Original Link: http://www.anandtech.com/show/6527/mydigitalssd-smart-bp3-msata-ssd-review
MyDigitalSSD SMART & BP3 mSATA SSD Reviewby Kristian Vättö on January 22, 2013 1:24 PM EST
The consumer SSD market is quite similar to the DRAM market. There are only a handful of NAND manufacturers (most of which make DRAM as well) but there are dozens, if not hundreds of SSD OEMs. Compared to DRAM there are obviously more components involved because on top of the NAND you'll also need a controller and possibly DRAM as well. Thanks to Marvell, Phison and especially SandForce you don't need a huge team of engineers to make an SSD because you can buy and license everything from third parties. Even manufacturing can be outsourced so basically what you're left with is distribution and marketing. That, of course, is if you choose the easiest route, which isn't necessarily the ideal option because there are already plenty of other companies using the exact same strategy.
MyDigitalSSD is one of the not-so-well-known SSD companies. They don't have a presence on NewEgg or many of the other major online stores, though you can find some of their products at Amazon. Since MyDigitalSSD doesn't have the resources it takes to build their own controller or firmware, they are left with using commercial controllers, SandForce and Phison in this case. Unlike many other SSD OEMs, MyDigitalSSD's aim is to provide something for everyone. Typically SSD OEMs, regardless of how big they are, only offer a few products that are almost without exception 2.5" SATA drives. MyDigitalSSD's approach is totally different as they offer SSDs ranging from standard 2.5" SATA drives to PATA SSDs and half-slim SATA SSDs. We don't often see such form factors used but there are laptops that rely on some of these uncommon SSD solutions. Of course if you're buying in volumes big enough (like Apple), then anyone will build you whatever you like; that makes finding upgrade parts difficult, so MyDigitalSSD is specifically targeting that market.
MyDigitalSSD sent us their 256GB SATA 6Gbps mSATA SSDs in for reviewing. Complete specifications are in the table below:
|Capacities (GB)||64, 128, 256||32, 64, 128, 256|
|NAND||25nm synchronous MLC (IMFT?)||24nm Toshiba Toggle-Mode MLC|
|Controller||SandForce SF-2281||Phison PS3108-S8|
|4KB Random Read||35K IOPS||30K IOPS|
|4KB Random Write||86K IOPS||45K IOPS|
MyDigitalSSD's SMART SSD is a standard SF-2281 based mSATA SSD and there are other OEMs such as Mushkin and ADATA offering similar products. What is more interesting (at least from a novelty standpoint) is the BP3 ("Bullet Proof 3"). It uses a new SATA 6Gbps controller from Phison, a company that's more known for their USB flash stick controllers. Our first encounter with Phison was with Crucial's v4 SSD, which wasn't very pleasant as the v4 was one of the slowest SSDs we have reviewed in years. As far as the specs go, the PS3108 seems to provide a much needed improvement to the random IO performance segment; we'll see how the PS3108 holds out in real world in just a second.
There aren't all that many commercially available mSATA SSDs because most are sold directly to OEMs, so most SSD manufacturers have chosen not to have a retail mSATA SSD lineup. MyDigitalSSD doesn't have presence at NewEgg or other major online resellers, but they do have their own store called MyDigitalDiscount which is also at Amazon. I took MyDigitalSSD prices from MyDigitalDiscount whereas the rest are from NewEgg:
|Price Comparison (1/21/2013)|
|Crucial M4 mSATA||$70||$115||$185|
|ADATA XPG SX300||$80||$125||$260|
In terms of pricing, the BP3 is very appealing. It's easily the cheapest mSATA SSD that I could find and by a fairly large margin. The SMART, on the other hand, is one of the most expensive mSATA SSDs so MyDigitalSSD is clearly trying to position the BP3 at the low-end while offering the SMART for the high-end.
The heart of the SMART is SandForce's SF-2281 controller, which is used by dozens of manufacturers. SandForce doesn't use dedicated DRAM cache, leaving the controller and NAND as the only components on the PCB. The drive came with firmware 5.0.2a but there's a newer 5.0.4 available at MyDigitalSSD's website. I ran a few tests with the new firmware but didn't see any noticeable difference in performance so the scores you'll see are with the 5.0.2a firmware.
On the flip side we find another two NAND packages. These are all octo-die packages to yield the capacity of 256GB, which is currently the maximum for mSATA form factor (though Mushkin has announced a stacked mSATA design with up to 480GB). Once 128Gb dies start to make their way into the market sometime in 2013, we should see mSATA SSDs of up to 512GB without special designs.
The actual NAND in the SMART is a bit of a mystery. Googling the part number yields no results but from what I was able to find, it seems that the NAND is running in synchronous mode and manufactured using IMFT's 25nm process but packaged by ADATA (for example their SX300 uses the same NAND). The idea is that you buy whole wafers of NAND and then use your own binning methodology to pick the best dies. When you buy pre-packaged NAND, it's unlikely that you'll be getting the highest quality NAND because usually the manufacturer reserves that NAND for their own products. If you buy NAND in whole wafers, you will also get a small share of the highest quality chips, although you are also left with a big chunk of NAND that is not suitable for SSDs (but it may be fine for memory sticks and cards). This is a fairly common practice in the SSD industry and for example OCZ and Kingston do it as well.
It's unclear to me why MyDigitalSSD chose ADATA but the SMART is actually recognized as an ADATA SX300 by the system, so the SMART may be a rebrand of the SX300 or at least MyDigitalSSD has worked very closely with ADATA.
Not much is known about Phison's PS3108 controller. Phison says the controller supports both SLC and MLC NAND, along with 1Xnm class NAND. The controller is fabricated using 55nm CMOS process, but the actual manufacturer is unknown. The controller is coupled with 256MB of DDR3-1333 DRAM from Nanya. The drive came with firmware version 3.2, which is currently the latest.
In the NAND department we have four of Toshiba's 24nm Togge-Mode MLC NAND packages, each consisting of eight 8GB dies yielding a total capacity of 64GB per package.
The motherboard in my testbed does not have an mSATA slot, so MyDigitalSSD also sent us an mSATA to 2.5" SATA adapter.
|CPU||Intel Core i5-2500K running at 3.3GHz (Turbo and EIST enabled)|
|Motherboard||AsRock Z68 Pro3|
|Chipset Drivers||Intel 126.96.36.1995 + Intel RST 10.2|
|Memory||G.Skill RipjawsX DDR3-1600 2 x 4GB (9-9-9-24)|
XFX AMD Radeon HD 6850 XXX
(800MHz core clock; 4.2GHz GDDR5 effective)
|Video Drivers||AMD Catalyst 10.1|
|Desktop Resolution||1920 x 1080|
|OS||Windows 7 x64|
Random Read/Write Speed
The four corners of SSD performance are as follows: random read, random write, sequential read and sequential write speed. Random accesses are generally small in size, while sequential accesses tend to be larger and thus we have the four Iometer tests we use in all of our reviews.
Our first test writes 4KB in a completely random pattern over an 8GB space of the drive to simulate the sort of random access that you'd see on an OS drive (even this is more stressful than a normal desktop user would see). I perform three concurrent IOs and run the test for 3 minutes. The results reported are in average MB/s over the entire time. We use both standard pseudo randomly generated data for each write as well as fully random data to show you both the maximum and minimum performance offered by SandForce based drives in these tests. The average performance of SF drives will likely be somewhere in between the two values for each drive you see in the graphs. For an understanding of why this matters, read our original SandForce article.
Compared to the PS3105 based Crucial v4, the BP3 offers a significant improvement in random read performance. It's still nowhere close to the fastest SATA 6Gbps drives, but at least the performance is no longer horrible. As for the SMART, its random read performance is similar to other SF-2281 based SSDs, as expected.
Random write speed is still the Achilles' Heel of Phison, though. The PS3108 is able to offer twice the performance of PS3105, but 36.4MB/s is still far slower than what most controllers are able to offer today. What's interesting is the fact that random write performance does not scale up at all when queue depth is increased—usually the throughput increases as the queue depth goes up as you can see in the graphs above.
Sequential Read/Write Speed
To measure sequential performance I ran a 1 minute long 128KB sequential test over the entire span of the drive at a queue depth of 1. The results reported are in average MB/s over the entire test length.
Similar to the Crucial v4, the BP3 has strong sequential performance. SMART's performance characteristics are once again typical SandForce, no surprises here.
AS-SSD Incompressible Sequential Performance
The AS-SSD sequential benchmark uses incompressible data for all of its transfers. The result is a pretty big reduction in sequential write speed on SandForce based controllers, while other drives continue to work at roughly the same speed as with compressible data.
Performance vs. Transfer Size
ATTO is a handy tool for quickly measuring performance at various transfer sizes. While IOs in real world usually happen at specific sizes (mainly 4KB), it important to get a big picture of performance. The BP3 is able to keep up with read performance, but write speed is once again lacking. The SMART is unsurprisingly on par with Intel SSD 335, which is based on the same SF-2281 controller.
Click for full size
AnandTech Storage Bench 2011
Last year we introduced our AnandTech Storage Bench, a suite of benchmarks that took traces of real OS/application usage and played them back in a repeatable manner. Anand assembled the traces out of frustration with the majority of what we have today in terms of SSD benchmarks.
Although the AnandTech Storage Bench tests did a good job of characterizing SSD performance, they weren't stressful enough. All of the tests performed less than 10GB of reads/writes and typically involved only 4GB of writes specifically. That's not even enough exceed the spare area on most SSDs. Most canned SSD benchmarks don't even come close to writing a single gigabyte of data, but that doesn't mean that simply writing 4GB is acceptable.
Originally we kept the benchmarks short enough that they wouldn't be a burden to run (~30 minutes) but long enough that they were representative of what a power user might do with their system. Later, however, we created what we refer to as the Mother of All SSD Benchmarks (MOASB). Rather than only writing 4GB of data to the drive, this benchmark writes 106.32GB. This represents the load you'd put on a drive after nearly two weeks of constant usage. And it takes a long time to run.
1) The MOASB, officially called AnandTech Storage Bench 2011—Heavy Workload, mainly focuses on the times when your I/O activity is the highest. There is a lot of downloading and application installing that happens during the course of this test. Our thinking was that it's during application installs, file copies, downloading, and multitasking with all of this that you can really notice performance differences between drives.
2) We tried to cover as many bases as possible with the software incorporated into this test. There's a lot of photo editing in Photoshop, HTML editing in Dreamweaver, web browsing, game playing/level loading (Starcraft II and WoW are both a part of the test), as well as general use stuff (application installing, virus scanning). We included a large amount of email downloading, document creation, and editing as well. To top it all off we even use Visual Studio 2008 to build Chromium during the test.
The test has 2,168,893 read operations and 1,783,447 write operations. The IO breakdown is as follows:
|AnandTech Storage Bench 2011—Heavy Workload IO Breakdown|
|IO Size||% of Total|
Only 42% of all operations are sequential; the rest ranges from pseudo to fully random (with most falling in the pseudo-random category). Average queue depth is 4.625 IOs, with 59% of operations taking place in an IO queue of 1.
Many of you have asked for a better way to really characterize performance. Simply looking at IOPS doesn't really say much. As a result we're going to be presenting Storage Bench 2011 data in a slightly different way. We'll have performance represented as Average MB/s, with higher numbers being better. At the same time we'll be reporting how long the SSD was busy while running this test. These disk busy graphs will show you exactly how much time was shaved off by using a faster drive vs. a slower one during the course of this test. Finally, we will also break out performance into reads, writes, and combined. The reason we do this is to help balance out the fact that this test is unusually write intensive, which can often hide the benefits of a drive with good read performance.
There's also a new light workload for 2011. This is a far more reasonable, typical every day use case benchmark. It has lots of web browsing, photo editing (but with a greater focus on photo consumption), video playback, as well as some application installs and gaming. This test isn't nearly as write intensive as the MOASB but it's still multiple times more write intensive than what we were running last year.
We don't believe that these two benchmarks alone are enough to characterize the performance of a drive, but hopefully along with the rest of our tests they will help provide a better idea. The testbed for Storage Bench 2011 has changed as well. We're now using a Sandy Bridge platform with full 6Gbps support for these tests.
AnandTech Storage Bench 2011—Heavy Workload
We'll start out by looking at average data rate throughout our new heavy workload test:
Despite the limitations of the PS3108 controller, the BP3 performs decently. The gap between it and the SandForce based SMART is only 12.6MB/s in our Heavy suite. Neither of the drives is able to keep up with standard 2.5" SSDs, but it's worthwhile to note that some of this is due to the limited NAND bandwidth (4 packages versus 8-16 in 2.5" SSDs).
The next three charts just represent the same data, but in a different manner. Instead of looking at average data rate, we're looking at how long the disk was busy for during this entire test. Note that disk busy time excludes any and all idles, this is just how long the SSD was busy doing something:
AnandTech Storage Bench 2011—Light Workload
Our new light workload actually has more write operations than read operations. The split is as follows: 372,630 reads and 459,709 writes. The relatively close read/write ratio does better mimic a typical light workload (although even lighter workloads would be far more read centric).
The I/O breakdown is similar to the heavy workload at small IOs, however you'll notice that there are far fewer large IO transfers:
|AnandTech Storage Bench 2011—Light Workload IO Breakdown|
|IO Size||% of Total|
Our Light workload test presents scores similar to the Heavy test: The SMART is faster but not by much.
Performance Over Time & TRIM
As the SMART is a SandForce drive with the stock firmware, I decided not to test its TRIM performance because we have tested that so many times, the latest being our article on SandForce TRIM issue. Hence this part will only cover the BP3 but since it's a new, untested controller, this data should be rather interesting. Let's start with HD Tach ran on a secure erased drive:
For our torture test, I filled the drive with sequential data and hammered it with 4KB random writes (100% LBA space, QD=32) for 40 minutes:
Performance is actually fairly good. The worst drop is around 70MB/s right at the beginning but after about half of the LBAs have been written with sequential data, the average data rate is around 200MB/s. Part of the cause for this is the fact that the BP3 has slow random write performance, which means it will write less data to the drive than what faster drives would, resulting in a less fragmented drive.
Next I secure erased the drive, filled it and tortured it again and TRIM'ed after torture:
TRIM performance is rather dubious. Performance is definitely better than without TRIM but it's still far from clean state performance. I have a feeling that the controller itself is not very powerful so it may simply not be able to do garbage collection on all the blocks without some idle time between torture and HD Tach run (I immediately TRIM'ed the drive after torture and began the HD Tach run). It would explain why the performance is over 300MB/s for a large part of the drive. The controller may have cleaned some of the blocks and the first writes go there but in the end it faces a situation where it has to read-modify-write because all blocks are full, and write performance drop to figures similar to after torture test.
Since mSATA SSDs will almost without a doubt be used in a notebook, power consumption plays an even bigger role than it usually does. The BP3 has an advantage in idle power, although it can't achieve numbers similar to Crucial v4. SandForce SSDs have always consumed quite a few electrons while idling, and the SMART is no exception.
Load power consumption for both SSDs is rather high, although nothing out of the ordinary. The mSATA to SATA 2.5" adapter might add a little to the power consumption but it shouldn't be significant.
Phison's PS3108 and the BP3 are a pleasant surprise. After reviewing the Crucial v4, I didn't have high hopes with the BP3 given the miserable performance of Phison SATA 3Gbps controller, but I was proven wrong. That's not to say that the PS3108 is anywhere close to challenging today's high-end controllers such as Samsung MDX and OCZ Barefoot 3, but at least the performance is no longer from 2009. I also doubt Phison is aiming the PS3108 for high-end market as in general their controllers have been found in low-cost SSDs, and shifting the market target would require big investments and possibly more R&D time as well. The low-end SSD market is also interesting in the sense that there is less competition than in the high-end market: the main options are Samsung's SSD 840, a bunch of SSDs from OCZ, and some SF-2281 based SSDs that compete at the lowest prices.
Pricing is definitely the biggest advantage of the BP3. It's considerably cheaper than other mSATA SSDs and is in fact cheap in terms of 2.5" SSD prices as well. However, Phison's (and MyDigitalSSD's) long term reliability is a big question mark as Crucial v4 is really the first Phison based SSD from a big SSD manufacturer. If you're willing to be a guinea pig, the BP3 is an affordable mSATA SSD choice and due to its price, I would recommend it over the SMART—and either way, it's still significantly faster than any hard drive you'd have in a laptop.
There is one general thing about the SSD market that I've been wondering for quite a while, though: is there really a market for mSATA SSDs? I see absolutely no reason why a desktop user would pick an mSATA SSD over a regular 2.5" SSD because 2.5" SSDs are usually cheaper and also faster (mSATA form factor is limited to four channels while most controller have eight). Only very few motherboards come with an mSATA slot anyway, so that diminishes the market even more.
The notebook market is fairly limited too as most laptops are still using standard 2.5" hard drives. The manufacturer may offer SSDs in some models or as built to order but those are still 2.5", an empty mSATA slot is a rarity due to space constraints. Popular SSD-only notebooks such as ASUS' Zenbook series and Apple's MacBook Air use custom SSDs, so mSATA is of no use with those either. That basically leaves us with a very limited amount of notebooks that even have mSATA capability to begin with. Out of those systems, most will likely already have an mSATA SSD installed when the system is purchased, so really the market for retail mSATA SSDs is consumers who have bought a notebook with an mSATA SSD (probably a small caching-only SSD) and want to upgrade the SSD and make it the OS/Apps drive. That's not a very big market if you ask me, at least right now.
I may be missing some scenarios where mSATA SSDs are used but I think it's safe to say that mSATA has not really taken off and the market for retail mSATA SSDs is close to non-existent. MyDigitalSSD does have the potential to grab a large share of that market, particularly for cost-conscious users where the BP3 might be attractive, but given the early nature of the controller I'm not sure most people would be comfortable buying the BP3. Looking at the SMART, we generally know what to expect from SF-2281 SSDs, which is good, but the pricing can't compete with other SF-2281 based mSATA SSDs. I would pick Crucial's M4 mSATA SSD over the SMART since it's considerably cheaper, but the BP3 is definitely worth a consideration if you're looking for an affordable mSATA drive and are willing to settle for an unproven drive.