transaction, for example, we find that it’s a much better value than any other drive we tested today (with the partial exception of the X25-M, but then only with the web server test pattern).
For those who put a premium on storage subsystem performance, the X25-E actually represents rather good value. And as the fastest drive we’ve ever tested, solid-state or otherwise, bragging rights are guaranteed. Finally, a product deserving of not only its Extreme moniker, but Editor’s Choice distinction, too.