SanDisk X300s (512GB) Review
by Kristian Vättö on August 21, 2014 2:15 PM ESTPerformance Consistency
Performance consistency tells us a lot about the architecture of these SSDs and how they handle internal defragmentation. The reason we do not 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 or cleanup routines directly impacts the user experience as inconsistent performance results in application slowdowns.
To test IO consistency, we fill a secure erased SSD with sequential data to ensure that all user accessible LBAs have data associated with them. Next we kick off a 4KB random write workload across all LBAs at a queue depth of 32 using incompressible data. The test is run for just over half an hour and we record instantaneous IOPS every second.
We are also testing drives with added over-provisioning by limiting the LBA range. This gives us a look into the drive’s behavior with varying levels of empty space, which is frankly a more realistic approach for client workloads.
Each of the three graphs has its own purpose. The first one is of the whole duration of the test in log scale. The second and third one zoom into the beginning of steady-state operation (t=1400s) but on different scales: the second one uses log scale for easy comparison whereas the third one uses linear scale for better visualization of differences between drives. Click the dropdown selections below each graph to switch the source data.
For more detailed description of the test and why performance consistency matters, read our original Intel SSD DC S3700 article.
 |
|||||||||
| Default | |||||||||
| 25% Over-Provisioning | |||||||||
The IO consistency is good but obviously not as good as the Extreme Pro due to lower over-provisioning (7% vs 12%). The architecture is still the same, though, as first the performance drops to around 10K IOPS, which is followed by a higher throughput burst. At steady-state the X300s averages about 5K IOPS, which is actually similar to the Crucial MX100 but with added over-provisioning the X300s gets close to the Extreme Pro level.
 |
|||||||||
| Default | |||||||||
| 25% Over-Provisioning | |||||||||
 |
|||||||||
| Default | |||||||||
| 25% Over-Provisioning | |||||||||
Facebook
Twitter
RSS
34 Comments
View All Comments
hojnikb - Friday, August 22, 2014 - link
*at a cost of capacity :) :)Kristian Vättö - Friday, August 22, 2014 - link
Yeah, fundamentally SLC, MLC and TLC are the same. Of course there are some silicon level optimizations to better fit the characteristics of each technology but the underlaying physics are the same.I'm thinking that pseudo-SLC is effectively just combining the voltage states of MLC/TLC. I.e. output of 11 or 10 from the NAND would read as 1, which allows for higher endurance since it doesn't matter if the actual voltage state switches from 11 to 10 due to the oxide wear out.
Spoony - Friday, August 22, 2014 - link
I believe you'd lose half the capacity on your drive. The MLC drives store two bits per cell, so they would store a 1 and a 0 for example. If you now are only allowing it to store a 1, then you've halved the capacity of the cell. Across the entire drive, this would thus halve the total drive capacity.As far as performance (read/write speed) I think this would be affected less. SSDs rely on parallelism to extract performance from NAND. The array is just as parallel as before. There might be an impact to performance having to do with extracting less information from each cell, how much this would be I'm not sure.
I think the changes to firmware would have to be much more substantial than just re-programming how many bits per cell are stored. There is most likely a lot of interesting logic around voltage handling at very small scales. Perhaps even looking at how voltages from neighbouring cells influence each other. I'm not sure how serious this firmware gets regarding physics, but it must have to do some sort of compensation because the drives seem pretty reliable.
hojnikb - Friday, August 22, 2014 - link
Yeah, ive "edited" the post to reflect the loss of capacity. Obviously capacity drops, but its still waay cheaper than real SLC solutions.I bet write speeds would actually go up (since this is the exact reason why samsung and sandisk are doing pSLC) but read would stay unaffected (since this is controller/interface limited anyway).
BillyONeal - Thursday, August 21, 2014 - link
Erm, what is MBAM for then? http://technet.microsoft.com/en-us/library/hh82607... My work PC has remotely managed BitLocker.
Zink - Thursday, August 21, 2014 - link
MBAM is "Malwarebytes Anti-Malware" malware removal toolBillyONeal - Thursday, August 21, 2014 - link
@Zink: It is also "Microsoft Bitlocker Administration and Management"Kristian Vättö - Thursday, August 21, 2014 - link
Looks like I should have done my research better. Thanks for the heads up, I've edited the review to remove the incorrect reference.thecoolnessrune - Thursday, August 21, 2014 - link
Yep, I the company I work with also has all of our drives encrypted with Bitlocker. It's managed by MBAM and integrated right into the rest of Active Directory Management. Really simple for the Domain Administrators (and relevant IT HelpDesk personnel) to use and manage.eDrive can fit in the Enterprise environment quite well.
cbf - Thursday, August 21, 2014 - link
Yup. As the other commenters indicate, the only thing we care about in the Enterprise is BitLocker. Hell, even if it was my personal drive, I'd probably only use BitLocker. I just trust it more than the third party solutions.So why don't you review this drive's encryption features using BitLocker. Anand showed how to do this last April: http://www.anandtech.com/show/6891/hardware-accele...