When I reviewed MyDigitalSSD's BP3, I have to say I was positively surprised. A relatively unknown manufacturer combined with a Phison controller is not the most promising mix. With SandForce you at least know what to expect but our experience with Phison based SSDs had been very limited and Crucial's v4 definitely didn't build a golden image of Phison as a controller maker, which made me very skeptic about the BP3 when I first got it. Fortunately, MyDigitalSSD proved me wrong. The BP3 turned out to be not the highest performing drive, but rather a very good bang for the buck. It was noticeably cheaper than many other mSATA offerings in the market, which made it an alluring option for value orientated mSATA buyers.

Almost immediately after our BP3 and SMART review went up, MyDigitalSSD told me that the successor to the BP3 was already just around the corner: the BP4. From a hardware standpoint it's very similar to the BP3 and MyDigitalSSD has simply switched from 24nm to 19nm NAND, which is logical given the cost savings of smaller lithography NAND. The heart of BP4 is still the same Phison PS3108 controller but there's been some pretty significant changes in the firmware and MyDigitalSSD has also opted to increase the default over-provisioning from 7% to 12%.

With the BP4, MyDigitalSSD is also trying to be more aggressive in the 2.5" field. There was a 2.5" BP3 but MyDigitalSSD never made much noise about it. I raised my concerns about the retail mSATA market in the BP3 and SMART review, although MyDigitalSSD didn't exactly agree with me and they were very satisfied with their mSATA sales (which is absolutely not a bad thing). Either way, MyDigitalSSD is looking to get out of the niche market and the BP4 is aimed directly at the mainstream market. I have to say I'm quite pleased with this approach because most of the so called budget SSDs are fairly bad, so more competition in that market segment is always welcome. The Samsung SSD 840 has really been the only budget drive without any serious drawbacks.

	BP4	BP4 mSATA
Controller	Phison PS3108
NAND	19nm Toshiba Toggle-Mode MLC
Capacities (GB)	60, 120, 240, 480, 960	30, 60, 120, 240
Sequential Read	560MB/s	560MB/s
Sequential Write	530MB/s	520MB/s
4KB Random Read	50K IOPS	4.4K IOPS
4KB Random Write	80K IOPS	16.1K IOPS

MyDigitalSSD sent us both the 2.5" and mSATA versions of the BP4. As you can see in the table above, there is a fairly big difference in performance between the 2.5" and mSATA version. I'm not sure why the difference is that dramatic but it most likely has to do with the controller design. If your controller is designed for high parallelism, you will only get the best performance when all the channels are used. The decrease in performance with fewer channels depends on how heavily you rely on parallelism but it seems like Phison's PS3108 really needs all eight channels to be populated to achieve high performance. Especially as we move to smaller process nodes, parallelism becomes more and more important because read and program latencies increase at every process node.

Quite surprisingly, there's also a 960GB version of the 2.5" BP4 and unlike the other terabyte-class 2.5" solutions we have seen so far, this isn't just two 480/512GB SSDs in RAID 0 but a "real" 960GB SSD driven by a single controller. I actually have a sample in the house already so a review will be up soon.

 

The 2.5" BP4 has a total of sixteen NAND packages (16GB per die), eight on each side of the PCB. There's also Powerchip Technology's 512MB DRAM chip as a cache, 120GB and smaller capacities will have 256MB instead. 

 

The casing is very bare from inside. MyDigitalSSD isn't using any thermal pads to dissipate the heat, although they are rarely necessary in consumer SSDs since the controllers are relatively low power.

As the mSATA form factor only has space for four NAND packages, the highest capacity you can achieve with an 8GB die is still 256GB (which will increase to 512GB once 16GB die becomes available). Our review sample is a very early sample that uses 24nm Toshiba NAND (indicated by the 11th character which is an H, 19nm NAND has and I or J), but all retail units should ship with 19nm NAND. 

MyDigitalSSD also sent me their new mSATA to USB 3.0 adapter. For $30, it's definitely worth it if you happen to have a spare mSATA SSD. ASMedia's ASM1053 is acting as a bridge in the adapter and it features support for SATA 6Gbps, although I was only able to reach speeds of around 150MB/s.

 

Test System

CPU	Intel Core i5-2500K running at 3.3GHz (Turbo and EIST enabled)
Motherboard	AsRock Z68 Pro3
Chipset	Intel Z68
Chipset Drivers	Intel 9.1.1.1015 + Intel RST 10.2
Memory	G.Skill RipjawsX DDR3-1600 2 x 4GB (9-9-9-24)
Video Card	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.

Random read performance has not been dramatically improved. In fact, the mSATA version actually takes a small hit (~12%).

Random write speed, on the other hand, is substantially better. We are still nowhere near SandForce or OCZ Vector numbers but at least the performance is no longer sub-par. The performance also finally scales up as the queue depth increases (at least for the 2.5" model). The scaling could be more aggressive at higher QDs but I'll rather take good low QD performance as that will have a bigger impact in real world performance where most IOs are between QDs of 1 and 5.

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.

Sequential performance is also very good. Surprisingly the BP4 mSATA is again slower in write performance than the BP3 and compared to the 2.5" BP4, there is quite a big difference.

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.

Performance Consistency

In our Intel SSD DC S3700 review Anand introduced a new method of characterizing performance: looking at the latency of individual operations over time. The S3700 promised a level of performance consistency that was unmatched in the industry, and as a result needed some additional testing to show that. The reason we don't have consistent IO latency with SSDs is because inevitably all controllers have to do some amount of defragmentation or garbage collection in order to continue operating at high speeds. When and how an SSD decides to run its defrag and cleanup routines directly impacts the user experience. Frequent (borderline aggressive) cleanup generally results in more stable performance, while delaying that can result in higher peak performance at the expense of much lower worst case performance. The graphs below tell us a lot about the architecture of these SSDs and how they handle internal defragmentation.

To generate the data below I took a freshly secure erased SSD and filled it with sequential data. This ensures that all user accessible LBAs have data associated with them. Next I kicked off a 4KB random write workload across all LBAs at a queue depth of 32 using incompressible data. I ran the test for just over half an hour, no where near what we run our steady state tests for but enough to give me a good look at drive behavior once all spare area filled up.

I recorded instantaneous IOPS every second for the duration of the test. I then plotted IOPS vs. time and generated the scatter plots below. Each set of graphs features the same scale. The first two sets use a log scale for easy comparison, while the last set of graphs uses a linear scale that tops out at 40K IOPS for better visualization of differences between drives.

The high level testing methodology remains unchanged from our S3700 review. Unlike in previous reviews however, I did vary the percentage of the drive that I filled/tested depending on the amount of spare area I was trying to simulate. The buttons are labeled with the advertised user capacity had the SSD vendor decided to use that specific amount of spare area. If you want to replicate this on your own all you need to do is create a partition smaller than the total capacity of the drive and leave the remaining space unused to simulate a larger amount of spare area. The partitioning step isn't absolutely necessary in every case but it's an easy way to make sure you never exceed your allocated spare area. It's a good idea to do this from the start (e.g. secure erase, partition, then install Windows), but if you are working backwards you can always create the spare area partition, format it to TRIM it, then delete the partition. Finally, this method of creating spare area works on the drives we've tested here but not all controllers may behave the same way.

The first set of graphs shows the performance data over the entire 2000 second test period. In these charts you'll notice an early period of very high performance followed by a sharp dropoff. What you're seeing in that case is the drive allocating new blocks from its spare area, then eventually using up all free blocks and having to perform a read-modify-write for all subsequent writes (write amplification goes up, performance goes down).

The second set of graphs zooms in to the beginning of steady state operation for the drive (t=1400s). The third set also looks at the beginning of steady state operation but on a linear performance scale. Click the buttons below each graph to switch source data.

Impact of Spare Area
	MyDigitalSSD BP4 240GB	MyDigitalSSD BP4 mSATA 240GB	MyDigitalSSD BP3 mSATA 256GB	Plextor M5M 256GB	Intel SSD 525 240GB	Samsung SSD 840 Pro 256GB
Default
25% Spare Area

The BP4 takes a big step forward in terms of IO consistency. There's more OP by default, that's for sure, but even if you look at the BP3 with 25% OP results, the improvement is still huge. Instead of playing in the 100-1,000 IOPS field, the BP4 stays between 1,000 and 10,000 IOPS throughout the test. What you can't actually see in the logarithmic graphs are the data points where BP3's IOPS is zero. There are plenty of those and sometimes the BP3 stopped for 10-20 consecutive seconds. I'll be honest here and say that that's very bad because it basically means the SSD isn't completing any IO requests from the host for that period, which means your whole system will be waiting for the drive. Even with 25% OP, there were moments when the BP3 stopped completely, though fortunately the durations were only 1-5 seconds instead of 20.

Another interesting point is that the 2.5" and mSATA BP4s behave quite differently. The 2.5" obviously has higher average IOPS but the mSATA is more consistent. If you look at the default OP mSATA graph, it's almost like a straight line once steady state is reached, and performance actually improves over time. The 2.5" BP4 is more inconsistent and its garbage collection seems to behave differently in general. Basically, there are periods of lower IOPS, where the drive is most likely doing garbage collection, and then you got peaks where performance increases temporarily until the drive has to slow down to do garbage collection again.

Neither of the approaches is worse than the other and I should probably note that the mSATA BP4 has a newer firmware (v4.6 instead of v4.3 found in the 2.5"). MyDigitalSSD told me that the firmware 4.6 only addresses some compatibility issues and 2.5" BP4 should currently ship with 4.6 as well, so I doubt that's the reason for the difference. The 2.5" BP4 has higher overall throughput so the drive will also be in more fragmented state, so it's possible that the higher fragmentation simply puts the drive into a state where it has to slow down every once in a while to keep the performance. Anyway, the drop is rather small (I've seen some other SSDs dropping much more) so I wouldn't worry about it.

Impact of Spare Area
	MyDigitalSSD BP4 240GB	MyDigitalSSD BP4 mSATA 240GB	MyDigitalSSD BP3 mSATA 256GB	Plextor M5M 256GB	Intel SSD 525 240GB	Samsung SSD 840 Pro 256GB
Default
25% Spare Area

While the 2.5" BP4 has much higher out-of-the-box IOPS, the steady state performance is very close to the mSATA version. The 2.5" does have peaks in the 10,000 IOPS range but the baseline performance for both SSDs is around 2,500-5,000 IOPS. The non-logarithmic graph also really shows how the BP3 suffers—the performance is literally zero IOPS throughout the steady-state period.

Impact of Spare Area
	MyDigitalSSD BP4 240GB	MyDigitalSSD BP4 mSATA 240GB	MyDigitalSSD BP3 mSATA 256GB	Plextor M5M 256GB	Intel SSD 525 240GB	Samsung SSD 840 Pro 256GB
Default
25% Spare Area

 

Performance vs. Transfer Size

ATTO does a good job of showing us how sequential performance varies with transfer size. Most controllers optimize for commonly seen transfer sizes and neglect the rest. The optimization around 4KB, 8KB and 128KB transfers makes sense given that's what most workloads are bound by, but it's always important to understand how a drive performs across the entire gamut.

Based on IOmeter scores, the BP4 has pretty impressive sequential read performance at high IO sizes but I'm actually pretty surprised that the BP4 performs great across all transfer sizes. Write performance isn't as good but it's still average. As always, I tried to keep the graphs readable, you can find scores for all SSDs in Bench.

Click for full size

AnandTech Storage Bench 2011

Two years ago 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
4KB	28%
16KB	10%
32KB	10%
64KB	4%

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 heavy workload test:

The BP4 is definitely not a record breaker and it's about 10-15% slower than most SSDs. The mSATA version is in fact slower than the mSATA BP3, although that was somewhat expected given the synthetic tests (though that's a tradeoff I'm willing to make for much, much better IO consistency). If we look at the breakdown, the BP4 is the slowest in reads but not by a large margin, so it's really the write performance that needs improvement (especially when writing to random LBAs).

AnandTech Storage Bench 2011—Light Workload

Our 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
4KB	27%
16KB	8%
32KB	6%
64KB	5%

The rank of BP4 doesn't change much when we move to a lighter workload.

Performance Over Time & TRIM

Our new performance consistency tests give a pretty good indication of performance over time but HD Tach is still a good method for checking TRIM functionality. As usual, I took a secure erased 2.5" BP4 and ran it through HD Tach to get the baseline performance.

Next I filled the drive with sequential data and tortured with 4KB random writes (QD=32, 100% LBA space) for 60 minutes and reran HD Tach.

Performance drops to as low as 20MB/s for the earliest LBAs, which is similar to what the steady state performance looked like in our IO consistency test. Over time, performance gets better as the drive defragments itself and after about 150GB of sequential writes, peak performance gets over 300MB/s. Again, the behavior we're seeing here is related to what we saw in IO consistency tests: there are short peaks and then the drive goes back to do garbage collection.

To test TRIM, I secure erased the drive, filled it and tortured again for 60 minutes to make sure that it's fully fragmented. Then I TRIM'ed all the LBAs:

And TRIM works. Write speed is actually higher than after a secure erase but if we look at IOmeter scores, ~350MB/s is what the BP4 gets. Sometimes drives perform weirdly right after a secure erase. Especially if you're erasing a highly fragmented drive, it may take some minutes for the drive to complete the request, even though it shows up as secure erased.

Power Consumption

Given the increasing usage of SSDs in laptops, power consumption is playing an even bigger role than before. A power efficient SSD can potentially add an hour or more of extra battery life, whereas an inefficient SSD can consume more power than a traditional hard drive. MyDigitalSSD touted low power consumption being one of the main advantages of the BP4 and it definitely lives up to the claim. Idle power draw is similar to what we saw with Samsung's SSD 840/840 Pro while load power consumption is average. Especially for light workloads, idle power consumption is what you should look at because the drive will be idling most of the time.

The mSATA BP4 has higher power consumption but that's most likely due to the mSATA to SATA adapter. mSATA drives use 3.3V whereas the standard SATA provides a 5V rail, so the voltage has to be lowered with a voltage regulator, which uses some power as well. Anand did some measurements between 3.3V and 5V power consumption a while ago, but unfortunately our adapters are not the same so I'm not sure how much power my adapter is drawing. However, the 2.5" BP4 should serve as a good reference point since the mSATA version is essentially the same.

Final Words

There's one thing that tells more about the BP4 than anything else: price. There's no SSD that can really challenge it in price if we look at all available capacities and I even tried to gather budget SSDs in the table below. Especially the 240GB BP4 at $160 is a brilliant deal as 240GB/256GB SSDs are commonly in the $200 range. Sure there are sales every now and then and I've seen the 250GB Samsung 840 hitting $150 but keep in mind that MyDigitalSSD's prices are not temporary—they are retail prices and are unlikely to change dramatically (at least not by going up).

Price Comparison (4/3/2013)
Capacity	120/128GB	240/256GB	480/512GB
MyDigitalSSD BP4	$90	$160	$350
MyDigitalSSD BP4 mSATA	$110	$180	N/A
MyDigitalSSD BP3 mSATA	$100	$180	N/A
Samsung SSD 840	$100	$210	$350
Corsair Neutron	$120	$210	N/A
OCZ Agility 4	$115	$200	N/A
Mushkin Chronos	$110	$180	$360
Intel SSD 525	$170	$290	N/A
Crucial M4 mSATA	$130	$210	N/A
Mushkin Atlas	$115	$190	N/A

Of course, you always have to trade off performance for price. The BP4 is not the fastest SSD we have tested and most of the other SATA 6Gbps SSDs are faster, but the BP4 is not terribly slow. Any SSD (well, almost) is still much faster than a traditional hard drive so simply having an SSD is far more important than the brand or performance of the SSD.

We have seen some pretty bad budget SSDs over the years, such as Crucial v4 and OCZ Agility 4, but the BP4 is miles ahead. Typically budget SSDs have one big Achilles' Heel (oftentimes write performance due to the use of lower quality NAND) but the BP4 has none. Random IO performance is average, IO consistency is good and power consumption is great. I literally can't find any weak points in the BP4. There are aspects that could be better, such as random read/write speeds, but after all we are dealing with a budget SSD and I don't think it's even reasonable to expect high-end SSD performance.

The only question is long-term reliability. When a relatively unknown player steps into the market, it always takes at least a generation or two before any sort of reliability can be determined. MyDigitalSSD is claiming that their reliability is on par with Intel but I would take manufacturers' words with a grain of salt. That said, I haven't seen enough MyDigitalSSD's drives around to draw any scientific conclusions of their reliability but on the other hand, I haven't had any issues with mine nor have I heard of anyone else having issues. Only time will tell how the reliability plays out but then again, we also need a sufficient sample size as reliability can't be determined unless people actually buy and use the product. While the unknown reliability is definitely a con, I don't think it should be taken too seriously. Any drive can fail so you should always have a backup regardless of what your primary drive is.

All in all, I have no reason not to recommend the BP4. For once the price/performance ratio is reasonable as most of the time budget SSDs fail due to the fact that the price isn't cheap enough to compensate for the lack of performance. I would still choose Samsung SSD 840 over the BP4 if the price is the same, mainly because the SSD 840 is slightly faster and its reliability is more proven, but I wouldn't consider the SSD 840 to be worth much more (maybe ~$10-20).

All MyDigitalSSD needs to do now is to get their name out in the public as I'm sure not many have heard of them before. Expanding their sales channel to more vendors would be a good start because in the end you will always be limited in visibility if you only rely on Amazon and your own online store (although I do see the the enticement of keeping the distribution channel as simple as possible and it definitely helps to keep the prices low). I'm sure MyDigitalSSD has considered many options and made their decisions based on what's best for the company, but there's no doubt that the BP4 could be a great vehicle for taking a bigger slice of the market.