by Ilya Gavrichenkov
10/24/2012 | 09:44 AM
We are usually not very thrilled about reviewing SSDs with 120 or 128 gigabytes storage capacity. This capacity is just not good for showing what a particular SSD design can do. Today’s SSDs use eight-channel controllers and 64-gigabit flash memory chips, so the controller can only use 2-way interleave in a 120/128GB SSD. The controller’s bandwidth is thus not utilized fully, which explains the difference in practical performance between SSDs of different capacities. The gap may vary depending on the specific hardware platform, but there is a gap nevertheless. The SandForce platform is especially susceptible to this effect since it doesn’t incorporate a DDR SDRAM buffer. As a result, the difference in speed between 120 and 240GB models, for example, is quite conspicuous.
In fact, we don’t have to look far for examples. We only have to compare the specified sequential write speeds of different-capacity SSDs from the same series. Here’s what we have with the Corsair Force GT series which we often use in our tests:
The diagram doesn’t begin at 0 on the Y axis, yet you should keep it in mind that we compare the officially specified speeds which are measured with compressible data. In real-life applications the difference in performance is going to be much bigger. We can illustrate this by benchmarking 120GB and 240GB Corsair Force GT drives in CrystalDiskMark 3.0.1 using incompressible data:
The picture says a lot, we guess. The SSDs differ twofold in terms of sequential writing. Their performance with 4KB data blocks differs considerably, too. It is certainly going to show up in practical applications.
That said, we still can’t help testing 120 and 128GB SSDs. They are just very popular. Their capacity is optimal for a system disk, for example, and they will anyway be much faster than any conventional hard disk. But again, if you need maximum performance, such drives won’t be the best option.
As you may have already guessed, we’ll test 120 and 128GB SSDs today. These are products from Silicon Power and Transcend. Both represent somewhat unusual implementations of the second-generation SandForce platform. Their flash memory type isn’t typical and the Transcend also has a nonstandard capacity of 128 rather than 120 GB. It will give us a chance to check out the influence of the 13% reserved space on the performance of SandForce-based SSDs.
Silicon Power isn’t a first-tier brand yet. This Taiwan-headquartered firm has only got some reputation as a flash card maker while its solid state drives aren’t widespread. In fact, there’s nothing odd about a company that works closely with NAND flash manufacturers trying its luck on the SSD market as well. It’s the same business, actually. However, promoting SSDs is a much more daunting task than actually making them since their consumers are well versed in computer technologies and can only be attracted by something special and innovative.
And that’s where Silicon Power may have some problems. Like most other smaller SSD makers, the company uses LSI’s SF-2281 controller which is hardly the best solution available right now. The model range of SATA 3 drives offered by Silicon Power currently includes four models but, according to the not-very-informative official website, they only differ in exterior design and accessories but use the same hardware – synchronous flash memory with Toggle Mode interface.
The model we’re going to test is Velox V60 120GB. Its cardboard box is all covered with promo slogans but offers no useful information about the product itself.
The SSD is additionally packed into a plastic case. Besides it, the box contains mounting screws, an installation guide, and an adapter that helps install the 2.5-inch drive into a 3.5-inch bay of a system case.
The exterior design of the SSD itself is somewhat unusual. The case is made of brushed aluminum and has the cold gray color of the metal. The model name and capacity are written in white paint. The manufacturer’s logo is pressed out in the top panel. So, the SSD has an original appearance and lacks conventional stickers.
The case has a standard height of 9 millimeters, so the Velox V60 won’t fit into ultra-slim mobile computers. The manufacturer offers another SSD series for such applications. It is appropriately called Slim.
Now let’s take a look into the case. There seem to be nothing special there at first sight: a reference PCB with an LSI SF-2281 controller and eight memory chips mounted on one side. There are places for an additional eight chips on the other side of the PCB, but they are empty in our SSD. Obviously, they are used for higher-capacity products in the Velox V60 series.
The only interesting thing about the hardware configuration is the type of the NAND flash memory chips. These are SanDisk SDZNPQBHER-016GT components which have but recently got some popularity. LSI/SandForce engineers have recently proposed a new version of their platform which uses 24nm flash memory with Toggle Mode interface. SanDisk being a major manufacturer of such memory, recently released SSDs with SF-2281 controllers often come with SanDisk chips. We saw the same components in our Corsair Force GS review, for example, so the Velox V60 is overall similar to Corsair’s Force GS series.
Each of the eight memory chips inside the Velox V60 120GB contains two 64-gigabit 24nm NAND devices, so the SF-2281 controller uses 2-way interleaving to access them. Thus, the Velox V60 is similar in its logical structure to SSDs with synchronous ONFI flash but uses lower-level interleaving than SSDs with older 32nm Toggle Mode flash.
After our immersion into the inner world of the Silicon Power Velox V60 120GB, we can give you a sum of its specs:
The specified speed of sequential operations is expectedly too high as is typical of SandForce-based products. As for capacity, the user can only access 87% of the internal flash memory, which is normal, too. The remaining 13% is reserved for the RAISE technology, for replacing failed memory cells and for the garbage collection technique. Again, all of this is perfectly normal for a decent SandForce-based SSD.
Silicon Power currently suggests that you use reference firmware version 5.0.2 for the Velox V60 series. Unfortunately, that version has obvious problems with the TRIM command. SandForce has already issued newer modifications but Silicon Power doesn’t hurry to update its products. So, the Velox 60 is going to lose more speed after hard use than similar products from other brands until Silicon Power adapts the updated firmware to its SSDs.
Being larger and more renowned than Silicon Power, Transcend was started up in the late 1980-ies and has been focusing on memory-related products ever since. Of course, Transcend couldn’t help producing devices with NAND flash memory which are so popular nowadays. Like Silicon Power, Transcend has selected the SandForce platform, taking its place in the second tier of SSD makers. It means that there’s nothing original about Transcend SSDs. They are similar to many products of this class from other brands.
Currently Transcend offers two consumer-class SATA3 SSDs: SSD720 and SSD320. We’ve managed to get a senior model for our tests. Its packaging provokes no emotions. There’s a lot of captions around the black cardboard box which give you information about the product, both useful and not. The box is somewhat thicker than is usually the case with SSD packaging.
This doesn’t indicate some extra accessories, though. In fact, there’s nothing in the box except the SSD and its user manual. There is even no adapter to install it into a 3.5-inch bay of a system case. The interior of the box is just filled with air and the plastic wrap that protects the SSD from damage during transportation.
The SSD itself is an aluminum brick, powder-coated in the popular “wet asphalt” style. The height of the brick is 7 millimeters which means that the Transcend SSD720 can be used in ordinary computers as well as ultrabooks.
The sticker on the top of the case tells you the model’s name, part number and capacity. Overall, the Transcend SSD720 looks like many other same-class products.
There’s nothing exciting inside, either. Featuring an original design, the PCB carries a standard set of components: an LSI SF-2281 controller on one side and eight SanDisk SDZNPQBHER-016GT chips on the other side. In other words, the Transcend SSD720 has the same hardware configuration as the above-discussed Silicon Power Velox V60.
The second-generation SandForce platform employed in the Transcend SSD720 uses 24nm flash memory with Toggle Mode interface. The SanDisk chips have two 64-gigabit semiconductor dies each, so the controller can use 2-way interleaving on each of its eight channels in this 128GB SSD. From the hardware standpoint, the Transcend SSD720 is similar to the numerous SandForce-based products with synchronous memory from IMFT. Similar, but not identical.
Now let’s take a look at the official specs of the Transcend SSD720 128GB:
The specified speeds are indecently high and, considering the peculiarities of the SandForce controller, have little to do with reality. What’s interesting about them is the specified capacity. It is no typo: the storage capacity is 128 rather than 120 gigabytes.
Transcend engineers have somehow found an extra 8 gigabytes in the standard design of SSDs with SF-2281 controller. The amount of reserved space is reduced from 13 to 7%, which looks like a competitive advantage of the Transcend SSD720 over the Silicon Power Velox V60 despite the same hardware.
Transcend didn’t use some special magic for that. The capacity increase is currently permitted by the reference SandForce firmware. SSDs with SF-2281 controller used to reserve some of their flash memory for two purposes: 1) replacement memory cells and garbage collection and 2) RAISE technology. It is impossible not to lose some of the SSD’s total capacity in the first case, so there are actually no SSDs available that offer all of their flash memory to the user. As opposed to that, the RAISE technology can now be turned off and it is indeed turned off in the Transcend SSD720.
RAISE is a feature of second-generation SandForce controllers which improves reliability by adding checksums to data that help in error correction. It is similar to RAID5 but is applied to NAND flash devices rather than hard disks. The technology needs one unused NAND device to work, which explains the capacity reduction typical of SandForce-based SSDs. One semiconductor die (8 gigabytes) is used to store the checksums. The controller instead acquires the ability to correct errors and even prevent data loss in case of a failure of a whole NAND device.
RAISE is useful in two cases. First, it’s when high reliability is required, for example in corporate environments. Second, it’s when the SSD uses low-grade flash memory which can produce a lot of errors. As soon as the SandForce firmware permitted to disable RAISE, there appeared RAISE-less solutions like the Transcend SSD720. SanDisk’s 24nm flash installed in this SSD is high quality and needs no special technology for error recovery.
The Transcend SSD720 has one more unique feature. The manufacturer provides a special utility called SSD Scope for it. Its extensive functionality can only be matched by Intel’s SSD Toolbox.
SSD Scope can give you information about the drive and its S.M.A.R.T. status.
It can update the firmware, too. Besides that, it can control the OS’s TRIM command and clone data from an HDD to an SSD.
Some of the utility’s features are not implemented properly, though. For example, you can clone only to a larger partition, which isn’t a likely scenario with SSDs. SSD Scope is also supposed to support diagnostic scanning and Secure Erase, but these features didn’t work with our SSD720 120GB drive.
Transcend’s attention to the software part of this product can also be seen in regular firmware updates. The SSD720 can already be used with firmware version 5.0.4 which has been adapted for but a small number of SandForce-based products so far.
For our today’s SSD test session we put together a system on an Intel H67 based mainboard. This chipset provides support for two SATA 6 Gbit/s ports, which we use to connect the tested SSDs.
Since our today’s main heroes, Silicon Power Velox V60 and Transcend SSD720, come in 120 GB and 128 GB storage capacities, we selected SSDs of comparable size to compete against them in our test session. This allowed us to add another very interesting SSD product in our today’s test session, which we haven’t had a chance to discuss earlier – Intel SSD 330. The thing is that Intel SSD 330 240 GB drive started shipping just recently and hasn’t yet reached our lab. However, this flash-drive is very appealing for a number of reasons. On the one hand it is very much like Intel SSD 520. Namely, it is also based on the LSI SF-2281 controller, uses 25 nm synchronous memory with ONFI interface and sophisticated high-performance firmware from Intel. On the other hand, Intel SSD 330 is an inexpensive drive therefore the resource of its flash-memory has been reduced from 5000 P/E cycles to only 3000, and the warranty has been shortened to 3 years. In other words, Intel SSD 330 is a very interesting product in terms of price-to-performance ratio, but at the same time it is a little risky as far as reliability goes.
Besides Intel SSD 330, we also included other more familiar products in this test session, though this time they will stand before us in their 120/128 GB incarnations. First, these are a pair of standard SandForce based products from Corsair – Corsair Force GT and Corsair Force 3, i.e. a drive built with 25 nm synchronous ONFI-memory and another one with asynchronous memory. Second, we took a Crucial m4 drive on Marvell 88S9174 with 25 nm synchronous memory with ONFI interface. And third, we took two OCZ SSDs on Everest 2 controller: vertex 4 with synchronous ONFI-memory and Agility 4 based on asynchronous NAND-flash.
Overall our testbed was configured as follows:
We use CrystalDiskMark 3.0.1 benchmark to test the random- and sequential read and write speed. This benchmark is convenient to work with as it can measure the speed of an SSD with both incompressible random and fully compressible recurring data. This feature is important for testing SSDs based on SF-2281/2282 controller, which tries to compress the data before writing it into the flash-memory. So, there are two numbers in the diagrams below that reflect the maximum and minimum SSD speed. The real-life performance of an SSD is going to be in-between those two numbers depending on how effective the controller data compression is.
Note that the performance tests in this section refer to SSDs in their “Fresh Out-of-Box” state (FOB). No degradation could have taken place yet.
We’ve mentioned above that the capacity of 120/128 gigabytes isn’t optimal for SandForce-based SSDs. The numbers you can see in the diagram confirm our point. While 256GB SSDs with SF-2281 controller deliver average performance in comparison with their opponents, their 120/128GB cousins are much slower. The only scenario in which such SSDs can outperform Marvell- or Indilinx-based ones is when they process compressible data.
The Silicon Power Velox V60 and the Transcend SSD720 behave like typical SandForce-based products with synchronous flash. They differ somewhat from the Corsair Force GT, yet this difference is too small, even though their Toggle Mode NAND flash can ensure higher performance than ordinary synchronous flash from Intel or Micron. The best SandForce-based SSD here is Intel’s SSD 330, though. As with the SSD 520 model, Intel’s exclusive firmware does the trick.
We can also see no fundamental difference between the Transcend SSD720 and the other SandForce-based products. It looks like the turning off of RAISE technology doesn’t affect its performance much, at least when the SSD is in its out-of-box state.
Unfortunately, SSDs are not always as fast as in their “fresh” state. In most cases their performance goes down after some time and in real life we deal with completely different write speeds than what we see on the diagrams in the previous chapter of our review. The reason for this phenomenon is the following: as the SSD runs out of free pages in the flash memory, its controller has to clear memory page blocks before saving data into them, which causes substantial delays. Although, modern SSD controllers can alleviate the performance drop by erasing unused flash memory pages ahead of time, when idle. They use two techniques for that: idle-time garbage collection and TRIM.
Of course, users are more interested in the consistent performance of their SSDs over a long period of time rather than the peak speed they are going to see only during the initial short-term usage period, while the drive is still “fresh”. The SSD makers, however, declare the speed characteristics of “fresh” SSDs for marketing reasons. That’s why we decided to test the performance hit that occurs when a “fresh” SSD becomes a “steady” one.
To get a complete picture of SSD performance degradation we ran special tests based on the SNIA SSSI TWG PTS (Solid State Storage Performance Test Specification) methodology. The main idea of this approach is to measure write speed consecutively in four different cases. First we measure the “fresh” SSD speed. Then we measure the speed after the SSD has been fully filled with data twice. The third test occurs after a 30-minute break during which the controller can partially restore performance by running the idle-time garbage collection. And finally, we measure the speed after issuing a TRIM command.
We ran the tests in synthetic IOMeter 1.1.0 RC1 benchmark, where we measured random write speed when working with 4 KB data blocks aligned to flash memory pages at 32 requests queue depth. The test data were pseudo-random. The following diagram shows the history of the relative speed changes, where 100% refers to the SSD performance in “fresh-out-of-box” state.
Strange as it might seem, but SandForce engineers have found it very hard to properly implement the TRIM command, which restores the SSD performance during writes. The SF-2281 controller was released over a year and a half ago, but TRIM still cannot restore the performance of a SandForce-based drive to its original level. After filled up with data, SandForce-based SSDs slow down by 10 to 15% even in TRIM-supporting OSes. The Silicon Power Velox V60 shows the worst scenario: its performance plummets by 25% due to its firmware (version 5.0.2) having a TRIM implementation error.
So, if you’re going to use the full capacity of your SSD, you may want to choose products with newer firmware, like the Transcend SSD720, yet even this drive is a typical SandForce, although with disabled RAISE technology. You should look at products other than SF-2281 based ones for a perfect TRIM implementation.
Since the characteristics of most SSDs do change once they transition from fresh out-of-the-box state into steady state, we measure their performance once again using CrystalDiskMark 3.0.1 benchmark. The diagrams below show the obtained results. We use random data writing and measure only performance during writes, because read speed remains constant.
That’s a good illustration of how weak SandForce-based SSDs are at writing in their steady state. Take note that the two drives with 24nm Toggle Mode flash, Silicon Power Velox V60 and Transcend SSD720, are somewhat slower than the reference Corsair Force GT with 25nm ONFI flash. We saw the opposite when these drives were in their out-of-box state.
The popular PCMark 7 contains an individual disk subsystem benchmark. It is not a synthetic test, but is based on real-life applications. This benchmark reproduces typical disk usage scenarios and measures how fast they are completed in popular applications. Moreover, the disk access commands are not executed as a steady uninterrupted flow, but in a more realistic manner – with certain pauses caused by the need to process the data. The benchmark generates an overall disk subsystem performance rating as well as speed readings in MB/s in individual usage scenarios. Note that the absolute speed in these scenarios is not too high because of the above mentioned pauses between individual input/output operations. In other words, PCMark 7 shows you the speed of the disk subsystem from the application’s point of view. Numbers like that show us not only the pure performance of an SSD, but mostly how big of a performance gain a certain SSD can guarantee in real life.
We ran PCMark 7 on “steady” SSDs, which is what they are going to be in actual computer systems most of the time. Their performance in this case is affected not only by their controller or flash memory speed but also by the efficiency of their internal algorithms that fight performance degradation.
The Silicon Power Velox V60 and the Transcend SSD720 are in the leading group here, the Velox V60 beating the OCZ Vertex 4 and almost catching up with the Intel SSD 330 which takes first place thanks to its exclusive firmware.
Now let’s check out the individual tests to get a more detailed picture of what our SSDs are capable of under various types of operational load:
Where the SSDs differ in performance more, the Silicon Power Velox V60 and Transcend SSD720 are almost always faster than the Corsair Force GT but slower than the Intel SSD 330. They are also similar to each other because they use the same hardware components: SF-2281 controller with 24nm flash from SanDisk. The slightly different capacity and different firmware versions do not interfere with their performance.
Intel NASPT is another disk sub-system test that uses real-life usage scenarios. Like PCMark 7, Intel NASPT reproduces predefined disk activity traces and measures how fast they are executed. However, the default traces are designed for network attached storage devices rather than for SSDs. Therefore during our test session we replace them with the specially developed SSD Benchmarking Suite which offers more relevant usage scenarios such as compressing and decompressing files, compiling large projects, copying files and folders, loading 3D game levels, installing software, batch-processing photos, searching a digital library for data, mass-launching applications, and transcoding video.
Like PCMark 7, this benchmark gives us a true-to-life illustration of disk subsystem performance. Here the SSDs are again tested in their “steady” state.
We guess that Intel NASPT provides a more realistic picture of SSD performance and it says that the best results are delivered by the OCZ Vertex 4 and the Crucial m4. The Transcend SSD720 is the best among the SandForce-based products, leaving even the Intel SSD 330 behind. This must be due to its Toggle Mode memory which has higher interface bandwidth than typical synchronous ONFI flash. Slowed down by its firmware flaws, the Silicon Power Velox V60 isn’t that fast, yet beats the Corsair Force GT. Thus, the new implementations of the SandForce platform with 24nm SanDisk flash are generally a little better than their predecessors.
Besides the average benchmark score, we would also like to offer you the results of individual usage scenarios, which will show where Silicon Power Velox V60 and Transcend SSD720 can really shine. Note that the data-transfer rate is higher than the SATA III interface bandwidth in some subtests. That’s because INASPT is a high-level test that uses standard Windows functions to access the disk subsystem. As a result, the OS caching mechanisms also affect the results.
Although 120- and 128-gigabyte SandForce-based SSDs can’t beat any performance records, there are a few usage scenarios where the Silicon Power Velox V60 and the Transcend SSD720 are very fast thanks to their fast memory. For example, the Velox V60 wins the test of writing folders with lots of small files and the test of starting up the OS and applications. Hopefully, this SSD will get a firmware update to solve its TRIM problem and increase its performance even more. The Transcend SSD720 seems to be potentially slower due to its disabled RAISE technology.
We use AS SSD version 1.6.4237.30508 test to benchmark the speed of copying files within a single partition the size of the whole SSD. The SSDs are tested in their steady state.
SandForce controllers have never been good at doing writing and reading in parallel. That’s why SandForce-based SSDs are so slow at copying files. The Transcend SSD720 is the best among them, outperforming the Intel SSD 330, Silicon Power Velox V60, and Corsair Force GT.
We have already tested SSDs with second-generation SandForce controllers and 24nm Toggle Mode flash memory before and today’s testing of Silicon Power Velox V60 and Transcend SSD720 drives didn’t produce any surprises since they have the same hardware platform. These SSDs are somewhat faster than popular SandForce-based drives with 25nm flash and synchronous ONFI interface, but often find themselves lagging behind the Intel SSD 330 which features exclusive firmware. It means that the Velox X60 and SSD720 are typical SSDs with SF-2281 controller which have all the highs and lows of such products.
The nonstandard 128GB capacity of the Transcend SSD720 didn’t change anything. The disabled RAISE technology didn’t affect the drive’s performance much. SandForce-based SSDs with a reserved space of 7% rather than 13% do not offer any benefits other than the extra user-accessible storage. Some people may worry about potential reduction in their reliability, but we don’t share such apprehensions because today’s MLC NAND flash is high quality.
Considering the respectable age of the second-generation SandForce controller and the fact that the capacity of 120/128 GB isn’t optimal for it, we must admit that the Silicon Power Velox V60 120GB and the Transcend SSD720 128GB can hardly be counted among the most interesting solutions today. The former is especially poor due to its old firmware with TRIM implementation flaws. The Transcend SSD720 is better in this respect. It has up-to-date firmware and useful accompanying software. The price factor must also be taken into account, though. Priced affordably, these SSDs seem to be quite competitive in the mainstream market segment. They can be recommended for all-purpose desktop and mobile computers.
Some time ago we started to fill in a summary table with test results of various SSDs. It contains basic hardware information about the tested SSDs and allows to quickly determine the general position of a particular model among its competitors in terms of relative performance.