Samsung SSD 850 EVO 250GB Review

I consider myself as a mainstream user. My computer needs do not require the latest and the fastest at a premium price. After the release of the 850 PRO line which I reviewed last July, I was eager to see what Samsung would come up with the 850 EVO line since it is a better match for my computing usage and wallet.

I am familiar with the Samsung SSD products. I knew that Samsung would release an 850 EVO version. As expected, the 850 EVO is built around what made the success of the 840 EVO, Turbowrite, TLC NAND and now is equipped with the latest 32 layers 3D V-NAND.

As we will see, V-NAND technology is a big boost for TLC mostly because of its higher bit per cell.
The most obvious one is endurance, 850 EVO SSD drives are backed by a five year limited warranty. Available capacities are 120GB, 250GB, 500GB and 1TB.

Specifications

Usage Application	Client PCs
Capacity	128GB, 256GB, 512GB, 1TB(1024GB)
Dimensions (LxWxH)	100 x 69.85 x 6.8 (mm)
Interface	SATA 6Gb/s (backward compatible)
Form Factor	2.5 inches
Controller	120/250/500GB: Samsung MGX Controller 1TB: Samsung MEX Controller
NAND Flash Memory	Samsung 32 layer 3D V-NAND
DRAM Cache Memory	256MB (120GB) or 512MB (250GB&500GB) or 1GB (1TB) LPDDR2
Performance*	Sequential Read:	Max. 540 MB/s
	Sequential Write**:	Max. 520 MB/s
	4KB Random Read (QD1):	Max. 10K IOPS
	4KB Random Write (QD1):	Max. 40K IOPS
	4KB Random Read (QD32):	Max. 98K(500GB/1TB) Max. 97K(250GB) Max. 94K(120GB)() IOPS
	4KB Random Write (QD32):	Max. 90K(500GB/1 TB) Max. 88K(120GB/250GB)
TRIM Support	Yes (Requires OS Support)
Garbage Collection	Yes
S.M.A.R.T	Yes
Data Security	AES 256-bit Full Disk Encryption (FDE)
	TCG/Opal V2.0, Encrypted Drive (IEEE1667)
Weight	Max. 66g (1TB)
Reliability	MTBF: 1.5 million hours
TBW	120/250GB: 75 TBW 500GB/1TB: 150 TBW
Power Consumption	Active Read/Write (Average): Max. 3.7W (1TB)/Max. 4.4W(1TB)
	Idle: Max. 50mW
	Device Sleep: 2mW(120/250/500GB), 4mW(1TB)
Temperature	Operating:	0°C to 70°C
	Non-Operating:	-40°C to 85°C
Humidity	5% to 95%, non-condensing
Vibration	Non-Operating:	20~2000Hz, 20G
Shock	Non-Operating:	1500G, duration 0.5m Sec, 3 axis
Warranty	5 years limited

Pricing

Description	Capacity (GB)	Model	MSRP	$/GB
850 EVO Series SATA III 2.5″	120	MZ-75E120B/AM	$99.99	$0.84
	250	MZ-75E250B/AM	$149.99	$0.60
	500	MZ-75E500B/AM	$269.99	$0.54
	1TB	MZ-75E1T0B/AM	$499.99	$0.50

Internals and Features

MGX Controller

The 120/250/500GB are powered by a “new” controller, the MGX. Unfortunately, there is not much information about it. Rumors are it is a dual core which is a downgrade compared to the MEX 3 core ARM. Per Samsung, the MGX has been optimized for low capacity model and sequential read/write and random low level QD. Which are the most relevant I/O type regarding user experience.

TLC, 3D V-NAND

Introduced with the 850 PRO, the 3D V-NAND revolutionized the SSD industry. I already detailed the specific in the 850 PRO review, but here is the essentials. Up to the V-NAND, the race was about shrinking the die, cramming more and more cells in the same space. On one hand, costs are reduced and prices drop. On the other hand, as the cells are getting closer and closer, error rates increased and writes required higher voltage.

Instead of shrinking, V-NAND stacks vertical layers. By going vertical, there are now more real estates for the cells in size and in between. Which translates into more room for voltage state changes, less interference and lower programing voltage. The benefits are, less retries due to interferences, lower power consumption. The benefits are, less retries due to interferences, lower power consumption, increase NAND endurance and overall I/O performance

TLC or MLC 3 bits per cell really take advantage of the V-NAND since there are more bits cells compared to the MLC (2 bits per cell). Beside the performance increased, the endurance rating was greatly improved allowing Samsung to offer a 5 year warranty compared to the 3 years for the previous EVO generation.

TurboWrite

By reserving a small portion of the V-NAND as emulated SLC, that space acts as a buffer, all I/O writes perform at SLC speed. As long as the writes stay under the buffer size, write I/O should perform as SLC speed. Once the buffer is full, subsequent writes fall back to TLC performance.

TurboWrite SLC Buffer Size
Capacity	120GB	250GB	500GB	1TB
TurboWrite Size	3GB	3GB	6GB	12GB

Outside of very specific softwares, it is unlikely that a home computer usage exceeds 3GB of continuous I/O writes.  In most cases, users would be able to take advantage of the SLC speed at the TLC price.

RAPID

The RAPID technology uses the computer host SDRAM as an extra level of cache. The cache size is dynamic, 25% of the RAM up to 4GB, split 50/50 between read and write. The most noticeable about the write cache is it focuses on small random I/O by collating the data and writes it back in larger blocks. The read cache is persistent, meaning a copy of the data map is written to disk every so often.

Dynamic Thermal Guard

Heat is a concern when dealing with an enclosed environnement such as a laptop or notebook when air flow is premium. The Dynamic Thermal Guard throttles the controller clock if the temperature reach a certain threshold.

Power Consumption

Active power measurements during writing operations are shown below

Software Package

“Samsung Data Migration” is a convenient disk cloning software by Clonix. The utility will always detect the OS drive as the source, which is a good failsafe. It only works if at least one SSD is a Samsung. At this time, it is Windows only. The latest version 2.7 puts a new restriction, it only recognized Samsung SSD as the target drive, meaning I can not chose a non-samsung SSD drive as a target. The previous version would only care if at least one drive was a Samsung product. Although the Samsung Data Migration will get the job done, truth to be told, I like the Intel Data Migration better. It has a couple more features  such as, create a bootable media rescue disk and a drive cleanser tool. Although “secure erased” would be as efficient.

The Magician 4.5, also Windows only, is well designed and all vital information is readily available. Firmware updates and customized over-provisioning can be set up with one click of the mouse. With another couple of mouse clicks, the O.S. is optimized for the SSD. There is no need to navigate through Windows registry keys and make changes. In doubt, “Maximum Reliability” is a good option.

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 7 (from an image), filled with about 120GB of data total and benchmarks were run from the tested unit acting as the OS drive.

Steady state: This state occurred overtime 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.

Benchmark Workstation Main Components
CPU	Intel Core i3-2120 Processor @3.30GHz
Motherboard	ASUSTeK Computer INC. P8Z77-V LX
RAM	8,192 MB (4,096 x 2) PNY Optima DDR3 – PC3-10666
GPU	NVIDIA GeForce GTS 450
OS	Windows 7 Pro 64bits
Storage Driver	iaStorA 12.9.0.1001
OS Hard Drive	The reviewed SSD unit

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

Strangely enough the writes test did not fully reflect the use of RAPID with Anvil’s Storage Utilities. I reran a series of test with CrystalDiskMark. CMD output is closer to what I expected to see regarding writes with RAPID enabled.

The EVO 850 performance is neck to neck with the 850 PRO. Not surprising considering ASU tests are performed with a 1GB, which is well with the 3GB SLC buffer size builtin the 850 EVO 250GB.

Beside synthetic benchmarks, I feel important to temper the expectations by testing the product in a context comparable to a home usage pattern. “PCMark 8 storage” and “PC Mark Vantage” are based on traces recorded from the most popular applications and games to provide real-word performance.

Conclusion

Once again, Samsung delivers a great product, taking the lead in the SSD space with its advanced technology. 3D V-NAND is really a game changer. The 3rd TCL generation offers pretty much everything the consumer would want to see from an SSD, performance, high endurance, encryption, low power consumption, great software package and backed by a 5 years limited warranty."3D V-NAND IS REALLY A GAME CHANGER"

I did leave out the price, didn’t I? I was hoping for a more aggressive pricing from Samsung. I understand that pricing is “subjective”. One can argue that for the new 3D V-NAND architecture, the performance and the 5 years warranty the pricing is right. The truth is, there are other options for the price range.

In the final words of my 850 PRO review, I suggested to either wait for the 850 EVO or take a look at the MX100. At roughly the same capacity, the MX100 256GB is priced for $110 while the EVO 850 250GB cost $140, on Amazon as of 12/2014.

Sure, 2 more warranty years could worth the extra $30. But if the EVO 850 250GB were priced at or below $120, it would be a clear winner, so watch out for “street prices” as well.

The other thing I was expecting to see was a 2TB SSD version. I am pretty sure Samsung can produce one. Maybe they didn’t because the market is not ready for it, but I wish they released a 2TB, just because they can.

Michael Nguyen
Categories: Computers, Reviews
Tags: Samsung, Samsung EVO, Samsung Reviews, ssd, Storage