Samsung SM951 (512GB) PCIe SSD Review
by Kristian Vättö on February 24, 2015 8:00 AM ESTAnandTech Storage Bench - Light
The Light trace is designed to be an accurate illustration of basic usage. It's basically a subset of the Heavy trace, but we've left out some workloads to reduce the writes and make it more read intensive in general.
AnandTech Storage Bench - Light - Specs | ||||||||||||
Reads | 372,630 | |||||||||||
Writes | 459,709 | |||||||||||
Total IO Operations | 832,339 | |||||||||||
Total GB Read | 17.97 GB | |||||||||||
Total GB Written | 23.25 GB | |||||||||||
Average Queue Depth | ~4.6 | |||||||||||
Focus | Basic, light IO usage |
The Light trace still has more writes than reads, but a very light workload would be even more read-centric (think web browsing, document editing, etc). It has about 23GB of writes, which would account for roughly two or three days of average usage (i.e. 7-11GB per day).
AnandTech Storage Bench - Light - IO Breakdown | |||||||||||
IO Size | <4KB | 4KB | 8KB | 16KB | 32KB | 64KB | 128KB | ||||
% of Total | 6.2% | 27.6% | 2.4% | 8.0% | 6.5% | 4.8% | 26.4% |
The IO distribution of the Light trace is very similar to the Heavy trace with slightly more IOs being 128KB. About 70% of the IOs are sequential, though, so that is a major difference compared to the Heavy trace.
AnandTech Storage Bench - Light - QD Breakdown | ||||||||||||
Queue Depth | 1 | 2 | 3 | 4-5 | 6-10 | 11-20 | 21-32 | >32 | ||||
% of Total | 73.4% | 16.8% | 2.6% | 2.3% | 3.1% | 1.5% | 0.2% | 0.2% |
Over 90% of the IOs have a queue depth of one or two, which further proves the importance of low queue depth performance.
The SM951 yet again provides roughly twice the data rate compared to the XP941 and with a full drive the difference is even more significant.
The same goes for average latency where the SM951's score is about one third of the XP941's. The SM951 can without a doubt boost performance with lighter IO loads as well, although in very light workloads the bottleneck tends to be the speed of user input (think about document creation for instance).
128 Comments
View All Comments
Kristian Vättö - Wednesday, February 25, 2015 - link
I didn't test any OCZ drives for this review because the Vector 180 NDA is due in a couple of days, so stay tuned.IlikeSSD - Thursday, February 26, 2015 - link
Vertex 460A is good enough to show the difference (maybe even Arc))Shadowmaster625 - Wednesday, February 25, 2015 - link
You should clip on a simple heatsink and run the tests again.squatsh - Wednesday, February 25, 2015 - link
Based on the orientation of the drive in it's M2 slot. (Ie is the controller facing up or down) Can it be made clear if we could just put a small thermal pad and one of those tiny heatsinks (like you can get for VRMs) on the controller (assuming that is what is overheating) in order to stop the thermal throttling on desktops and laptops with enough space.Luke212 - Wednesday, February 25, 2015 - link
what happened to the Intel P3500? It's been a year and nothing.Mikemk - Wednesday, February 25, 2015 - link
Comment 100Sorry, couldn't resist
awall13 - Thursday, February 26, 2015 - link
Thanks for the great review. I like the changes in how the data is presented. One idea: In the performance consistency section, the steady-state performance could be presented with a single two-color bar per drive, with the full bar representing the average performance, and the shorter bar representing the average - 1 standard deviation. I'd still keep the second chart showing the standard deviation by itself, perhaps. Just a thought.gseguin - Thursday, February 26, 2015 - link
Performance performance performance. Ok, I get it, its fast and hard to find.Since we saw many issues crop up with the predecessors, with eventual fixes by Samsung, how much confidence is there in the drive itself from Anandtech... You know the controllers, you know the memory, you've covered different company's testing methodologies, how much confidence do you have in this product, and why is that absent from the introduction and final thoughts?
Kristian Vättö - Thursday, February 26, 2015 - link
All Samsung's issues have been related to SSD with planar TLC NAND (i.e. 840 & 840 EVO). There is absolutely no reason to believe that the SM951 shares the issues because it's based on MLC NAND and Samsung's MLC NAND based SSDs have been flawless (830 & 840 Pro).kgh00007 - Saturday, February 28, 2015 - link
Hey, nice article. I like the new SSD test suite. Is there any way to add a long term read consistency test, or any sort of read consistency test in light of the Samsung 840 EVO issues?I'm an 840 EVO user with drops in read speed to below 50Mb/s. In fact I have never gotten the advertised 500MB/s read speeds on any SSD I own. Is that because I am using them as the boot drive?
Do you test with the SSD as a second drive in the system? So taking out the overhead of the OS running on the drive as well as the test suite?
Cheers, keep up the good work, I've been learning from this site for a long time....