Samsung NVMe SSD 960 PRO Review

It’s that time of the year, Samsung released its new NVMe SSD line, the 960 Pro. As we are familiar with Samsung’s nomenclature, the “PRO” version is MLC based and is equipped with the best components available to the consumer market, whereas the EVO is TLC based.

Last year model, the 950 PRO was available in two sizes, this year, the 960 PRO line comes in three capacities, starting at 512 GB, a 1 TB and a 2 TB model.

"40% SPEED INCREASE IN READ AND WRITE"With the 960 PRO 2 TB, product in hands, let’s look at the specification sheet. The 960 Pro 2 TB clocks at 3,500 MB/s sequential read, 2,100 MB/s sequential write. Remember, the 950 Pro reached 2,500 MB/s sequential read and 1,500 MB/s sequential write. That’s a whooping 40% speed increase in read and write!

Compared to the best SATA model, we are looking at a 536% improvement for the sequential read and a 281% for the sequential write. Crazy numbers right, but so is the pricing. From Amazon, the 512 GB retails for $329.99 ($0.65/GB), the 1 TB for $629.99 ($0.61/GB) and the 2 TB goes for $1299.99 ($0.63/GB)

Pricing

Internals and Features

The 960 Pro is an M.2 form factor and will require PCI-Express 3.0 x4 lanes either with a motherboard with M.2 connectors or an M.2 PCIe adapter.

48 layer MLC V-NAND, 256Gb,  twice the capacity per die compare to the 32 layer version. Samsung Polaris, replaced the UBX, with its 5 cores, one dedicated to the communication with the host.

The unit is still single sided, which is quite impressive considering a total capacity of 2 TB. To fit all the chips on the NMVe stick, Samsung was able to pack the 2GB of LPDDR3 on top of the controller.

High performance is often associated with high heat output. Looking at the small size of the device and the where it is located on the motherboard, it does raise some concerns. Not because of the risk of frying the M.2, but it’s about sustained performance. These storage devices are capable of throttling by lowering the controller frequency.

To address the heat issue, a few layers of copper have been added to the backside of 960 PRO.

Software Package

Unlike the 950 PRO, there were no Samsung’s NVMe drivers provided for the 960 PRO. It will come later, sometime in the 4th quarter, with the Samsung Magician 5. I would expect a better performance or at least more functionalities via the Magician toolkit software.

Data Migration is very reliable and easy to use. It is really a matter of selecting the source and the destination drive. Once the process is complete, just restart the PC and set your new boot drive in the BIOS and that is it.

I used to be able to use it for other SSD brands as long as one drive is a Samsung, but they changed it, the destination drive must be a Samsung drive.

As it stands right now, there’s not much about it. It provides basic system information and a performance test utility.

Testing Protocol

I went through most of the popular benchmark tools, AS SSD, CrystalDiskMark, ATTO, IoMeter, Anvil’s Storage Utility v1.1.0 and PCMark Vantage. But I also used performance monitoring tools such as DiskMon and hIOmon, primarily to validate the tests.

Instead of posting chart after chart, I believe, as a consumer, what is important is how the product fits the needs and not chasing after uber high numbers which are only attainable during benchmarking. For this review, I narrowed it down to Anvil’s Storage Utility, PC Mark Vantage Licensed Pro version, CrystalMark and PCMark8.

Drive conditioning: The SSDs were prepped with Windows 10 (from an image), filled with about 150 GB of data total and benchmarks were run from the tested unit acting as the OS drive.

Steady state: This state occurred over time when the drive went through enough write cycles, or to be more specific program/erase (P/E) cycles, that write performances were consistent or stable. It may take a few weeks before the SSD reaches it, depending on the computing usage, but it can be accelerated using IoMeter.

In summary, Steady State is: Written Data = User capacity x 2, at least.

What numbers are relevant in a real world usage?

Keep in mind that unlike synthetic benchmarks which perform only one specific operation at the time for a predetermined duration, seq read, then seq write then random read, and so on and so forth, real world usage paints a different picture.

All four access types can occur at any time, and different transfer rates and different (I/O access) percentages. For instance, a storage subsystem on a streaming server would mostly see high seq read I/O, large block reads, with very little to none write.

Looking at a database server without blob data type, we would probably see 75% random read, 20% random write and 5% random and seq write. I could either guesstimate the different ratios or figure a method to define a more accurate I/O usage baseline.

I/O Baseline

While it is entertaining to run a bunch of benchmarking tools, expecting huge numbers, the purpose of testing the units is to get a good look at how they perform under realistic desktop usage pattern. That is why I picked PCMark Vantage suite as my usage pattern. By capturing and analyzing I/O during the PCVM run, disk operations are breakdown to percentage read vs. write, random vs. sequential, queue depth and average file transfer size.

With that information, benchmarking makes more sense since all the numbers do not carry the same importance, thus some results are more valuable than others.

In summary, I/O pattern defines what I need from the device vs. what can the device do overall.

The I/O baseline process was explained in the Intel 525 mSATA review.

From the numbers, I rated the I/O usage by activity as follow: Random Read > Random Write > Seq Read > Seq Write and average file size is 128K.

To cover Queue Depth, I used hIOmon during the PC Vantage full run. There is a trial version for a week, which is enough time to build the baseline. Based on the chart below, it is obvious that a benchmark score from a QD 16 (or more) does not carry the same weight as a score from a QD 1.

Performance

It is not surprising that the Samsung NMVe 960 PRO takes a huge lead when dealing with high read I/O thanks to the PCIe 3.0 x4 lane bus. The higher the QD, the better the 960 PRO performs.

Anvil’s Storage Utilities. Higher is better

Anvil’s Storage Utilities. Higher is better

Anvil’s Storage Utilities. Higher is better

PCMark Vantage shows very little improvement between last year product, the 950 Pro versus the 960 Pro.

PCMark Vantage. Higher is better

PCMark8. Higher is better

Interesting numbers here, although the bandwith score is about twice the SATA III models, the storage scores are quasi identicle.

PCMark8. Higher is better

Even at different file sizes, 4GB and 32GB, CDM scores stay very consistent.

CrystalDiskMark. Higher is better

Conclusion

With a street price at almost $1300, I don’t see how to get the return on the investment. Traced back benchmarks show very little improvements, even when compared to SATA III devices. That has nothing to do with Samsung. At this time, Operating Systems and applications are not taking advantage of what the SSD has to offer. Also, not all applications are I/O heavy, certainly not your everyday PC computing usage. Take gaming for instance, WoW traced benchmark in PCMark8 shows only a 1 second difference, over 58 seconds, between the 960 Pro and the 850 EVO.

I can only think of a few applications that would begin to take advantage of the 960 Pro, like virtualization, databases and any kind of heavy imaging or editing like music, video, 3D. At that point, I guess, we are entering the realms of  high budget hobbies or professionals.

Once again, Samsung does what Samsung does best, putting together another fantastic consumer product.  Let’s recap, 2 TB of pure awesomeness V-NAND, a 5 cores controller, 2 GB of DDR3 cache on a 2280 form factor is beyond impressive. As far as I am concerned, Samsung has been leading the SSD SoHo market for several years now and is unlikely to give up its lead.

Filed in Reviews. Read more about Editorspick, Samsung SSD and Ssd.