Storage Bench v1.0 (Background Info)
Our Storage Bench incorporates all of the I/O from a trace recorded over two weeks. The process of replaying this sequence to capture performance gives us a bunch of numbers that aren't really intuitive at first glance. Most idle time gets expunged, leaving only the time that each benchmarked drive is actually busy working on host commands. So, by taking the ratio of that busy time and the amount of data exchanged during the trace, we arrive at an average data rate (in MB/s) metric we can use to compare drives.
It's not quite a perfect system. The original trace captures the TRIM command in transit, but since the trace is played on a drive without a file system, TRIM wouldn't work even if it were sent during the trace replay (which, sadly, it isn't). Still, trace testing is a great way to capture periods of actual storage activity, a great companion to synthetic testing like Iometer.
Incompressible Data and Storage Bench v1.0
Also worth noting is the fact that our trace testing pushes incompressible data through the system's buffers to the drive getting benchmarked. So, when the trace replay plays back write activity, it's writing largely incompressible data. If we run our storage bench on a SandForce-based SSD, we can monitor the SMART attributes for a bit more insight.
| Mushkin Chronos Deluxe 120 GB SMART Attributes | RAW Value Increase |
|---|---|
| #242 Host Reads (in GB) | 84 GB |
| #241 Host Writes (in GB) | 142 GB |
| #233 Compressed NAND Writes (in GB) | 149 GB |
Host reads are greatly outstripped by host writes to be sure. That's all baked into the trace. But with SandForce's inline deduplication/compression, you'd expect that the amount of information written to flash would be less than the host writes (unless the data is mostly incompressible, of course). For every 1 GB the host asked to be written, Mushkin's drive is forced to write 1.05 GB.
If our trace replay was just writing easy-to-compress zeros out of the buffer, we'd see writes to NAND as a fraction of host writes. This puts the tested drives on a more equal footing, regardless of the controller's ability to compress data on the fly.
Average Data Rate
The Storage Bench trace generates more than 140 GB worth of writes during testing. This tends to penalize drives smaller than 180 GB and favor those with more than 256 GB of capacity. So, it's not wise to take a trace from a 240 GB SSD and wrap it around something, say, 40 GB-large. Using a small trace on a larger SSD is no problem, but going the other way results in different trace timing, affecting the results.

If you've followed us so far, you won't be surprised to see Crucial's M550s and the 1024 GB SP920 tripping over one another. Nor should the 128 and 256 GB model's outcome require much explanation. But the 512 GB SP920? That doesn't end up where we might have guessed.
First, we tried running our trace a second time. Then we tried other systems, all of which spat back the same result. Really though, losing 15 MB/s doesn't mean much at the end of the day. But we do come away knowing that the 512 GB SP920 isn't exactly the same as the 1024 GB model. Nor is it identical to the M550s. Something is slightly different on what should be an identical drive. Perhaps a look at mean service times on the next page will shed some light...
- Adata's SP920: Quite Literally, A Familiar Face
- A Primer: The Art Of The Platform, SMART, And You
- Test Setup And Benchmarks
- Results: Sequential Performance
- Results: Random Performance
- Results: Tom's Hardware Storage Bench v1.0
- Results: Tom's Hardware Storage Bench v1.0, Continued
- Results: TRIM Testing With ULINK's DriveMaster 2012
- Results: Power Consumption
- Adata SP920: Adding Value With A Nice Bundle
The X210 is pretty awesome, but newer Marvell implementations are built with Haswell-style power features in mind. If you're looking for a drive to use in mobile applications, mind the heat and power consumption stats.
Regards,
Christopher Ryan
I have to say, the plastic or metal chassis a drive comes in doesn't mean much. In the lab, I like a nice heavy metal SSD casing, but in a laptop? You probably want a flimsy plastic chassis. It's not conductive and doesn't add much weight.
Regards,
Christopher Ryan
I agree that a substantial chassis tends to reinforce the perception of a drive's build quality, but much of the time its aesthetic. The component choice on the PCB speaks more to quality. I've seen some downright terrible drives in the fanciest of cases.
Regards,
Christopher Ryan
Impressive ... power consumption is a bit high though, compared to the Samsung 120GB Evo (my current $80 fav)
Are 'microseconds' considered 'milliseconds' ??
Case in point. I read ADATA had released driver updates back in February. I have a SX 900 64G I use for benchmarking. After downloading and running the 525 FieldUpdater it was obvious that no matching driver was included in the newly released 525 driver update package I had downloaded despite the SX 900 being on the list.
I emailed tech support and didn't receive a reply for 7 days despite the confirmation of receipt stating a response would come in 1 to 3 days. I did get a reply after a posting the issue to their Facebook page in which I was told "Please see the attached for the signature file you needed for the firmware update as it took us some time to obtain it from our headquarters"
Indeed a matching 525 FieldUpdater driver had been included in this email and I installed it.
The result was something to see, erratic and downright wonky, come to mind. A malfunction in the installation I presumed and so I downloaded the ADATA software suite SSD Tool which wouldn't and couldn't Security Erase the disk. Parted Magic was then installed and did the job allowing me to clean reinstall the driver I'd been emailed and hopefully resolve the erratic performance issue. In part it resolved erratic reads and writes but now every ATTO write above the 128k mark had lost 40% of its performance capability from the benchmarks using the previous driver.
Support were emailed including ATTO screen captures, and their response was to send it back for replacement.
_Wait a minute just sent me the old driver I stated in my response
Can't "After consulting with our headquarters, it’s confirmed that once the firmware is updated, it cannot be removed or fallback to the older version".
Having no choice I hinted that as this debacle had originated with them that an upgrade from 64G would be an appropriate compensation for the enormous waste of time, and additional costs I was incurring.
The disk was packaged and insured then sent to ADATA for a cost of 20$ and it should be noted that locally this additional $20 represented the retail difference between a 64G and a 128G capacity.
Tracking informed me ADATA received the disk in 5 days and later confirmed by them. Once the 3 to 5 days turn over stated by them came and passed and on day seven a status inquiry was sent and 6 hrs later, confirmation a package to me had been put in the hands of mail services.
That was March 21 and I received the disk 11 days later April 1st in my snail mail box no signature required (a less scrupulous person could easily exploit that).
The round trip experience total is six weeks with a dollar output of over 20% of the sticker price the disk was purchased for. Total time wasted on their error, fiddling with downloads installing and removing ADATA brand software that didn't accomplish the required tasks (additionally the Acronis suite that was packaged with the drive never successfully completed the key transaction on their site) research, installations removal... ya'll get the picture.
Does the SX 900 perform well/competitively... all the published reviews and accompanying benchmarks hold true, so yes it's a cracker even in the 64G capacity.
Would I recommend it?
Only to a tech masochist with tons of time to waste as well a a few bucks.
Now you have a more complete picture.
Impressive ... power consumption is a bit high though, compared to the Samsung 120GB Evo (my current $80 fav)
Are 'microseconds' considered 'milliseconds' ??
Remember (at least for idle) that these are active numbers. That is, the drive and host aren't collaborating to put the drive in a lower power state. In a mobile application, most every SSD is going to drop to lower sleep-state levels, but at the cost of higher latency when returning to idle. For the sake of consistent testing we choose to use active idle.
Did we mix up our units somewhere?
Regards,
Christopher Ryan
I disagree. I actually like seeing the authors chime in. I think it builds a better community as the authors seem more accessible and thus reliable and relatable.
I also don't think Christopher Ryan is trolling in the strictest definition of the word; He's not trying to stir things up for the sake of starting a post war. He's simply continuing to give his opinion and replies in threads. Why should they not be able to make comments on articles they've written? It is expected that any knowledge they can impart to the community is considered of value, regardless if it's in the original article or in the comments section. I think there's also a potential consideration where maybe they have more to share, but due to time constraints or simply available space, they didn't say all they really wanted. So they chime in in the comments section. Kinda like the "extras" on a DVD or BR... "Oh, Director Commentary. Cool!"
I disagree. I actually like seeing the authors chime in. I think it builds a better community as the authors seem more accessible and thus reliable and relatable.
I also don't think Christopher Ryan is trolling in the strictest definition of the word; He's not trying to stir things up for the sake of starting a post war. He's simply continuing to give his opinion and replies in threads. Why should they not be able to make comments on articles they've written? It is expected that any knowledge they can impart to the community is considered of value, regardless if it's in the original article or in the comments section. I think there's also a potential consideration where maybe they have more to share, but due to time constraints or simply available space, they didn't say all they really wanted. So they chime in in the comments section. Kinda like the "extras" on a DVD or BR... "Oh, Director Commentary. Cool!"
I'm making an effort to try and encourage more discussion, which means becoming more active in the comment section. Previously, I tended to let them be. Now, I think I can increase the utility of the comments section over time by more active participation.
As someone who used to read and post comments on Tom's as a reader, I always thought it was awesome that I had a place where I could interact with the author. I want to see more of that with my reviews, so my participation is the best way to make that happen.
Regards,
Christopher Ryan
mSATA SSDs are still using the SATA 3.1 host spec, as are most M.2s. There are a few M.2 PCIe SSDs, but there are currently next to no applications for them. We have more M.2 and mSATA reviews on the way, so you should be able to judge for yourself.
Regards,
Christopher Ryan