 |
Review: Samsung 860 PRO SATA 256GB Reviewed by: J.Reynolds Provided by: Samsung Firmware: RVM01B6Q |
Introduction
Welcome to Myce’s review of the Samsung 860 PRO 256GB SATA
SSD (hereafter referred to as the Samsung 860 PRO).
The Samsung 860 PRO is the successor to Samsung’s very
successful 850 PRO.
The Samsung 860 PRO uses Samsung’s latest 64 layer, V-NAND 2bit
MLC technology, and the newly designed MJX controller.
It will be interesting to see how the Samsung the 860 PRO compares
to the Samsung 860 EVO, which recently set new records for an SATA drive in
many of our tests.
Let the battle of 2 bit, MLC NAND (the 860 PRO) versus 3
bit, TLC NAND with clever SLC Write Cache technology (the 860 EVO) commence...
Packaging
To get started let’s have a look at the Samsung 860 PRO’s
packaging -
Market Positioning and Specification
This is how Samsung positions their 860 PRO –
Here is Samsung’s specification for the 860 PRO –
Now let's head to the next page, to look at my approach
to testing Client SSDs.....
Testing Approach
When reviewing the performance of a Storage solution there
are three basic metrics to look at:
1.
IOPS – the number of Input/Output Operations per Second
2.
Bandwidth – the number of bytes transferred per second (usually measured
in Megabytes per second, ‘MB/s’)
3.
Latency – the amount of time each IO request will take to complete
(usually, in the context of Solid State Storage solutions, measured in
Microseconds, which are millionths of a second).
It is true to say that IOPS and Bandwidth had both been
growing rapidly before the advent of Solid State Storage, but Latency can only
be significantly decreased by eliminating mechanical devices, and thus Latency
is the single most important improvement that Solid State solutions deliver to
enhance performance.
Latency in a technical environment is synonymous with delay.
In the context of a Solid State solution it is the amount of time between an IO
request being made, and when the request is serviced.
Bandwidth, also commonly referred to as ‘Throughput’, is the
amount of data that can be transferred from a storage device to a host, in a
given amount of time. In the context of Solid State solutions it is typically
measured in Megabytes per second (MB/s).
A great Solid State solution offers an effective balance of
all three metrics. High IOPS and Bandwidth is simply not enough if Latency
(the delay in an IO operation) is too high.
Queue Depth is the average amount of IO requests
outstanding. If you are running an application and the Average Queue Depth is
one or higher and CPU utilisation is low, then the application’s performance is
most probably suffering from a ‘Storage Bottleneck’.
It is true to say that a typical PC user will very rarely
cause a modern SSD to see a Queue Depth greater than 1 or 2. So for Client SSDs
we need to primarily focus on performance at low Queue Depths.
Another important aspect to consider with an SSD is the
state of its NAND when an IO task begins. When an SSD is new, or immediately
following a purge (a Secure Erase for an SATA device) being performed, it is in
a Fresh out of Box (‘FOB’) state and in this state all of its Blocks of NAND
are clean and able to immediately accommodate the writing of new data.
Typically SSDs are supplied with a greater capacity of ‘Total NAND’ than their stated
‘User Capacity’ and the difference between them (Total NAND – User Capacity) is
known as an Over Provision (‘OP’) at the firmware level.
If an IO Task that involves writing new data can complete
without the supply of clean blocks being exhausted it will complete more
quickly than if blocks must first be cleaned on the fly before writes can be
accommodated. The number of free blocks available may also impact on performance
(think of it this way - the more free blocks there are the easier it is to find
one to write to). An SSD will continue to write to clean blocks until there
are no more available after which it must then free up blocks by completing an
Erase/Write cycle on the fly before it can write new data. Blocks that have
been written to are flagged as being able to be cleaned when either the logical
address they are associated with in an Operating System is written to again or
when a Trim instruction is sent by the OS to indicate that a range of logical
addresses (which map to physical blocks) no longer contain live data (for
example, in Windows, Trim instructions are sent to an SSD, when a file is
logically deleted, to indicate that all of the physical blocks which contained
the file no longer hold live data).
An SSD’s controller performs a process known as ‘Garbage
Collection’ which gathers together spaces that no longer hold live data so that
it can create clean blocks in preparation for accommodating the writing of new
live data. Blocks are contained within Pages and only complete Pages can be
erased in preparation to accommodate new writes, so one of the responsibilities
for Garbage Collection is to shuffle blocks out of partially filled pages so
that whole pages can then be cleaned. Garbage Collection can be performed as a
regular background task and on the fly. The effectiveness of an SSD’s Garbage
Collection has a significant impact on its long term performance. It is
important to note that a Trim command does not itself clean blocks and it will
always take a bit of time for Garbage Collection to follow up and actually
complete the cleaning process.
An SSD maintains a table, that can be used by an OS, which
holds the mapping of its physical blocks to logical addresses. Effectively,
the OP is increased above that set at the firmware level whilst the drive’s
user capacity is not full of live data. In Windows a user can effectively
choose to underline their commitment to increasing the level of OP by not fully
allocating the drive’s user capacity to partitions.
When a drive is compelled to clean blocks on the fly to
accommodate new data it moves from an FOB state towards what is known as a ‘Steady
State’. A Steady State is achieved when performance is steady and no longer
changes significantly over time. Testing of Enterprise SSDs is always
performed when a drive is in a Steady State. It is fair to say that typically
a Client SSD will spend most of its time in an FOB state (or near to FOB state)
and it's in this state that our testing is performed using the Desktop PC.
Remember though that one can expect to see a performance drop when the drive
holds increasing amounts of live data, as the pool of free blocks (the
effective OP) becomes smaller.
Whilst most Client SSD users need not be overly concerned
about Steady State performance we do push an SSD to its limits as part of our
testing on the OakGate Test Platform.
So what performance characteristics make for an excellent
Client SSD?
Put simply, we look for a solution that provides both
excellent Sequential IO performance and excellent Random IO performance.
Excellent Sequential performance supports the rapid transfer of large amounts
of data from one place to another, such as when copying a movie, loading a game,
or running a backup. Excellent Random IO means that a drive will support the
rapid reading, writing, and updating of relatively small files that are
randomly placed on a drive (such as is required by the Windows Operating System),
the launching of applications, or by a database based application. Sequential
performance is most often measured in terms of MB/s (Megabytes per Second) and
Random IO is most often measured in terms of IOPS (IO Operations per Second).
Modern SSDs deliver low Latency and support tens of thousands of Random IOPS
and whilst very few PC users really need support for such a high level of IOPS
it does mean that every IO will be fast.
Manufacturers most frequently quote the headline maximum
Sequential Read and Write Bandwidth for a drive. They also regularly cite the
maximum IOPS level for 4K Random Reads and Writes. Operating Systems are known
to make extensive use of the 4K IO size and this is why strong 4K Random Read
and Write performance is considered important.
I use two test platforms for testing Client Storage
solutions –
Firstly, a Desktop PC, with the following specification: CPU
– Intel Core I7 6700K, Motherboard – Asus Maximus VIII Extreme (Z170), System
Drive – Intel 750 400GB, GPU – EVGA GeForce GTX970 FTW, RAM – 32GB Corsair
Dominator Platinum, Cooler – Corsair H110i GTX, Windows 10 using Intel RST
15.7.1.1015 and with C States disabled in the BIOS, as this ensures reasonable
consistency from storage benchmarks.
Secondly, an OakGate Storage Test Platform, which is
introduced in an article that you can view by clicking here.
The OakGate Test Platform can be thought as a professional, laboratory
instrument where the test environment is managed strictly and consistently so
that test results from multiple solutions can be compared with great confidence
and precision.
The Desktop PC is used to run a cross section of the most
respected and commonly used storage performance benchmark software, including
AS SSD, Anvil, Crystal Disk Mark, ATTO, and PCMark 8 Storage, together with a
number of real world file copies. Most of these benchmark programs are freely
and easily available for you to run on your own PC. There is a good case for reviewers
to test an SSD as a System Drive, as arguably this is the way in which most
people will use an SSD. However, I choose to test drives as a spare as I
believe this makes it far easier to provide a consistent basis for product
comparisons, which I feel is most important.
The OakGate Test Platform is used to provide an accurate
baseline for a drive’s performance in all of the key aspects of performance,
including Sequential Reads and Writes, Random Reads and Writes, and Random
Mixed Reads and Writes. The OakGate Test Platform is also used to investigate
how a drive behaves when it is pushed to its limits and to measure a drive’s
power consumption characteristics. (All testing on the OakGate Test Platform
is conducted with fully random data and is aligned to 4K boundaries)
In the presentation of test results I include comparisons
with other products I have tested in the same way on the same platforms.
Now let's head to the next page, to look at the results
for the Desktop PC Synthetic Benchmarks.....
Desktop PC – Synthetic Benchmarks
AS SSD
As its name suggests AS SSD was developed specifically to
measure the performance of SSDs. It measures Sequential Read and Write
performance with an IO Size of 16MB and a Queue Depth of 1. It measures Random
4K Read and Write for a Queue Depth of 1 and for 64 Threads. 64 Threads
generates a Queue Depth of 64 (please note that SATA drives support a maximum
Queue Depth of 32, so they are at a disadvantage in this test to NVMe devices,
which support queue depths of 128 or more). The Access Time AS SSD reports is
for 512Byte sequential reads and writes.
The 4K random Reads and Writes performance is particularly
relevant to a drive’s ability to act as a Windows system drive. I use the
default test file size of 1GB.
AS SSD produces a score for Read Performance, Write
Performance and an Overall Score.
The scores are calculated as –
Overall score = (Seq Write x 0.15) + (Seq Read
x 0.1) + (4K Read * 2) + 4K Write + 4K-64Thrd Write + (4K-64Thrd Read * 1.5)
Read score = (Seq Read * 0.1) + 4K Read + 4K-64Thrd
Read
Write score = (Seq Write *0.1) + 4K Write + 4K-64Thrd
Write
For Client SSDs, I feel that there should be an
even greater loading given to the Queue Depth 1 4K Read and 4K Write results
but nevertheless AS SSD is a quick and useful benchmark. I always use a 1GB
test file. We would expect a modern SATA SSD to achieve an overall score
of 1000+.
The latest version of AS SSD can be downloaded here.
Here is the AS SSD result for the Samsung 860 PRO -
This is an excellent result for an SATA drive.
Here is a comparison of the overall AS SSD score with the
other products I have tested –
You can see that the Samsung 860 PRO 256GB is a little
slower than the 860 EVO 250GB in this benchmark.
Anvil’s Storage Utilities
Anvil’s Storage Utilities tests Sequential Reads and Writes
with an IO Size of 4MB, Random 4K Reads and Writes at Queue Depths of 1, 4 and
16 and Random 32K and 128K Writes.
The scores are calculated as –
Overall Score = Read Score + Write Score
Read Score = (Seq 4MB = MB/s x 1) + (4K = MB/s
x 4.5) + (4K QD4 = MB/s x 2.75) + (4K QD16 = MB/s x 1.75) + (32K = MB/s x 1) +
(128K = MB/s x 1.5)
Write Score = (Seq 4MB = MB/s x 1) + (4K =
MB/s x 4) + (4K QD4 = MB/s x 3) + (4K QD16 = MB/s x 3)
I always use a Test size of 1GB and 100%
Incompressible data.
The latest version of Anvil’s Storage
Utilities can be downloaded here.
Here is the Anvil result for the Samsung 860 PRO
-
Here is a comparison of the Anvil Total score with the other
products I have tested -
This is a great score for an SATA SSD but it is again slower
than the Samsung 860 EVO with its SLC Write Cache technology.
Crystal Disk Mark
Crystal Disk Mark is a widely respected benchmark, which is
often used by manufacturers as a basis for publishing their ‘headline’ sequential
read and write speeds. I always run the test with One Thread and a Queue Depth
of 32 (which generates a Queue Depth of 32, being the maximum Queue Depth
supported by SATA drives), a 1GB test file, Random data and 5 passes. The
benchmark performs sequential IO with an IO Size of 512K for the Seq Q32T1
test, sequential IO with an IO Size of 1MB for the Queue Depth 1 Seq test and
Random IO with an IO Size of 4K for the 4K (Queue Depth 1) and the 4K Q32T1
test.
Crystal Disk Mark can be downloaded here (I use the
standard edition).
Here is the CDM result for the Samsung 860 EVO -
You can see that the Sequential Read and Write speeds as
specified by Samsung, of 560MB/s and 530MB/s respectively, have both been
exceeded.
ATTO
The ATTO benchmark tests Sequential IO for a large range of
IO Sizes. I always run the test with the default Queue Depth of 4.
ATTO can be downloaded here.
Here is the ATTO result for the Samsung 860 EVO -
Again, you can see that the maximum Sequential Read and
Write speeds, as specified by Samsung, have both been exceeded.
Now let's head to the next page, to look at the results
for the Desktop PC Real World Benchmarks.....
Desktop PC – Real World Benchmarks
PCMark 8 Storage Benchmark 2.0
This is how Futuremark describes the PCmark 8 Storage
Benchmark –
PCMark 8 Storage benchmark is ideal for testing the
performance of SSDs, HDDs and hybrid drives.
Using traces recorded from Adobe Creative Suite,
Microsoft Office and a selection of popular games, PCMark 8 Storage highlights
real-world performance differences between storage devices. You do not need to
have these applications installed on your system to run the Storage benchmark.
The PCMark 8 Storage benchmark test contains the
following workload traces: Adobe Photoshop light, Adobe Photoshop heavy, Adobe
Illustrator, Adobe InDesign, Adobe After Effects, Microsoft Word, Microsoft
Excel, Microsoft PowerPoint, World of Warcraft and Battlefield 3
You can read a detailed description of each storage test and
how the overall score is calculated in the PCMark 8 Technical Guide by clicking
here.
The results from this benchmark are, I feel, a valuable
insight into how a drive will support real world applications.
I thank Futuremark for providing Myce with a license to use
PCMark 8 Pro.
Here is the result for the Samsung 860 PRO -
Exceeding an overall score of 5,000 is very impressive.
Here is a comparison of the overall score with the other
client products I have tested -
You can see that the 860 PRO 256GB has equalled the score
achieved by the 860 EVO.
File Copy Benchmarks
FastCopy is a useful program for recording how long copying
files to and from a drive takes. FastCopy can be downloaded here.
A Ram Disk (a virtual drive held in RAM) is used as the
source drive when a file is ‘copied to’ the test drive and is then used as the
destination when a file is ‘copied from’ the test drive. This ensures that the
test drive is on the critical path for the time taken.
Here are the results -
Copy a Blu-ray Movie to the Samsung 860 PRO
You can see that the 860 PRO 256GB is faster than the 860
EVO 250GB in this test. This is because the size of the Movie is bigger than the
860 EVO 250GB’s maximum SLC Write Cache size of 12GB and therefore the 850
EVO’s transfer speed drops to its post ‘TurboWrite’ speed for part of the
transfer.
Copy a Blu-ray Movie from the Samsung 860 PRO
Copy a Game to the Samsung 860 PRO
Copy a Game from the Samsung 860 PRO
Copy a folder of JPEGs to the Samsung 860 PRO
Copy a folder of JPEGs from the Samsung 860 PRO
Now let's head to the next page, to look at the results
for the OakGate FOB Tests.....
OakGate Platform - ‘Fresh out of Box’ Tests
These tests provide a highly consistent basis for comparing
solutions. The sequence of tests begins with a purge of the drive to ensure
that it starts in a FOB state.
The tests cover all of the essential IO performance
characteristics.
Sequential Writes
This test performs 20 seconds of Sequential Write IOs for
each combination of Queue Depths 1, 4, and 32, and IO Sizes of 4K, 128K, and
1024K. IO traffic is limited to an IO Range of 1GB (which is equivalent to a
test file size of 1GB).
Here are the results for the Samsung 860 PRO –
[masterslider id="68"]
Here is a comparison of the 1024K, Queue Depth 32,
Sequential Write performance with the other products I have tested to date –
This is an excellent result which exceeds Samsung’s
specification of 520 MB/s.
Let’s also have a look at how the Sequential Writes Power
Consumption compares, but to do this fairly we must divide the average MB/s by
the average Milliwatts to get a value for the effective work done. Here is the
result –
You can see that the Samsung 860 PRO has an outstanding
level of power efficiency.
Sequential Reads
The test performs 20 seconds of Sequential Read IOs for each
combination of Queue Depths 1, 4, and 32, and IO Sizes of 4K, 128K, and 1024K.
IO traffic is limited to an IO Range of 1GB.
Here are the results for the Samsung 860 PRO –
[masterslider id="69"]
Here is a comparison of the 1024K, Queue Depth 32,
Sequential Read performance with the other products I have tested to date –
This is an excellent result for an SATA drive. Just like the
WD Blue 3D and the Samsung 860 EVO, the Samsung 860 PRO is pushing to the very
limit of the SATA Bus bandwidth and it exceeds Samsung’s specification of 560
MB/s.
Random Writes
The test performs 20 seconds of Random Write IOs for each
combination of Queue Depths 1, 4, and 32, and IO Sizes of 4K, 16K, and 32K. IO
traffic is limited to an IO Range of 1GB.
Here are the results for the Samsung 860 PRO –
[masterslider id="70"]
Here is a comparison of the 4K, Queue Depth 1, Random Write
performance with the other products I have tested to date –
This is an excellent result for an SATA drive.
The 4K Random Write IOPS for an IO Size of 4 and Queue Depth
of 32 was 70,000 which is a long way short of Samsung’s specified value of 90,000
IOPS.
Let’s have a look at the Latency Distribution for the 4K, QD
1 performance –
This graph shows the Latency for every IO that was performed
in the 20 seconds of traffic. It shows the Number of IOs (IO Count) that fell
within a particular period of Time (Microseconds). The red line plots the Time
against the percentage of total IOs performed.
You can see that the Samsung 860 PRO achieves a remarkably high
level of consistency and that 99.9% of all IOs have a Latency of 40 Microseconds
or less.
Random Reads
The test performs 20 seconds of Random Read IOs for each
combination of Queue Depths 1, 4, and 32, and IO Sizes of 4K, 16K, and 32K. IO
traffic is limited to an IO Range of 1GB.
Here are the results for the Samsung 860 PRO –
[masterslider id="71"]
Here is a comparison of the 4K, Queue Depth 1, Random Read
performance with the other products I have tested to date –
This is an excellent result for an SATA drive but falls
behind the remarkable result achieved by the Samsung 860 EVO.
The 4K Random Read IOPS for an IO Size of 4, and Queue Depth
of 32, was 86,000, which is a long way short of Samsung’s specified value of 100,000
IOPS.
4K Random Mixed Reads/Writes
The test performs 20 seconds of 4K Random Mixed Reads/Writes
for each combination of Queue Depths 1, 4, and 32, and Read/Write ratios of
0/100, 30/70, 50/50, 70/30, and 100/0. IO traffic is limited to an IO Range of
1GB.
Here are the results for the Samsung 860 PRO –
[masterslider id="72"]
Here is a comparison of the 4K Mixed Random 50% Read/50%
Write, Queue Depth 1 performance, with the other products I have tested to date
–
This is an outstanding result.
Now let's head to the next page, to look at the results
for the Oakgate Steady State Tests.....
OakGate Steady State Tests
Sequential Writes to Steady State
This test starts with a purge (Secure Erase), so that the
drive is in a FOB state, and then performs 128K Sequential Writes until twice the
drive’s User Capacity has been written to.
Here is a graph showing the resulting Write Bandwidth over
time -
You can see that the Samsung 860 PRO 256GB very quickly
settles into a steady state of 523MBs, which remains remarkably consistent
throughout the test. This can be compared to the Samsung 860 EVO 250GB, whose
speed crashed to around 300 MB/s after its maximum write cache size of 12GB was
exhausted. Here’s a reminder of what the 860 EVO 250GB looked like –
So, in this test a clear victory for the Samsung 860 PRO
256GB versus the EVO 250GB with its SLC Write Cache technology (though I must
mention that the difference would be far less significant with the higher
capacity EVO drives)
4K Random Writes, FOB to Degraded (Steady State) to Recovered
This test is designed to fully degrade the drive’s
performance and then see how it recovers following a Trim and a period of Rest.
In this test I start with a purge
of the drive to take it to a FOB state.
- FOB Performance – 4K Random
Writes Bandwidth
I then test the FOB 4K Random
Write Performance at Queue Depths of 1 and 32, in an IO Range of 16GB, the
result was as follows –
- Sequential Writes to two times User Capacity
I then performed 128K Sequential Writes to the drive for
twice the drive’s user capacity (as in the previous Sequential Writes to Steady
State test).
- 4K Random Writes for 1 Hour
This was immediately followed by performing 4K Random Writes
to the drive for 1 hour. Here is a graph showing the resulting Bandwidth over
Time –
You can see that the Random 4K Write performance drops
significantly after around 50 seconds and the performance then settles heading
towards a Steady State of around 27 MB/s.
- Fully Degraded, 4K Random Write ‘Steady State’
Performance
I then immediately test the 4K Random Write Performance, in
an IO Range of 1GB (as we did in the initial FOB test), and the result was as
follows -
You can see that performance has dropped significantly
compared to the test performed in the FOB step. At this stage it is fair to
say that the drive’s performance is fully degraded and in a Steady State. It’s
as bad as it can get!
- Trim and Rest for 5m
I then liberally sent Sequential Trim commands to the drive
for a minute (to ensure that the drive’s entire range of logical blocks (= User
Capacity) was trimmed).
I then let the drive rest for 5 minutes before retesting 4K
Random Write performance again. The result was –
- ‘Recovered’ 4K Random Write Performance after Trim and
5 minutes rest
The outcome was –
You can see that the level of performance has improved but
it has not yet approached full recovery. This result is disappointing (as was
the result for the Samsung 860 EVO) when it is compared to the WD Blue 3D,
which had nigh-on fully recovered after 5 minutes.
Now let's head to the next page, to look at the
Conclusions from this review.....
Conclusions
Let’s get straight to the point – why would one buy a
Samsung 860 PRO when a Samsung 860 EVO often has better performance, as can be
seen in many of our test results, and the 860 EVO is significantly less
expensive? There are two reasons – firstly, the 860 PRO has twice the level of
endurance and secondly, the consistency of its sustained sequential writes is
superior.
In truth the Samsung 860 EVO, and indeed its competitors,
such as the WD Blue 3D, is a better proposition for the vast majority of PC
users and logically the 860 PRO should only be considered by heavy duty
workstation users, such as, for example, users that are rendering large videos
most of the day.
So it seems that TLC NAND and clever SLC Write Caching
technology is winning the battle in the SATA Consumer market segment.
Samsung has indicated that the recommended UK price for the
860 PRO 256GB will be GBP £126.49, which can be compared to GBP £90.49 for the
Samsung EVO 250GB, and GBP £83.99 for the WD Blue 3D 250GB (which, I feel, does
make running a couple of either these drives in a Raid 0 configuration an
appealing proposition).
Finally to put the appeal of SATA drives into an overall
market perspective why would a storage performance enthusiast with an M.2 port
not prefer to buy a WD Black PCIe NVMe SSD 256GB (GBP £91.96) or a Samsung 960
EVO 250GB PCIe NVMe SSD (GBP £106.08)? - prices from Amazon UK at the time of
going to press. I suspect the enthusiast is now more often using SATA SSDs as
secondary drives rather than as system drives.
Nevertheless, the Samsung 860 PRO is an excellent and robust
SATA SSD and I am pleased to award our rating of ‘Excellent’ -
