review intels 520 series solid state drive

Intel’s 520 Series solid-state drive

Intel has gotten a lot of mileage out of the SSD controller it designed for the original X25-M. This chip made its debut in 2008, back when Justin Bieber was underground. Since then, the same controller architecture has been stretched over three generations of consumer drives. With each refresh have come firmware enhancements and new flash memory chips built using finer fabrication techniques. The latest in that lineage is the Intel 320 Series, which features 25-nm MLC NAND.

Despite this cutting-edge flash, the 320 Series can’t keep up with the performance of its contemporary rivals. That doesn’t seem to bother Intel, whose long-term relationship with the controller has become an open one. Last year, Intel hooked up with a Marvell controller for the high-end 510 Series SSD. Although the 6Gbps controller had already been used by Crucial in the C300, Intel brought its own flash and firmware to the party, plus faster performance.

The Intel 510 Series uses older 34-nm NAND, so it’s long overdue for a transition to cheaper, faster, 25-nm flash. Rather than sticking with the Marvell controller, Intel dumped it for a new muse: SandForce’s SF-2281, which you’ll recognize from SSDs like the OCZ Vertex 3, Corsair Force Series GT, Kingston HyperX, and a stack of others. The SandForce controller has been widely praised for its performance, but much of its early life was tarnished by a pesky BSOD bug.

Intel took its time with the controller, which is why Cherryville, otherwise known as the 520 Series SSD, is hitting the market long after rival drive makers released similar offerings. With only the finest Intel NAND and exclusive firmware improvements, this latest entry into the high-end SSD market promises better performance than existing SandForce implementations and solid reliability that Intel is backing up with a five-year warranty. Is this the SandForce SSD you’ve been waiting for? Let’s take a closer look to find out.

Scuffed outside, SandForce inside
On the surface, the Intel 520 series looks all but identical to the 320 Series. The case dates back to the second-generation X25-M, and there’s good reason for Intel to keep using it. See that black metal rim framing the face of the case? It’s a spacer that brings the drive up to the 9.5-mm thickness common among 2.5″ SSDs, mechanical hard drives, and notebooks. Without the spacer, the metal case measures only 7 mm thick, allowing it to slip into slimmer systems.

The utilitarian design isn’t particularly flashy, but the brushed metal top piece has a nice finish. Unfortunately, the same can’t be said for the bottom of the case, which looks like it’s been kicked across the pavement and run over with a belt sander.

The scuffs are much clearer in person, and they can’t be buffed out. This isn’t just a one-off thing, either; similar scuff patterns can be found on all the Intel SSDs we have with this case design.

Intel says the case meets its “form, fit and function” requirements, of course. For an SSD line that scales up to a 480GB model that sells for $999, aesthetics should probably be a part of that equation. Kingston and Samsung have put together particularly attractive cases for their SSDs, and I’m a sucker for the Ferrari-red paint that covers Corsair’s Force Series GT. Intel would do well to follow those leads or, at the very least, to clean up the design it already has. The distressed look may work for trendy furniture and designer denim, but it’s a poor fit for high-tech products.

But I digress. The real action occurs inside the case, which is where we find SandForce’s SF-2281 controller. Intel has long insisted it continues to develop its own proprietary controller technology, but there’s no telling when the firm will have a next-gen design ready for public consumption. Until then, Intel seems content to respin its existing controller for low-end drives and use third-party solutions to service the high end of the consumer market.

Intel maintains that the X25-M’s controller was designed in-house only because the alternatives available at the time weren’t up to snuff. The Marvell 88SS9174 proved itself robust enough to anchor the Intel 510 Series SSDs, and now the SandForce SF-2281 has been deemed worthy of powering its replacement, the 520 Series.

As far as controllers go, the SF-2281 is about as exotic as they come. We took a closer look at its architecture in our early peek at the OCZ Vertex 3, so I won’t recount all of the nerdy details here. The chip has, after all, been making the rounds for about a year now. It does have a few characteristics worth highlighting, though. Like the Marvell chip behind the Intel 510 Series, the SandForce controller sports eight memory channels and a 6Gbps Serial ATA interface. All the usual memory types are supported by the chip, whose internal buffers are large enough that a separate DRAM cache isn’t required.

Pulling off a cacheless design is a neat trick, but it’s not nearly as intriguing as SandForce’s DuraClass technology, which includes everything from write compression to RAID-like redundancy. DuraClass has persisted through two generations of SandForce controllers, and its inner workings remain a closely guarded secret. One component is DuraWrite, a lossless, on-the-fly compression scheme used to reduce the size of incoming writes from the host. Writing less data to the flash should speed write performance provided the data is sufficiently compressible. Reducing the NAND footprint of incoming writes can also increase the lifespan of the flash, which is bound by a limited number of write-erase cycles. The less data is written to the flash, the fewer write-erase cycles are consumed.

Should individual flash cells or even an entire die burn out, the Intel 520 Series will be protected by RAISE, a RAID-like technology that falls under the DuraClass umbrella. Short for Redundant Array of Independent Silicon Elements, RAISE behaves similar to a RAID 5 array by spreading data and parity bits across multiple NAND dies. Like RAID 5, you have to give up storage capacity in exchange for redundancy. RAISE consumes the capacity of one flash die, which is why SandForce-based SSDs come in capacities like 60, 120, and 240GB rather than 64, 128, and 256GB.

In addition to RAISE, the SandForce controller features an ECC error-correction engine. Another engine handles 256-bit AES encryption. Because this encryption can’t be turned off, we suspect it’s an integral part of the whole DuraClass bit-scrambling process. By default, SandForce-based SSDs are configured with blank passwords to ensure users have unfettered access to their data.

Custom firmware, cherry-picked NAND
The Intel 520 Series uses the very same SandForce controller available to other drive makers. Tellingly, Intel says it’s been testing and validating the chip for more than a year now. Other SSD makers started pushing out drives as early as last spring, and the volume of user complaints and firmware updates that followed make a good argument for Intel’s conservative approach.

By the fall, new firmware code from SandForce claimed to fix the infamous BSOD bug. It seems to have done the trick, but Intel’s response was measured when asked about BSOD errors during a conference call with the press. The company stated that the 7A and 4F blue screens commonly associated with the BSOD bug were “absolutely going to be extremely reduced” on its 520 Series SSD. That sounded a little too measured, but Intel assured us after the call that this new model has been held to the same quality and reliability standards as its other SSDs. Indeed, the 520 Series is purportedly solid enough to appear in systems sold by “tier one” PC makers—a first for SandForce-based SSDs, Intel says.

Firmware, rather than hardware, seems to have been the culprit behind the SandForce BSOD issue. Those tentative about springing for any SandForce-based drive may find some comfort in the fact that the 520 Series’ firmware was “co-defined” with Intel. The firmware enhancements contributed by Intel won’t be shared with other SandForce drive partners, making the 520 Series unique in a market filled with largely cookie-cutter designs. Looks like Intel is following the same playbook it did with the 510 Series, which had its own set of exclusive firmware tweaks. Intel wasn’t keen on revealing details about the finer points of the 520 Series’ firmware other than to say that its garbage collection and NAND management have been improved versus other SandForce solutions.

Intel also differentiates the 520 Series SSD at the NAND level. The chip giant has been making the flash in its own SSDs since the X25-M. Recently, Intel NAND has started appearing in drives from other vendors. Nearly half of the SandForce-based SSDs we’ve reviewed lately have used Intel flash—specifically, synchronous NAND built on a 25-nm fabrication process.

The very same NAND appears on Intel’s own 520 Series SSD, but the individual dies have been cherry picked for the drive. When your fabs crank out a steady stream of 25-nm NAND, you have the luxury of sorting through it and saving the best stuff for yourself. Intel has several grades of 25-nm NAND, including one sold to other drive makers and a high-endurance variety reserved for enterprise-grade SSDs.

All of the models in the Intel 520 Series lineup use 64Gb NAND dies. The number of dies ramps up with the total storage capacity, and the higher-capacity models have more dies per package. Only one die is squeezed into each NAND package for the 60GB model, while there are two dies per package for the 120, 180, and 240GB variants. The 480GB model has four dies per package. These details are important because we’ve found that NAND die configurations can have a big impact on SSD performance, especially with SandForce-based drives.

Intel’s spec sheets for the 520 Series set some expectations for the relative performance of different models in the lineup. They also highlight the performance deltas one can expect going from compressible to incompressible data. SandForce-based SSDs only achieve their peak write performance with data that’s amenable to compression. Here’s how the performance ratings for the 520 Series stack up:

Capacity Max sequential (MB/s) Max 4KB random (IOps) Price
Compressible Incompressible Compressible Incompressible
Read Write Read Write Read Write Read Write
60GB 550 475 430 80 15,000 23,000 12,000 6,900 $149
120GB 550 500 550 150 25,000 40,000 24,000 13,000 $229
180GB 550 520 550 170 50,000 60,000 46,000 13,000 $369
240GB 550 520 550 235 50,000 60,000 46,000 16,500 $509
480GB 550 520 550 275 50,000 42,000 46,000 9,500 $999

As you can see, sequential and random write performance take a big hit with incompressible data. Intel contends that 75% of the file types accessed by “typical office user[s]” can be compressed by “60% or more,” so real-world performance should fall somewhere between the two extremes. PC enthusiasts deal with a lot of data that’s already been compressed—music, image, and video files, for example—but most of those files probably reside on secondary mechanical storage. There isn’t much benefit to having one’s MP3 collection sitting on an SSD.

While the Intel 520 Series’ performance ratings climb with the capacity, the 480GB model’s random-write performance is purportedly lower than that of its 240GB sibling. Intel says higher-capacity SSDs can be slower due to the larger address space required to cover additional storage. That issue apparently doesn’t affect sequential write speeds, which ramp up by 40MB/s for the top model, at least with incompressible data.

Before we move on from the table, check out the prices on the right. Those are for 1,000-unit quantities, so street prices won’t necessarily be lower. Ouch! Comparable SandForce SSDs with 240GB of synchronous NAND can be found selling for less than $400 these days. 120GB models have fallen under $200, and their 60GB siblings are routinely listed at around $100.

Intel is no stranger to selling SSDs at premium prices, and the 520 Series’ five-year warranty at least justifies some of the additional cost. Most solid-state drives, including all of the SandForce-based designs we’ve seen, are covered by three-year warranties. Longer warranties don’t guarantee better reliability, of course, but Intel points to this article as evidence of the low return rate of its SSDs. Intel says the failure rate of the “millions” of its SSDs out in the wild is less than the company’s 0.75% target.

If the promise of better reliability isn’t enough to cover the added expense associated with the 520 Series, perhaps software will help. We first played with the new version of Intel’s SSD Toolbox software at IDF in September. The application comes with 520 Series SSDs and provides an easy way to secure-erase drives and update their firmware. Intel includes a handy life meter to give users a sense of how many write-erase cycles are left in the flash. There’s also a TRIM-powered optimizer that can be set to run on a schedule if you don’t want to wait for the drive to empty eligible flash pages on its own.

Our testing methods
We tried to get a 120GB version of the Intel 520 Series to test, but the company was only able to provide us with 60GB and 240GB models. They’ll give us a sense of how the 520 Series’ performance compares at the extremes of the spectrum, especially versus comparable SandForce-based drives. We have a full suite of test results for 60GB and 240GB flavors of Corsair’s Force Series GT, which uses the same SandForce controller and synchronous Intel NAND. SandForce-based SSDs are also available with slower asynchronous memory, a configuration represented in our test results by the Corsair Force Series 3.

Our mountain of test data also includes a much broader collection of SSDs based on different controller technologies and flash configurations. If you’ve been keeping up with our storage coverage here at TR, the rest of this page will be old news; feel free to skip ahead to the performance results. For the rest of you, we’ve detailed each drive’s essential characteristics in the table below. You’ll want to pay particular attention to how the Intel 520 Series SSDs compare to the Force Series GT drives, which are brothers from a different mother. Also, note how the 240GB model fares against the other high-capacity SSDs, which represent the highest performance levels for their respective drive families

  Interface Cache Spindle speed Areal density Flash controller NAND
Corsair Force Series 3 60GB 6GBps NA NA NA SandForce SF-2281 25-nm Micron async MLC
Corsair Force Series 3 120GB 6GBps NA NA NA SandForce SF-2281 25-nm Micron async MLC
Corsair Force Series 3 240GB 6Gbps NA NA NA SandForce SF-2281 25-nm Micron async MLC
Corsair Force Series GT 60GB 6Gbps NA NA NA SandForce SF-2281 25-nm Intel sync MLC
Corsair Force Series GT 120GB 6GBps NA NA NA SandForce SF-2281 25-nm Intel sync MLC
Corsair Force Series GT 240GB 6GBps NA NA NA SandForce SF-2281 25-nm Intel sync MLC
Corsair Performance 3 Series 128GB 6GBps 128MB NA NA Marvell 88SS9174 34-nm Toshiba MLC
Crucial m4 128GB 6Gbps 128MB NA NA Marvell 88SS9174 25-nm Micron sync MLC
Crucial m4 128GB 6GBps 128MB NA NA Marvell 88SS9174 25-nm Micron sync MLC
Crucial m4 256GB 6Gbps 256MB NA NA Marvell 88SS9174 25-nm Micron sync MLC
Intel 320 Series 120GB 3Gbps 64MB NA NA Intel PC29AS21BA0 25-nm Intel MLC
Intel 320 Series 120GB 3GBps 64MB NA NA Intel PC29AS21BA0 25-nm Intel MLC
Intel 320 Series 300GB 3Gbps 64MB NA NA Intel PC29AS21BA0 25-nm Intel MLC
Intel 510 Series 120GB 6GBps 128MB NA NA Marvell 88SS9174 34-nm Intel MLC
Intel 510 Series 250GB 6Gbps 128MB NA NA Marvell 88SS9174 34-nm Intel MLC
Intel 520 Series 60GB 6Gbps NA NA NA SandForce SF-2281 25-nm Intel sync MLC
Intel 520 Series 240GB 6Gbps NA NA NA SandForce SF-2281 25-nm Intel sync MLC
Kingston HyperX 120GB 6GBps NA NA NA SandForce SF-2281 25-nm Intel sync MLC
OCZ Agility 3 120GB 6GBps NA NA NA SandForce SF-2281 25-nm Micron async MLC
OCZ Octane 512GB 6Gbps 512MB NA NA Indilinx Everest 25-nm Intel sync MLC
OCZ Vertex 3 120GB 6GBps NA NA NA SandForce SF-2281 25-nm Intel sync MLC
Samsung 830 Series 256GB 6Gbps 256MB NA NA Samsung S4LJ204X01 2x-nm Samsung Toggle DDR
Seagate Momentus 5400.4 25GB 3Gbps 8MB 5,400 RPM 204 Gb/in² NA NA
Seagate Momentus XT 500GB 3Gbps 32MB 7,200 RPM 394 Gb/in² NA* 4GB SLC
Seagate Momentus XT 750GB 6Gbps 32MB 7,200 RPM 541 Gb/in² NA* 8GB SLC
WD Caviar Black 1TB 6Gbps 64MB 7,200 RPM 400 Gb/in² NA NA
WD Scorpio Black 750GB 3Gbps 16MB 7,200 RPM 520 Gb/in² NA NA

Our performance data also includes a number of more traditional hard drives. I’ve grayed out the latter in the table and in the graphs on the following pages to focus our attention on how the Intel 520 Series stacks up against its solid-state competition. Neither the mechanical drives nor the hybrids are in the same league, at least in terms of performance.

We used the following system configuration for testing:

Processor Intel Core i7-2500K 3.3GHz
Motherboard Asus P8P67 Deluxe
Bios revision 1850
Platform hub Intel P67 Express
Platform drivers INF update
Memory size 8GB (2 DIMMs)
Memory type Corsair Vengeance DDR3 SDRAM at 1333MHz
Memory timings 9-9-9-24-1T
Audio Realtek ALC892 with 2.62 drivers
Graphics Asus EAH6670/DIS/1GD5 1GB with Catalyst 11.7 drivers
Hard drives Corsair Force Series 3 60GB with 1.3.2 firmware
Corsair Force Series 3 120GB with 1.3 firmware
Corsair Force 3 Series 240GB with 1.3.2 firmware
Corsair Force series GT 60GB with 1.3.2 firmware
Corsair Force Series GT 120GB with 1.3 firmware
Corsair Force Series GT 240GB with 1.3.2 firmware
Crucial m4 64GB with 0009 firmware
Crucial m4 128GB with 0009 firmware
Corsair m4 256GB with 0009 firmware
Intel 320 Series 40GB with 4PC10362 firmware
Intel 320 Series 120GB with 4PC10362 firmware
Intel 320 Series 300GB with 4PC10362 firmware
Intel 510 Series 120GB with PPG4 firmware
Intel 510 Series 250GB with PWG2 firmware
Kingston HyperX 120GB with 320ABBF0 firmware
Corsair Performance 3 Series 128GB with 1.1 firmware
OCZ Agility 3 120GB with 2.15 firmware
OCZ Vertex 3 120GB with 2.15 firmware
WD Caviar Black 1TB with 05.01D05 firmware
Seagate Momentus 5400.4 250GB with 3.AAB firmware
Seagate Momentus XT 500GB with SD22 firmware
WD Scorpio Black 750GB with 01.01A01 firmware
Seagate Momentus XT 750GB with SM12 firmware
OCZ Octane 512GB with 1313 firmware
Samsung 830 Series 256GB with CXM03B1Q firmware
Intel 520 Series 60GB with 400i firmware
Intel 520 Series 240GB with 400i firmware
Power supply Corsair Professional Series Gold AX650W
OS Windows 7 Ultimate x64

Thanks to Asus for providing the systems’ motherboards and graphics cards, Intel for the CPUs, Corsair for the memory and PSUs, Thermaltake for the CPU coolers, and Western Digital for the Caviar Black 1TB system drives.

We used the following versions of our test applications:

Some further notes on our test methods:

  • To ensure consistent and repeatable results, the SSDs were secure-erased before almost every component of our test suite. Some of our tests then put the SSDs into a used state before the workload begins, which better exposes each drive’s long-term performance characteristics. In other tests, like DriveBench and FileBench, we induce a used state before testing. In all cases, the SSDs were in the same state before each test, ensuring an even playing field. The performance of mechanical hard drives is much more consistent between factory fresh and used states, so we skipped wiping the HDDs before each test—mechanical drives take forever to secure erase.

  • We run all our tests at least three times and report the median of the results. We’ve found IOMeter performance can fall off with SSDs after the first couple of runs, so we use five runs for solid-state drives and throw out the first two. The Hybrid drives have also been subjected to five runs, but only in tests that show their performance improving after the first one.

  • Steps have been taken to ensure that Sandy Bridge’s power-saving features don’t taint any of our results. All of the CPU’s low-power states have been disabled, effectively pegging the 2500K at 3.3GHz. Transitioning in and out of different power states can affect the performance of storage benchmarks, especially when dealing with short burst transfers.

The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at a 75Hz screen refresh rate. 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 present transfer rates in a couple of different ways. Using the benchmark’s “full test” setting gives us a good look at performance across the entire drive rather than extrapolating based on a handful of sample points. The data created by the full test also gives us fodder for line graphs.

To make the line graphs more readable, we’ve split up the high- and low-capacity drives. The mid-range SSDs have been dropped from the line graphs, as have the mechanical drives. The results for all the graphs have been colored by drive maker, with the Intel 520 Series set apart from the company’s other SSDs in a brighter shade of blue. Our only exceptions are the hybrid and mechanical drives, which have been greyed out in all of the graphs to focus our attention on the solid-state drives.

The Intel 520 Series 60GB scores a second-place finish in HD Tune’s read speed test, but it’s not quick enough to catch the Samsung 830 Series. Curiously, the 240GB Intel SSD is slower than the 60GB model. The 240GB drive sits in the middle of a huge pack of SSDs that all offer about the same sequential read performance.

Apart from the 60GB drive’s strong showing, the Intel 520 Series doesn’t really differentiate itself from its most direct rival, the Force Series GT. Intel’s latest SSD does, however, have a definite edge in read performance over its 510 Series predecessor.

This time, it’s the 240GB version of the Intel 520 Series that comes out on top. However, its average sequential write speed is still much lower than that of the Samsung SSD, which tops the standings by a huge margin.

The Intel 520 Series suffers from the same problem as other SandForce-based SSDs in this test. If you look at the line graphs, you’ll see the write speed oscillate between incredibly high peaks and much lower valleys. The SandForce drives spend a lot more time in those valleys, which is why their average write speeds are so much lower than their maximum rates. We’ve seen this behavior from two generations of SandForce controllers.

Although Intel’s latest SSD doesn’t really set itself apart here, it does look like an improvement over the outgoing model. The 520 Series 240GB may be only slightly faster than the 510 Series 250GB, but the new 60GB drive is faster than the old 120GB one. Impressive.

HD Tune’s burst speed tests are meant to isolate a drive’s cache memory.

The SandForce controller’s cacheless design certainly doesn’t hurt its performance in these tests. The Intel 520 Series 240GB doesn’t break away from the Corsair Force Series GT, but the 60GB model behaves a little differently. While the smaller Intel SSD’s read burst speed is 18MB/s slower than the 60GB Corsair drive, its write burst speed is 32MB/s faster.

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. Once again, we’ve split up our results for the line graphs.

The Intel 520 Series SSDs offer comparable access times to their SandForce-based relatives. All the solid-state drives are pretty quick in these tests, though. Things only start to spread out in the 1MB test, which runs slower on the older Intel 510 and 320 Series SSDs than it does on the latest 520 Series drives.

At least among the SandForce-based SSDs, there isn’t much to see here. They’re the fastest of the bunch with random writes, and the Intel 520 Series is right there with the leaders.

Again, note the slower access times for the other SSDs in the 1MB test. Also notice just how much longer the access times are for the hybrid and mechanical drives. The differences in access times between the SSDs are much smaller than the gaps between solid-state drives and their mechanical competition.

TR FileBench — Real-world copy speeds
Concocted by resident developer Bruno “morphine” Ferreira, FileBench 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 typical real-world performance. We tested using the following five file sets—note the differences in average file sizes:

  Number of files Total size Average file size
Movie 6 4.1GB 701MB
RAW 101 2.32GB 23.6MB
MP3 549 3.47GB 6.48MB
TR 26,767 1.7GB 64.6KB
Mozilla 22,696 923MB 39.4KB

The names of most of the file sets are self-explanatory. The Mozilla set is made up of all the files necessary to compile the browser, while the TR set includes years worth of the images, HTML files, and spreadsheets behind my reviews.

To get a sense of how aggressively each SSD reclaims flash pages tagged by the TRIM command, we’ve run FileBench with the solid-state drives in two states. We first test them in a fresh state after a secure erase. The SSDs are then subjected to a 30-minute IOMeter workload, generating a tortured used state ahead of another batch of copy tests. We haven’t found a substantial difference in the performance of mechanical drives between these states.

In a factory fresh state, the Intel 320 Series’ copy speeds shadow those of the Corsair Force Series GT. The Intel SSDs are a little bit quicker in some of the file sets, while the Corsair drives pull slightly ahead in the others.

Only with the smaller files of the Mozilla and TR sets does the Intel 520 Series 240GB boast the fastest copy speeds. The old Intel 510 Series 250GB actually offers higher copy speeds in the movie and MP3 sets, which are made up of much larger files. Those larger files copy faster on the OCZ Octane and Samsung 830 Series, as well.

How’s this for an interesting subplot? While the Crucial m4 SSDs are always slower than the Intel 520 Series drives with the smaller sets, their relative performance with the larger sets depends on the capacity. The Intel 520 Series 240GB is always faster than the m4 256GB with larger files, while the 60GB Intel drive is consistently slower than the 64GB Crucial one.

When we switch to a used state, the Intel 520 Series continues to trade blows with the Corsair Force Series GT. One is faster in some sets, while the other has an edge in the rest—unless they’re tied.

The Intel 520 Series’ used-state copy speeds fall farther behind those of its rivals in the movie, RAW, and MP3 sets. Although the leaders don’t change, the Crucial m4 256GB edges out the 520 Series 240GB in the movie and RAW sets. The Crucial SSDs are much slower in the Mozilla and TR sets, though. When copying smaller files, the Intel 520 Series is top dog.

To get a better sense of how copy speeds change between fresh and used states, we’ve graphed the performance deltas as percentages. These graphs are meant to characterize the slowdown in copy speeds that usually results in a used state. We’ve found that some SSDs turn in faster copy speeds after our IOMeter torture test, which is why negative percentages appear in the graphs below—those are the SSDs that are running faster in a used state.

Only a handful of drives post faster copy speeds in a used state, and the deltas are relatively small compared to the performance drops suffered by some of the SSDs. The SandForce-based drives tend to slow down more than alternative configurations. They’re more prone to slower copy speeds with larger files than they are with smaller ones, too.

Once again, it’s hard to pick a clear favorite between the Intel 520 Series and the Corsair Force Series GT. The 520 Series SSDs slow down by smaller margins most of the time, but there are definitely a few exceptions to that trend.

TR DriveBench 1.0 — Disk-intensive multitasking
TR DriveBench allows us to record the individual IO requests associated with a Windows session and then play those results back as fast as possible 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. The individual workloads are explained in more detail here.

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.

DriveBench 1.0 runs particularly well on SandForce-based SSDs, so it’s no surprise to see the Intel 520 Series at the front of the pack. The Intel SSD has to share the top billing with the Corsair Force Series GT, though.

While the Intel 520 Series 240GB is comfortably ahead of the competition, its 60GB counterpart is actually slower than the Crucial m4 64GB. The gap between them isn’t huge, but it does echo what we saw in some of the FileBench results. Let’s see if DriveBench’s individual tests illuminate things a little more.

Through all of DriveBench’s individual tests, the Intel 520 Series 240GB is faster than the Crucial m4 256GB, while 520 Series 60GB is slower than the m4 64GB. We may have a trend developing.

TR DriveBench 2.0 — More disk-intensive multitasking
As much as we like DriveBench 1.0’s individual workloads, the traces cover only slices of disk activity. Because we fire the recorded I/Os at the disks as fast as possible, solid-state drives also have no downtime during which to engage background garbage collection or other optimization algorithms. DriveBench 2.0 addresses both of those issues with a much larger trace that spans two weeks of typical desktop activity peppered with multitasking loads similar to those in DriveBench 1.0. We’ve also adjusted our testing methods to give solid-state drives enough idle time to tidy up after themselves. More details on DriveBench 2.0 are available on this page of our last major SSD round-up.

Instead of looking at a raw IOps rate, we’re going to switch gears and explore service times—the amount of time it takes drives to complete an I/O request. We’ll start with an overall mean service time before slicing and dicing the results.

Only 0.02 milliseconds separate the mean service times of the top three drives in our two-week trace. The Intel 520 Series 240GB drive sits right in the middle of that pack, sandwiched between similarly sized members of the Samsung 830 Series and the Corsair Force Series GT families. The 60GB Intel SSD looks pretty well placed, too. It has a slight advantage over the Force Series GT 60GB and bigger leads over the other low-capacity SSDs.

The Intel 520 Series and Corsair Force Series GT remain close when we split DriveBench 2.0 service times into reads and writes. Those drives have the fastest read service times overall, but the Samsung 830 Series has a lower mean write service time by 20 milliseconds.

Remember when the Crucial m4 was slower than the Intel 520 Series at 240-256GB but faster at 60-64GB? That’s not the case here. The m4’s read and write service times are much lower than those of the equivalent 520 Series capacities.

There are millions of I/O requests in this trace, so we can’t easily graph service times to look at the variance. However, our analysis tools do report the standard deviation, which can give us a sense of how much service times vary from the mean.

Not only does the Intel 520 Series enjoy some of the lowest mean service times in DriveBench 2.0, but it also has some of the most consistent service times. The 240GB model has the lowest variance of the two 520 Series drives, but the 60GB drive easily holds its own against SSDs with similar capacities.

Our IOMeter workloads feature a ramping number of concurrent I/O requests. Most desktop systems will only have a few requests in flight at any given time (87% of DriveBench 2.0 requests have a queue depth of four or less). We’ve extended our scaling up to 32 concurrent requests to reach the depth of the Native Command Queuing pipeline associated with the Serial ATA specification. Ramping up the number of requests also gives us a sense of how the drives might perform in more demanding enterprise environments.

We run our IOMeter tests using the fully randomized data pattern, which presents a particular challenge for SandForce’s write compression scheme. We’d rather measure SSD performance in this worst-case scenario than using easily compressible data.

As we’ve done with the other line graphs in this review, we’ve split our results into two sets of graphs, along capacity lines.

In its 240GB capacity, the Intel 520 Series continues to offer performance comparable to the Corsair Force Series GT. The two have nearly identical transaction rates in the file server, database, and workstation tests. However, the Intel SSD has a clear advantage in the web server test, which is made up entirely of read requests.

Versus the rest of the field, the 520 Series tends to deliver the highest transaction rates at the lower loads levels typical of desktop systems. The Samsung 830 Series boasts better performance when the number of concurrent I/O requests really ramps up, though. The Samsung also dominates the web server test, in which the Intel 520 Series struggles to compete with the Crucial m4. At least the 520 Series delivers substantially higher transaction rates than the 510 Series across the board.

The Intel 520 Series 60GB is thoroughly outclassed by the Crucial m4 64GB in the web server test. However, the 520 Series gets the better of the m4 elsewhere, and it’s comfortably ahead of the old 510 Series throughout.

At this lower capacity, the 520 Series has higher transaction rates than the Force Series GT. The gaps are relatively small, but they’re consistent across each test and load.

Boot duration
Before timing a couple of real-world applications, we first have to load the OS. We can measure how long that takes by checking the Windows 7 boot duration using the operating system’s performance-monitoring tools. This is actually the first time we’re booting Windows 7 off each drive; up until this point, our testing has been hosted by an OS housed on a separate system drive.

The Intel 520 Series SSDs have slight advantages over their Corsair Force Series GT counterparts in our Windows 7 boot test. Really, though, all of the SSDs are very close. Most of the drives are within just one second of each other.

Level load times
Modern games lack built-in timing tests to measure level loads, so we busted out a stopwatch with a couple of reasonably recent titles.

The same trend holds true in our level load tests. Although the Intel 520 Series SSDs are among the fastest, the majority of the solid-state drives are closely matched.

Power consumption
We tested power consumption under load with IOMeter’s workstation access pattern chewing through 32 concurrent I/O requests. Idle power consumption was probed one minute after processing Windows 7’s idle tasks on an empty desktop.

The Intel 520 Series SSDs consume more power than the equivalent Corsair Force Series GT drives at idle but less under load. Those differences will probably only amount to a few minutes worth of battery life in a typical notebook.

Notice how the SSDs aren’t dramatically more power-efficient than the 2.5″ mechanical and hybrid hard drives. The WD Scorpio Black and Seagate Momentus XTs may draw more wattage than most of their purely solid-state rivals, but they’re competitive with an awful lot of SSDs.

The value perspective
Welcome to our famous value analysis, which adds capacity and pricing to the performance data we’ve explored over the preceding pages. We used Newegg prices to even the playing field for everything but the Intel 520 Series, and we didn’t take mail-in rebates into account when performing our calculations. For the 520 Series SSDs, we’ve used Intel’s 1,000-unit pricing.

First, we’ll look at the all-important cost per gigabyte, which we’ve obtained using the amount of storage capacity accessible to users in Windows.

So, yeah, the Intel 520 Series is a tad expensive. The 60GB model costs $0.65 more per gigabyte than the equivalent Corsair Force Series GT. At 240GB, the 520 Series commands at 56-cent premium per gigabyte. The graph plainly illustrates that this is just the latest in a string of pricey Intel SSDs.

Our remaining value calculations use a single performance score that we’ve derived by comparing how each drive stacks up against a common baseline provided by the Momentus 5400.4, a 2.5″ notebook drive with a painfully slow 5,400-RPM spindle speed. This index uses a subset of our performance data described on this page of our last SSD round-up. Some of the drives were actually slower than our baseline in a couple of the included tests, so we’ve fudged the numbers a little to prevent those results from messing up the overall picture.

The Intel 520 Series 240GB slots in between the Corsair Force Series GT and the Samsung 830 Series overall. Those three, er, series are around 100 percentage points ahead of their closest rival, the OCZ Octane. The 520 Series’ advantage over the Force Series GT is razor thin in both the 60 and 240GB capacities, though.

Now for the real magic. We can plot this overall score on one axis and each drive’s cost per gigabyte on the other to create a scatter plot of performance per dollar per gigabyte. To cut down on some of the clutter, I’ve dropped the labels from the mechanical drives, which are cloaked in gerbilesque anonymity. Their data points have been included only to illustrate the overall trend.

Good luck making a case for the Intel 520 Series from a value perspective. Despite performance that is slightly better than or equivalent to competing products, the high cost per gigabyte of the 60 and 240GB variants sours the deal.

Although this analysis is helpful when evaluating drives on their own, what happens when we consider their cost in the context of a complete system? To find out, we’ve divided our overall performance score by the sum of our test system’s components. Those parts total around $800, which also happens to be a reasonable price for a modern notebook.

Even within the context of a complete system, the Intel 520 Series’ price premium is readily apparent. The 240GB model looks particularly expensive on this plot.

Is the Intel 520 Series the SandForce SSD you’ve been waiting for? Maybe. Intel’s latest is a little bit faster than comparable drives based on the same controller and flash memory. The performance gaps between the 520 Series and its synchronous SandForce counterparts seem to be wider at 60GB than they are at 240GB. That said, there isn’t enough of a difference to make a compelling argument for the Intel drive over the alternatives on performance alone.

Factor in the five-year warranty, and the 520 Series starts to look more attractive. Then there’s the peace of mind that comes from knowing Intel spent a year validating the drive before releasing it to the public. Only time will tell if the 520 Series is truly as solid as Intel claims. There’s room for optimism given Intel’s history, but also room for doubt given the SandForce controller.

The notion that the 520 Series will develop a good reputation for reliability may be the only shot the drive has at justifying its exorbitant price tag. Intel’s new high-end SSD isn’t just a little more expensive than comparable SandForce-based SSDs—it’s marked up by a huge margin. The 60 and 240GB models we tested are set to cost $150 and $509, respectively. Corsair’s Force Series GT, which offers nearly identical performance, costs $110 for 60GB and $375 for 240GB. The GT isn’t the only high-end SSD with a much lower price tag than the 520 Series. The Samsung 830 Series scored slightly higher in our overall performance index, and the 256GB version runs only $360 right now. You’ll have to pay $9 more to get the 180GB version of the Intel 520 Series.

Don’t get me wrong. The Intel 520 Series is a great SSD. The models we tested are among the fastest to pass through the Benchmarking Sweatshop, and if Intel has done its homework, they should be free of troublesome issues. The problem is the 520 Series costs 35-40% more than competition that offers equivalent performance. I’d love to be running a 520 Series SSD in my own desktop or notebook, but I’d have a hard time actually buying one—or recommending that others do the same.

0 responses to “Intel’s 520 Series solid-state drive

  1. I’m guessing it’s because there is a good deal of protocol overhead in SATA. Since each read/written chunk needs to be a different transaction, they can’t be as efficiently bundled together like large sequential reads/writes can be. It might be illuminating if someone with an FPGA that can do SATA3 to write a dummy device that responds instantly to any request. Then do some performance testing on it. You’d be completely testing protocol overhead+transfer speed. Compare those results to sequential reads/writes and see what difference there is.

  2. Sure, hybrid drives may make sense for a lot of users, just like HDDs generally make sense for most users due to the relatively high cost of SSDs. But we’re talking about the single slowest component of the computer, which is even slower today relatively speaking than it was 10 years ago.

    I mean, 10 years ago we had HDDs pushing 40MB/s, while the fastest CPU was something like a 2.8Ghz Pentium 4. Today’s quad cores will something like 10x more work in somewhat multithreaded benchmarks. Do you see consumer HDDs pushing 400MB/s? Look back 20 years and you’ll see a similar pattern. SSDs relieve that bottleneck, and that is why we discuss them.

    If you find that the performance of your HDD is not limiting what you can do with your computer, then more power to you. However, please don’t judge us for drinking the kool-aid if you haven’t had a taste yet.

  3. Still going completely one-sided, seems like you are highly biased. You invested in an expensive SSD, right? I have not yet, so I consider the options.

    From what I see, the hybrid technologies are the best of both worlds now – fast, high capacity, low price. Like the Momentus drives, or much better – the Intel SSD cache technology, Smart Response, providing read and write cache. Seems this tech or some variant of it is already in use in few Samsung and Lenovo laptops.

    As a tradeoff I accept the risk of a laptop falling down – I would know it is my fault. But to pay so much more and risk the syndrom of sudden SSD death?

    I use Ubuntu at home since 5 years, so EXT2/EXT3, lately on EXT4. Almost never hearing the drive in my laptops (unlike the fans). No need to defragment – the current drive is 640GB, only 40GB used.

    At work I do use Windows and NTFS – but defragmentation is not on option: 700GB of data, about 30 GB of many many files changed daily. The disk access is slow, true. When the price of 1TB SSDs comes down to 200 euro, my company may consider buying me SSD 🙂

  4. There are numerous SSD makers using basically the same combinations of controller, NAND, and firmware. The testing we do is very time-consuming, so we tend to restrict our coverage to drives that offer something unique. As far as I’ve seen, OWC drives behave just like other SandForce-based offerings that we’ve already reviewed.

  5. Just curious, does TR have something against OWC SSDs? They seem to perform pretty well in other reviews, but I would like to see some of their drives get the Tech Report Treatment… A cursory glance in the Storage subsection shows no mention anywhere of OWC. What gives?

  6. [i<]- I do not defragment my HDDs. SSDs have their own problems with filling up.[/i<] Then you have [b<]massive split I/Os[/b<] and you suffer greatly from it. Sucks to be you. Fragmentation is an issue on several fronts: 1. It increases as drives fill up and its impact is felt more and more as free space dwindles. You might as well never use a mechanical beyond 80% use for anything serious. Go ahead and fill up a mechanical with 81% filled up small and large files, then do anything write heavy with them. Measure performance. DISMAL. Enthusiast benchmarks never go into this, from what I've seen. 2. It is a performance penalty that is impossible to fully remove on NTFS systems. You either have zero fragmentation because you defragmented (which incurs a performance penalty) or you have split I/Os because you haven't defragmented. SSDs do have a performance penalty pre-trim but it still gives way better performance than any mechanical and doesn't result in split I/Os. 3. There is a performance penalty when removing fragmentation in Disk I/O AND CPU use. [i<]- Drop your laptop - the SSD would be ok, but the case and the screen?[/i<] Possibly fine. If you drop or jar a moving/in use mechanical you will almost certainly lose data/the drive. [i<]- Heat / noise - laptop drives from the last few years are silent, consume the same amount of power as SSDs = same heat.[/i<] They are absolutely [url=<]not silent[/url<].

  7. Well, compared to a magnetic drive or just about any other device in your computer the power consumption for SSDs range from extremely low to absurdly low.

    They probably just bothered to build in some power consumption features in the controller. This could even be firmware level stuff on drives with identical controllers. Others probably just skipped worrying about power consumption thinking the extra watt or two isn’t really that important compared to how much the entire PC uses. I would tend to agree. The Intel 520240GB is 2.8 watts at full tilt, and the worst SSD, the Samsung 830 is 4.8 watts at full tilt. A 2 watt difference running 24/7/365 is only ~17kwh for an entire year, which at our local power rate of $0.11/kwh is about $1.93 for an entire year. At idle the biggest difference is 0.5 to 1.5 watts, about $0.95 for an entire year. I don’t see how that difference could ever be a consideration for a consumer. Maybe if you’re trying to build a computer to shoot into space and run it off solar panels… I dunno.

  8. As indeego has illustrated, adding the hybrids to the value charts isn’t going to change the value proposition for the Intel 520 Series SSD, which is the focus of this review. If you want our take on the Momentus, check out our review from November: [url<][/url<] We like the drive and gave it a TR Recommended award based on its value proposition for notebooks. I think you're missing a crucial characteristic of the Momentus, though. Its solid-state cache only works with reads--writes are completely unaffected, and they make up a rather large component of real-world I/O. 36% of all the I/O requests in our two-week DriveBench 2.0 trace are writes.

  9. Fortunately most people spend their time reboot their PC. It’s almost like a game!

    Because of that you are correct there is no good reason to pick anything but a normal HD.

  10. PCIe “addon” cards are ripe for headaches for any OS with driver support. What the SSD makes are working on is something that sits on the PCI-e lane and appears to the EFI/BIOS/OS as storage. Physically, it won’t be SATA, but I’m sure it’ll have some form of sata-like ruggedness.

    SATA is really the limiting factor for these drives. SATA was built for mechanical drives and so many pieces of functionality are useless on a flash-based setup.

    The sooner some form of standard is adopted (i.e. free/low cost licensing on chipsets/boards, along with driver support for every OS), the sooner your performance numbers will be achievable.

  11. Because to do that, the peak ability on easier use cases would be so much greater than the I/O interconnect that the manufacturer would be silly to over design by so much. The drive would cost an arm and a leg, use a lot more power, and noone but you would buy it.

  12. Your points are very one-sided:
    – I do not defragment my HDDs. SSDs have their own problems with filling up.
    – Drop your laptop – the SSD would be ok, but the case and the screen?
    – Heat / noise – laptop drives from the last few years are silent, consume the same amount of power as SSDs = same heat.

    Boot time – [url<][/url<] - Momentus 750 GB => 9.7 sec - all SSDs tested => from 7.6 sec to 9.6 sec sec edit: 0.1 sec difference in boot time... Does it make sense now to have a drive of 6x less capacity? Not so much. Not to mention the danger of BSOD and complete data loss when using SSDs - which is everyday news.

  13. [url<][/url<] Does that not scale enough for you, or do you really think that removing synthetic [i<]really[/i<] makes all that big a difference? Look at the real-world benches (TR Drivebench 1/2 and boot speeds) You are saving 20 seconds per boot going SSD. The Seagate is great for mechanical-class drives, crap compared to SSD drives. Getting a seagate doesn't make sense when you factor in the [url=<]cost of the entire system as a whole.[/url<] Now something reviews don't go into much about SSDs: - Fragmentation overhead/maintenance on mechanicals is a serious pita. - Shock damage is a concern with mechanicals. It's not at all with SSDs. You can drop your laptop and your SSD (and data) will be fine. - Less heat/noise with SSD. Going back to mechnicals is like a "why did we ever put up with this before" moment.

  14. Actually my request is from yesterday, more than a day ago. Never mind.

    … Anyway, I see, Momentus XT drives will be very close to Caviar Black, as the overall performance includes all synthetics. I wanted to see them compared only in the “real-life” benchmarks / cost / GB…

  15. And once you’ve behaved well for as many months as you behaved badly, people may be willing to give you the benefit of the doubt. Reputations can be repaired, but it doesn’t happen overnight.

  16. Your first request was just 2 hours ago, it’s well after normal working hours even on the west coast, and now you’re already back to ask about this again? Your expectations seem a tad unrealistic..

  17. C’mon.. be fair! I haven’t been trolling here that much anymore; I’m saving that for SemiAccurate.

    I’ve realized that here I can get a lot of good stuff (info, viewpoints), so I’ll try to give some, too. I understand that it’ll take a while for some people to forget my “wild” past, but I’ve been behaving better lately, wouldn’t you say..?

  18. Thanks for the heads up. The [url=<]first page of the Cherryville review on AT[/url<] serves up some really tasty morsels. While I don't visit their site anywhere near as much as I do TR, Anand does make invaluable contributions to the PC hardware community. The respect he commands in the industry enables him to squeeze his contacts for all kinds of valuable information that we would not have access to otherwise.

  19. Not sure about SF using Intel’s firmware knowlege for non-Intel drives… At least not for any near future – I’ve read on Anandtech that Intel supposedly has some agreement with SF that Intel-modified firmware will remain exclusive for some time (supposedly until the release of 3rd generation of SF’s controller). Not sure if it’s actually true but makes perfect sense and I would do the same in Intel’s place.

  20. I suppose that method is as good as any. It’s a good idea and I’m happy that you’ve gone through the trouble.

    Obviously what this means, however, is that drives that do a lot better in one metric, yet only perform average on others will have a higher overall performance index that may not be reflective of it actually being an overall better drive in most scenarios. The Samsung 830 may be good example. While a good drive overall, it’s excellent leading write performance probably obscures it’s more average performance in other areas.

    From the benchmarks included in this index, it looks like you are tailoring the index toward best all around performance as a general workstation boot drive, which makes sense. It would be interesting, however, to investigate rankings of said benchmarks. IE, some benchmarks are likely to be more closely proportional to real-world performance deltas than others and could be weighted accordingly. Weighting may also allow for the creation of indices aimed at different usage scenarios (ie, boot drive for gaming PC, vs server drive for web hosting, etc).

    The utility obviously hinges on the validity the assigned weights, which in turn involve determining how closely a benchmark correlates to a given set of real-world applications – which, with the exception of load time benchmarks, is probably not that straightforward to quantify). However, if feasible, would provide single performance indices that may better reflect real-world performance differences across different usage models.

  21. When SATA4 gets here?

    There’s very little point in engineering these drives to be internally faster than SATA3. If you need more speed than that, there are many attractively priced PCIe solutions available.

  22. Dissonance –
    Really nice review, but I still miss having a Velociraptor included in the chart. The 600GB Velociraptor seems to have price parity with the 120GB Intel 520 SSD. Or is it that the other mechanical drives is close enough in all but access time that there isnt worth having yet another level of reference?

    Before I got my X25-M I had a velociraptor, I made a whole lot of sense then, and in some cases, it does now. My main comp still has the velociraptor as a secondary drive for the large storage and temporary stuff since the SSD is to tiny to both hold installed software and working data.

  23. Samsung’s 830 and Crucial’s M4 come to mind. In the SATAII only front, the older Indilinx drives performed well in their class and were quite reliable.

  24. Actually, I think you’ll find that most people here would trade some performance for lower prices.

    Afterall, [url=<]your needs have been met[/url<] for some time now.

  25. I band-aided an old Athlon XP box (running WinXP) with an SSD-based page file, because it could only handle 2GB of RAM reliably and the working sets of the tasks it was running frequently blew past that into the page file. Since this was purely a band-aid for a system that was slated to be retired relatively soon, I just used the 40GB Intel 320 (which happened to be a very good deal at the time). The difference was dramatic, especially in certain very-noticeable operations like starting a new app when the system was already overcommitted. (Since the page file was only occupying a fraction of the SSD, I also redirected temp files and things like the browser cache there, though I never investigated how much difference that made)

    Pages are 4K*, and access is often fairly random (pages tend to be non-contiguous in the page file), so the 4K Random tests that you see in many reviews are a reasonable indicator of performance. The low access time (latency) makes more of a difference than raw STR throughput, especially in the improved responsiveness you can actually see in interactive tasks. Access time hasn’t improved much with newer SSDs even as STR has skyrocketed, so for this usage case the newer SSDs don’t offer as much of an advantage as their raw specs might suggest.

    * Yes, there are other page sizes. But for the purposes of this discussion, 4K pages are all that matter.

  26. Yea its not really any faster conclusively than any other decent SF2281 based SSD right now, but it sure costs alot more. And its not like Intel have been immune to bugs and early deaths of recent years.

  27. No doubt that I was RAM constrained. And you’re probably right that buying more RAM makes more sense. What makes even more sense is just writing more efficient programs to begin with 🙂 but the academic question remains of whether there are noticeable differences between SSDs in their performance in this setting.

    My SSD is one of the first to use the SandForce controller. I bought it in the second half of 2010 from OWC.

    I’ve noticed that SSDs appear to be closing in on PC100 SDRAM specs, at least in terms of sequential read/write speeds (800 MB/sec for PC100 SDRAM, iirc). But I guess SSDs are probably still a long ways off on random read/writes?

  28. Yeah, that was my experience, too. I didn’t notice a responsiveness issue, but I saw that CPU-utilization had fallen dramatically for the 10 concurrently running programs, and that prompted me to check what was going on. It’s then that I realized I had blown through all the RAM. Activity Monitory was reporting disk read/writes of over 100 MB/sec, so I knew I virtual memory was kicking in.

  29. Intel must have heeded the call. The [url=<]510 series[/url<] has dropped $40, which still doesn't make it the best value ($2/Gig) in the pack, does make it more palatable, however.

  30. My box has 6GB of ram and an SSD. I can easily max out the amount of RAM in use (VMs make this cake) and found that my system is still very responsive even with heavy disk swapping. Heck, I didn’t know my disk was swapping until I checked on how much RAM was in use.

  31. My thoughts are you were possibly RAM constrained (you didn’t tell us cached/available/free), and more RAM is a better bang/buck in terms of speed than SSDs for memory constrained applications rather than increasing pagefile or finding fastest pagefile drive.

    [quote<]"But it left me wondering if some of the newer SSDs might provide a noticeable benefit over my older SSD in situations like this. "[/quote<] How old? Will latest provide faster perceived performance than the 2009 generation? Absolutely. Faster than 2010-2012 generation? Unlikely much of a perceived difference. The SSDs you see here thrive in multiple I/O environments, much harder on drives than a single app stressing a pagefile will provide.

  32. I’d guess it would strongly correlate with random reads and STR writes. Microsoft has said that that is how the majority of the IO to the pagefile happens, and notes that SSDs are very well suited for pagefiles. Which SSD works best is probably decided by how large of a block is usually read from the pagefile, some drives work very well at 4K reads but struggle a lot more with 1K reads.

  33. Sorry if this has been covered and I’ve missed it, but how about testing the performance of SSDs in situations where a computer is severely RAM constrained and starts going swap crazy?

    I ran into such a situation recently running 10 concurrent instances of a program and each instance unexpectedly started demanding between 2 and 5 GB of RAM (I only have 22 GB of RAM), so naturally it was time to hit the swap file. It ultimately turned out that there was a (*much*) more efficient way to write these particular programs, so I solved my problem that way. But it left me wondering if some of the newer SSDs might provide a noticeable benefit over my older SSD in situations like this.

    Any thoughts?

  34. Not really doing much for me Intel… since you can’t really test for the BSoD bug at TR, there really isn’t a whole lot of appeal to this drive besides the extra warranty. I guess that’s the way it is with all drives based on the same controller unless the manufacturer really botches things. SSD reviews are seemingly falling into the same category with video card reviews.

  35. Given the number of vendors peddling SSDs nowadays (containing chips from just a few manufacturers like Intel/Micron and Samsung), I’m not surprised. Anyway, why not check out [url<][/url<] They specialize in reviewing many mech HDDs or SSDs. If you want those kind of reviews, I think SR is a better place to visit (and no, please, I'm not saying this to make TR look bad in any sense).

  36. For most folks, I imagine the relevance of this announcement to their computing life would be making SSDs go down in price. Like anything else, there’s a ceiling as to how much people will pay for something like an SSD. This new offering from Intel will probably command a high price, forcing other similarly priced offerings from other companies (or Intel itself) to get priced lower, and consequently pushing lesser SSDs down in price as well. That’s good news for you and me, I imagine.

    As the prices of mech hard drives continue to fly at their current stratospheric levels (I guess we all got spoiled by the low prices before the Thailand flood), price drops in the world of SSDs can’t come at a better time. And [u<]IF[/u<] Intel prices these new babies well enough (read: affordable for mere mortals), it'll really do wonders for SSD affordability in general.

  37. Is there a reason that Mushkin drives are never tested?

    I’m really curious to see the numbers they pull, though I realise there is only so much “secret sauce” one can put into a sandforce SSD.

  38. Even with lesser companies than intel the component supplier usually is in no position to force such licensing terms on their customers. They’d get shoved out the door if they tried.

    Those intel-specific firmware tweaks are almost certainly closely guarded intel IP and Sandforce has zero access to them.

  39. Out of my budget.

    If I was in the market for a 256GB, I think it would be between the M4, 830, and corsair Pro

    [url<][/url<] Crucial seems to stay on top of any firmware, which is good support. I heard Corsair's customer support is good, but I don't know about technical support (firmware) and Sammy has good hardware, but whether they will support it down the road with good software is another question. Yet it's nice to see Intel using sandforce, it should help firmware optimization, and then Intel will buy LSI/Sandforce anyway lol

  40. Considering that price, it almost sounds like Intel decided to bundle the drive with an Extended Warranty, forcing every customer to buy it.

  41. Another bump for “reputation matters”. I paid something like a 40% price premium last fall for an Intel 320 and the main factors were Intel reputation, and five year warranty.

    That said, while I have no interest in jumping on the Neely-bashing bandwagon, if you give out a lot of crap, you get a lot of crap. Cause and effect.

  42. Wrong – I have an OCZ Onyx in one build. But when I can get an 80GB 320 from Newegg for $70, it’s kind of hard to resist…

    I wouldn’t buy 520 – it’s overpriced even compared to other high-end competition. I still believe that top SSD performance is overrated, and not worth the money. I agree with OAS that TLC has promise… for an SSD, a price reduction is much more attractive to me than performance upgrade.

  43. That’s such a good point. You could get an 830, and put the dollars difference into your next video card. Very good point.

  44. [quote<]then you can bet your bottom dollar that Sandforce will be using Intel's firmware knowledge to fix the issue.[/quote<] Why on earth would intel willingly provide Sandforce/LSI with the know-how that sets intel apart from their direct competitors?

  45. In the real-world benchmarks Seagate Momentus XT are just like SSDs, yet their price and capacity are on HDD levels – clear win on speed, capacity, price (and even reliability?)

    Can you also add them to the performance-per-dollar charts?

  46. [quote<]it's-command-line-so-no-average-person-will-ever-do-it simple (linux)[/quote<] Right there you show you have absolutely no clue to what I was referring to (or to the fact that pretty much in eveything linux has a gui as well).

  47. Agreed. Intel’s reputation counts for a lot. That said, when it comes to *my* home system, well, a Samsung 830-series drive is much more tempting due to the value proposition and the (perceived) reliability of the Samsung drive…

  48. If another sandforce firmware bug shows up in OCZ & Corsair drives but not Intel 520 drives, then you can bet your bottom dollar that Sandforce will be using Intel’s firmware knowledge to fix the issue.

    As it is, I’m happy not to pay a 45% premium for my SSD’s thanks. I treat all SSD’s as volatile anyway, which is realistically what you should do for any disk, SSD or mechanical.

    Backups > hope.

  49. Good news: The price for this SSD will not awaken The Slumbering Krogoth.

    More seriously: If you are really hellbent on getting a sandforce controller in an SSD, this one is more expensive but at least you get a real warranty and some assurance that somebody actually tested the drive before throwing it out on the market. (I’m looking at you OCZ!)

    Of course, a better option would be to get a good drive that doesn’t use the Sandforce controller, and fortunately there are a bunch of those available.

  50. The baseline is now a 5,400-RPM Momentus 5400.4. You’re right about the index, though. We do a percentage increase for the tests we pull from the suite and then take the harmonic mean of those percentages.

    I’m also looking into doing a Performance Pro review and a few other things on the SSD front to fill out holes in our dataset.

  51. At the price point Intel has chosen to operate, it does matter. It is perceived value. Given two equally performing products, the average Joe will take the cheaper product unless there is a perceived value in the more expensive offering. Intel seems to think their name is perceived value enough, and they may be right. They may also be very wrong.

  52. Is the constant Neelyprodding ever going to relent?

    I am not an Intel supporter, nor do I particularly love any of their products, but I can still answer why people would choose this product. This product has a Tier 1 vendor support. No other Sandforce controllers can say the same. There are plenty of people who will pay a price premium for a product with implied/endorsed reliability and high performance.

    Personally, I take Corsair or Samsung and throw ever more dollars into other areas of my system, but I get the sentiment.

  53. I’d rather see a 128 GB Performance Pro compared against a 128 GB 830, in a two-fer review, so that we’d also get to compare the Samsung against its larger sibling. And thus we’d have reviews of the “sweet spot” capacities more people are interested in (or can afford).

  54. Seems like even Intel’s vast resources couldn’t accomplish much more on the performance front than what others have done with the SF-2281 already. Maybe reliability will be better though.

  55. Thanks for the great review. Sorry if this has bee addressed before:

    I’m curious as to how you calculated the overall performance index showing percentage increase in overall performance from the Caviar Black baseline. Do you simply calculate the percentage increase in performance of the measure for each benchmark and then take a (harmonic) mean of this percentage increase? Or is there something more complicated you do with each score before aggregating them into a single percentage increase above the baseline?

    I agree with the conclusion that the value proposition makes these drives disappointing for the time being anyway. Any chance of a Corsair Performance Pro review (preferably in 256GB to compare with the Samsung 830 review)?

  56. Can someone explain the wide range of power consumptions for SSDs? There’s a difference of 2x for similar capacity SSDs.

    If Intel were at the bottom of the power consumption pile, then I could attribute this to their superior process engineering team, but Intel is middle of the pack. How can the active power consumption vary so much for products that are identical for 90+% of their silicon (NAND flash cells)?

  57. I disagree about the aesthetics of a drive being a priority: this really, really doesn’t matter to me at all, and that makes logical sense because it’s facing the bottom of my case. I care just as much what the bottom of my PC’s rubber “feet” look like.

    Aesthetics matter very much when it comes to mice, keyboards, screens and so on, but if there’s any money to be saved by scuffing the hell out of the bottom of the drives, I say do it.

  58. Not really interested in fighting over the difference between simple (Windows 7) and spit-for-brains simple (OS X) and it’s-command-line-so-no-average-person-will-ever-do-it simple (linux). The point is the disk image backups are f-word-ing simple and everyone should be doing them and that mitigates any “high risk” that any of these SSDs have of failing.

  59. Great performance, but at that price… I’ll stick with the even faster and cheaper Samsung 830 256GB.

  60. I would definitely not go with the 520 series drives at these prices. I’d get one of the other drives and I’d have a weekly or even daily hard drive image scheduled and I have a whole box of spare mechanical drives that I could restore the image to in the ( very unlikely ) even of failure.

    Everyone should have a good hard drive image schedule implemented. Now that the feature is build into Windows 7 there is absolutely no excuse. And that stands regardless of what hard drive you have in your machine – mechanical, solid state, Intel, or OCZ. With Windows 7 it is, in fact, very simple and totally painless to have a hard drive image made every day, so a drive failure would only set you back a couple of hours to totally recover from.

  61. I may be wrong, but I think Intel is one of only a couple manufacturer’s giving a 5 year warranty with their drives, so that is part of the equation. Still too much for me, but for some that warranty will help bump them to the/near the front.

  62. I could get a 7950 and have change to spare for jeans and a coffee rather than THIS!…..

    And I thought [i<]THAT[/i<] was too [s<]expensive[/s<] [s<]exorbitant[/s<] [s<]blood sucking[/s<] awesome.

  63. Intel isn’t free from the SSD unreliability stain. I know they spent time trying to get it right but I am going to reserve judgement on that until after this drive has been outdated in a few years.

  64. Price = ouch. However for me reliability > all when it comes to storage, so this drive has some appeal to me.