The new Samsung 970 EVO isn't their top of the line consumer SSD, but it might as well be. With their latest 3D TLC NAND flash memory and SSD controller, the 970 EVO offers almost all the performance of its PRO counterparts but without such a steep price premium.

Phoenix Rises

Today, Samsung is launching two new stacks of SSDs: the 970 EVO and the 970 Pro. The new Samsung 970 EVO is the more mainstream TLC-based option from Samsung's new generation of consumer NVMe SSDs, while the 970 PRO is using MLC NAND flash memory. From our perspective, the much more affordable 970 EVO will be the more interesting product.

The broad strokes of the 970 EVO have been obvious for months thanks to availability of the Samsung PM981 client SSD for OEMs. From that drive, we knew that the replacement for the 960 EVO would move from 48-layer 3D TLC to 64-layer TLC, and the Samsung Polaris controller would be replaced by the Phoenix controller. That combination was beating some records set by the 960 PRO when we tested the PM981 in November, so we've been looking forward to the 970 EVO for quite a while.

Read Our Samsung PM981 SSD Review

Samsung hasn't shared many architectural details of the new Phoenix controller, but like its earlier NVMe controllers it uses a PCIe 3.0 x4 interface and includes 8 channels for NAND flash access. Like the previous generation Polaris controller, there are five CPU cores, with one dedicated to host-side communications. We know the controller itself is capable of providing very high performance, because it is also used in Samsung's top of the line enterprise SSD, the Z-SSD SZ985. Performance with Samsung's 3D TLC NAND will be lower than with their Z-NAND, but not due to controller bottlenecks. The 970 EVO brings support for some of the more recent features from the NVMe 1.3 specification and otherwise is equipped to meet expectations for a high-end consumer SSD.

Most other SSD manufacturers have abandoned MLC NAND flash for their consumer product lines or for all of their SSDs with the transition to 64-layer 3D NAND. This means that Samsung's 970 PRO will have very little direct competition and the 970 EVO will be facing off against the flagship SSDs from most other brands. To better match the flagship products the 970 EVO will compete against, the warranty has been lengthened from 3 years to 5 years and write endurance ratings have increased by 50%. This is a substantial bump for anyone looking at Samsung's latest EVO drives.

The Samsung 970 EVO is a broad product family ranging from 250GB up to 2TB. Similar to the last generation, Samsung offered a 2TB model, but from the PRO line - that option has been dropped this time around in favor of the much cheaper EVO version for a 2TB TLC drive. That 2TB model will be the only 970 EVO model to use Samsung's larger 512Gb 64L TLC die, while the smaller drives use the 256Gb die. This prevents the smaller models from suffering the performance penalties that come with the reduced parallelism of having fewer NAND flash chips on the drive. The 512Gb die also enables Samsung to easily fit 2TB onto a single-sided M.2 2280 card without resorting to the expensive DRAM on controller chip stacking that was necessary with the 2TB 960 PRO.

Samsung 970 EVO Specifications
Capacity		250 GB	500 GB	1 TB	2 TB
Interface		PCIe 3 x4 NVMe 1.3
Form Factor		M.2 2280 Single-sided
Controller		Samsung Phoenix
NAND		Samsung 64-layer 256Gb 3D TLC			Samsung 64L 512Gb 3D TLC
LPDDR4 DRAM		512 MB		1 GB	2 GB
SLC Write Cache	Dedicated	4 GB	4 GB	6GB	6 GB
	Dynamic	9 GB	18 GB	36 GB	72 GB
Sequential Read		3400 MB/s			3500 MB/s
Sequential Write (SLC Cache)		1500 MB/s	2300 MB/s	2500 MB/s	2500 MB/s
Sequential Write (TLC)		300 MB/s	600 MB/s	1200 MB/s	1250 MB/s
4KB Random Read	QD1	15k IOPS
	QD128	200k IOPS	370k IOPS	500k IOPS	500k IOPS
4KB Random Write	QD1	50k IOPS
	QD128	350k IOPS	450k IOPS	450k IOPS	480k IOPS
Active Power	Read	5.4 W	5.7 W	6 W	6 W
	Write	4.2 W	5.8 W	6 W	6 W
Idle Power	APST On	30 mW
	PCIe L1.2	5 mW
Write Endurance		150 TB	300 TB	600 TB	1200 TB
Warranty		5 years
MSRP		$119.99 (48¢/GB)	$229.99 (46¢/GB)	$449.99 (45¢/GB)	$849.99 (42¢/GB)

The Samsung Phoenix controller introduces a nickel-coated heatspreader, and the 970 EVO retains the copper foil layer in the label on the back of the drive that was introduced with the 960 generation, but there are no more serious cooling measures on the drive. Samsung claims the 970 units are even less susceptible to thermal throttling thanks to a combination of higher performance before they throttle and a slightly higher temperature limit.

The 970 EVO and 970 PRO will be available for purchase beginning May 7, 2018.

Along with the 970 series, Samsung is launching a new version of their NVMe driver for Windows. As with almost all of our testing, this review sticks to just the NVMe drivers included with the operating system. We have generally not found vendor drivers to offer compelling performance improvements, though they have historically enabled some extra features that Microsoft's drivers don't allow for.

This review will compare the 500GB and 1TB Samsung 970 EVO against:

• The Samsung 960 EVO and 960 PRO, Samsung's previous generation of high-end consumer NVMe SSDs.
• The Samsung PM981, the OEM SSD based on the same controller and flash as the 970 EVO. The PM981 isn't officially available at retail as a standalone drive, but has been shipping to OEMs for several months.
• The Western Digital WD Black with 3D NAND, the first retail drive with WDC's new in-house NVMe SSD controller and their first consumer NVMe SSD with 3D NAND. The WD Black catapulted WD into competition at the top of the consumer SSD market after the first-generation WD Black SSD with planar TLC was a disappointment.
• The Intel SSD 760p, a more mid-range NVMe SSD based on Silicon Motion's SM2262 controller and Intel's 64L 3D TLC.

Results for several older NVMe SSDs and a few SATA drives are also included, as are results from Intel's ultra-premium Optane SSD 900P.

We will be reviewing the Samsung 970 PRO soon, and hope to include the rest of the 970 EVO capacity range in that review. In the near future, our SSD testbed will be updated with the OS and microcode patches for the Meltdown and Spectre vulnerabilities, so it is clear to point out that this review does not contain them yet. Since these updates will have enough of an impact on benchmark results to make new results incomparable to old results, I am also taking the opportunity to make some minor updates to the synthetic test suite. The outgoing 2017 test suite will also be run on several drives with the patched system to measure the performance impact of the patches. There will be several more SSD reviews over the next few weeks using benchmark results that have already been collected before the new test suite debuts.

AnandTech 2017/2018 Consumer SSD Testbed
CPU	Intel Xeon E3 1240 v5
Motherboard	ASRock Fatal1ty E3V5 Performance Gaming/OC
Chipset	Intel C232
Memory	4x 8GB G.SKILL Ripjaws DDR4-2400 CL15
Graphics	AMD Radeon HD 5450, 1920x1200@60Hz
Software	Windows 10 x64, version 1709
	Linux kernel version 4.14, fio version 3.1

• Thanks to Intel for the Xeon E3 1240 v5 CPU
• Thanks to ASRock for the E3V5 Performance Gaming/OC
• Thanks to G.SKILL for the Ripjaws DDR4-2400 RAM
• Thanks to Corsair for the RM750 power supply, Carbide 200R case, and Hydro H60 CPU cooler
• Thanks to Quarch for the XLC Programmable Power Module and accessories
• Thanks to StarTech for providing a RK2236BKF 22U rack cabinet.

AnandTech Storage Bench - The Destroyer

The Destroyer is an extremely long test replicating the access patterns of very IO-intensive desktop usage. A detailed breakdown can be found in this article. Like real-world usage, the drives do get the occasional break that allows for some background garbage collection and flushing caches, but those idle times are limited to 25ms so that it doesn't take all week to run the test. These AnandTech Storage Bench (ATSB) tests do not involve running the actual applications that generated the workloads, so the scores are relatively insensitive to changes in CPU performance and RAM from our new testbed, but the jump to a newer version of Windows and the newer storage drivers can have an impact.

We quantify performance on this test by reporting the drive's average data throughput, the average latency of the I/O operations, and the total energy used by the drive over the course of the test.

The Blue lines indicate the PM981, the OEM version of the 970 EVO.
The Orange lines are the 970 EVO.

The average data rates from the Samsung 970 EVO on The Destroyer are a slight step backwards from the Samsung PM981 OEM drive and from the 960 EVO. All of the TLC-based drives are still performing below even Samsung's older MLC-based NVMe drives, and of course the Intel Optane SSD. This year's Western Digital WD Black offers about the same performance as the 970 EVO.

Average and 99th percentile latencies for the 970 EVO are again very slightly worse than the PM981, but on these metrics the 960 EVO doesn't beat its replacement. The WD Black has notably better 99th percentile latency than the other flash-based SSDs.

There is a clear range of average read latency scores that make up the high-end NVMe market segment. The 970 EVO doesn't stand out from the other drives in that category. For average write latency, scores vary a lot more, and the 970 EVO outperforms its predecessor slightly but fails to match the very good score the PM981 obtained.

The 99th percentile read and write latency scores from the 970 EVO don't break new ground and mostly fail to match the PM981, though the differences aren't large enough to be a serious concern. The WD Black's notable QoS advantage is on the read side, where it is the only flash-based SSD to almost always keep read latency below 1ms.

We didn't have the opportunity to measure power usage of the Samsung PM981 on The Destroyer, so this is our first look at the power draw of the Samsung Phoenix controller on this test. The situation isn't good. The 970 EVO uses twice the energy as the WD Black to does despite both drives offering about the same level of performance on The Destroyer. The power efficiency of the 970 EVO seems to be a big step backwards from the previous generation and is not at all competitive.

AnandTech Storage Bench - Heavy

Our Heavy storage benchmark is proportionally more write-heavy than The Destroyer, but much shorter overall. The total writes in the Heavy test aren't enough to fill the drive, so performance never drops down to steady state. This test is far more representative of a power user's day to day usage, and is heavily influenced by the drive's peak performance. The Heavy workload test details can be found here. This test is run twice, once on a freshly erased drive and once after filling the drive with sequential writes.

Our initial runs of the Heavy test on the Samsung 970 EVO produced results similar to the Samsung PM981, with the 1TB model showing worse performance on an empty drive than a full drive. This seems to be related to the secure erase process used to wipe the drive before the test. Like many drives, the 970 EVO seems to lie about when it has actually finished cleaning up. Adding an extra 10 minutes of idle time before launching the Heavy test produced the results seen here, and in the future all drives will be tested with longer pauses after erasing (all other drives were given at least two minutes of idle time after each erase).

With the odd behavior eliminated, the Samsung 970 EVO comes close to setting a new record on the Heavy test. The empty drive performance of the 1TB model is up in Optane territory, though the full drive average data rate is not much higher than other TLC-based drives. The 500GB model is far slower, and its full-drive performance doesn't even keep pace with the Intel SSD 760p.

The average and 99th percentile latency scores from the Samsung 970 EVO are about normal and in line with its closest competitors, except for the particularly good empty-drive score from the 1TB 970 EVO.

The average write latency of the 970 EVO is fairly typical for a high-end NVMe SSD, but the average read latency of the 1TB 970 EVO in the best case is surprisingly quick. Both capacities show a larger than normal gap between empty and full drive performance, even after accounting for the fact that they are using TLC to compete against the best MLC drives.

The 99th percentile read latency scores from both tested capacities of the 970 EVO show a big difference between full drive and empty drive performance. The 500GB drive's read QoS doesn't seem up to par, but the 1TB model's scores would look pretty good if the WD Black hadn't recently shown up with an MLC-like minimal performance loss when full. The 99th percentile write latency scores of the 970 EVO are good but not substantially better than the competition, and the 500GB model is clearly worse at keeping latency under control than the 1TB model or MLC drives of similar capacity.

The 500GB 970 EVO continues the trend of relatively poor power efficiency from the Samsung Phoenix controller, but the 1TB model in its best case of running the test on an empty drive is fast enough that its overall energy usage is comparable to good SATA drives.

AnandTech Storage Bench - Light

Our Light storage test has relatively more sequential accesses and lower queue depths than The Destroyer or the Heavy test, and it's by far the shortest test overall. It's based largely on applications that aren't highly dependent on storage performance, so this is a test more of application launch times and file load times. This test can be seen as the sum of all the little delays in daily usage, but with the idle times trimmed to 25ms it takes less than half an hour to run. Details of the Light test can be found here. As with the ATSB Heavy test, this test is run with the drive both freshly erased and empty, and after filling the drive with sequential writes.

The peak performance of the Samsung 970 EVO on our Light test is record-setting, but that's only an improvement of 6% over the Samsung 960 PRO's average data rate—not a big enough difference to notice on such a light workload. When the test is run on a full drive, the 970 EVO loses more performance than most top drives, because it is one of the few TLC-based drives in that tier.

The average and 99th percentile latencies from the 970 EVO on the Light test are some of the best we've measured, still represent tiny improvements over Samsung's previous high-end SSDs.

The Samsung 970 EVO leads over other flash-based SSDs for both average read and write latency, but the differences are just a few microseconds and thus completely imperceptible.

The Samsung 970 EVO is the first drive to keep its 99th percentile write latency below 100µs whether or not the Light test is run on a full drive, showing that the 4-6GB static SLC cache is still quite useful even when the dynamically sized portion of the cache is unavailable. The 99th percentile read latency shows that Samsung has improved their full-drive QoS over the 960 EVO, but for the 500GB model in particular they still have room for improvement.

The energy usage of the Samsung 970 EVO is slightly higher than the PM981, putting the 970 EVO in last place for efficiency among flash-based SSDs. The energy usage of the 970 EVO doesn't show much variation between running the test on a full vs empty drive, despite the large performance differences between those scenarios.

Random Read Performance

Our first test of random read performance uses very short bursts of operations issued one at a time with no queuing. The drives are given enough idle time between bursts to yield an overall duty cycle of 20%, so thermal throttling is impossible. Each burst consists of a total of 32MB of 4kB random reads, from a 16GB span of the disk. The total data read is 1GB.

The burst random read performance of the Samsung 970 EVO is the best they've ever delivered from TLC NAND flash memory, but the Intel SSD 760p is a few percent faster still.

Our sustained random read performance is similar to the random read test from our 2015 test suite: queue depths from 1 to 32 are tested, and the average performance and power efficiency across QD1, QD2 and QD4 are reported as the primary scores. Each queue depth is tested for one minute or 32GB of data transferred, whichever is shorter. After each queue depth is tested, the drive is given up to one minute to cool off so that the higher queue depths are unlikely to be affected by accumulated heat build-up. The individual read operations are again 4kB, and cover a 64GB span of the drive.

On the longer random read test, the Samsung 970 EVO proves to be the fastest TLC-based drive, but Samsung's MLC-based drives offer up to 20% higher performance.

Power Efficiency in MB/s/W	Average Power in W

The Samsung 970 EVO and its OEM sibling PM981 have the worst power efficiency of any recent high-end SSD during the random read test. The 970 EVO is drawing over 2.5W while Samsung's previous generation high end drives averaged less than 2W for very similar performance.

Samsung 970 EVO 1TBIntel SSD 760p 512GBSamsung 950 PRO 512GBSamsung 960 EVO 1TBSamsung 960 EVO 250GBSamsung 960 PRO 1TBSamsung 960 PRO 2TBSamsung 960 PRO 512GBSamsung 970 EVO 500GBSamsung PM981 1TBSamsung PM981 512GBOCZ RD400 1TBOCZ RD400 512GBWestern Digital WD Black 1TB (3D NAND)Toshiba XG5 1TBSamsung 860 PRO 512GBCrucial MX500 1TBCrucial MX500 500GBIntel Optane SSD 900P 280GB

The performance scaling of the 970 EVO is almost identical to that of the 960 EVO, but the 970 EVO draws more power throughout the random read test.

Random Write Performance

Our test of random write burst performance is structured similarly to the random read burst test, but each burst is only 4MB and the total test length is 128MB. The 4kB random write operations are distributed over a 16GB span of the drive, and the operations are issued one at a time with no queuing.

The burst random write performance from the Samsung 970 EVO is disappointing compared to the PM981, especially for the 1TB 970 EVO. Meanwhile, recent Intel and WD drives have been raising the bar with very fast SLC write caches.

As with the sustained random read test, our sustained 4kB random write test runs for up to one minute or 32GB per queue depth, covering a 64GB span of the drive and giving the drive up to 1 minute of idle time between queue depths to allow for write caches to be flushed and for the drive to cool down.

On the longer random write test, the 1TB PM981 provided top-tier performance, but the 1TB 970 EVO is about 12% slower, putting it on par with the previous generation from Samsung. The 500GB 970 EVO is also slightly slower than its PM981 counterpart.

Power Efficiency in MB/s/W	Average Power in W

Power efficiency has also regressed for the 970 EVO on the random write test, leaving it well below the standard set by the WD Black and the slower but similarly efficient Toshiba XG5.

Samsung 970 EVO 1TBIntel SSD 760p 512GBSamsung 950 PRO 512GBSamsung 960 EVO 1TBSamsung 960 EVO 250GBSamsung 960 PRO 1TBSamsung 960 PRO 2TBSamsung 960 PRO 512GBSamsung 970 EVO 500GBSamsung PM981 1TBSamsung PM981 512GBOCZ RD400 1TBOCZ RD400 512GBWestern Digital WD Black 1TB (3D NAND)Toshiba XG5 1TBSamsung 860 PRO 512GBCrucial MX500 1TBCrucial MX500 500GBIntel Optane SSD 900P 280GB

The random write performance of the 1TB 970 EVO tops out at just over 1.5 GB/s at queue depths of 8 and higher. The 500GB 970 EVO starts running out of SLC cache and showing inconsistent performance past QD4. The 1TB PM981 was able to ramp up performance much faster than the 970 EVO and hit a maximum of about 1.8GB/s before running out of SLC cache near the end of the test. The 512GB PM981 behaved very similarly to the 500GB 970 EVO.

Sequential Read Performance

Our first test of sequential read performance uses short bursts of 128MB, issued as 128kB operations with no queuing. The test averages performance across eight bursts for a total of 1GB of data transferred from a drive containing 16GB of data. Between each burst the drive is given enough idle time to keep the overall duty cycle at 20%.

The Samsung PM981 set new records for burst sequential read performance, but the Samsung 970 EVO fails to live up to that standard. The 970 EVO is a substantial improvement over the 960 EVO, but doesn't manage to beat the last generation's fastest MLC drives.

Our test of sustained sequential reads uses queue depths from 1 to 32, with the performance and power scores computed as the average of QD1, QD2 and QD4. Each queue depth is tested for up to one minute or 32GB transferred, from a drive containing 64GB of data.

On the longer sequential read test, the Samsung 970 EVO performs far better than the Samsung PM981, indicating that Samsung has made significant firmware tweaks to improve how the drive handles the internal fragmentation left over from running the random I/O tests. The 970 EVO is the fastest TLC-based drive on this test, and the 1TB model even manages to beat the MLC-based 1TB 960 PRO.

Power Efficiency in MB/s/W	Average Power in W

The 1TB 970 EVO draws more power during this sequential read test than any other M.2 drive in this mix, but its performance is high enough to leave it with a good efficiency score. The 500GB 970 EVO ends up with below-average efficiency.

Samsung 970 EVO 1TBIntel SSD 760p 512GBSamsung 950 PRO 512GBSamsung 960 EVO 1TBSamsung 960 EVO 250GBSamsung 960 PRO 1TBSamsung 960 PRO 2TBSamsung 960 PRO 512GBSamsung 970 EVO 500GBSamsung PM981 1TBSamsung PM981 512GBOCZ RD400 1TBOCZ RD400 512GBWestern Digital WD Black 1TB (3D NAND)Toshiba XG5 1TBSamsung 860 PRO 512GBCrucial MX500 1TBCrucial MX500 500GBIntel Optane SSD 900P 280GB

Both capacities of the Samsung 970 EVO have very steady performance and power consumption across the duration of the sequential read test. This is in contrast to drives like the WD Black and Toshiba XG5 that don't reach full performance until the queue depths are rather high.

Sequential Write Performance

Our test of sequential write burst performance is structured identically to the sequential read burst performance test save for the direction of the data transfer. Each burst writes 128MB as 128kB operations issued at QD1, for a total of 1GB of data written to a drive containing 16GB of data.

The burst sequential write performance of the Samsung 970 EVO tops the charts, with the 500GB model almost reaching 2.5GB/s where the last generation of drives couldn't hit 2GB/s. The WD Black is only slightly behind the 970 EVO.

Our test of sustained sequential writes is structured identically to our sustained sequential read test, save for the direction of the data transfers. Queue depths range from 1 to 32 and each queue depth is tested for up to one minute or 32GB, followed by up to one minute of idle time for the drive to cool off and perform garbage collection. The test is confined to a 64GB span of the drive.

On the longer sequential write test, the 1TB 970 EVO takes a clear lead over everything else, even the 1TB PM981. The 500GB model is handicapped by its smaller capacity and smaller SLC cache, but still manages to be significantly faster than the 512GB PM981.

Power Efficiency in MB/s/W	Average Power in W

The 970 EVO and PM981 offer almost exactly the same power efficiency on the sequential write test. The 1TB model is slightly less efficient than the WD Black and 960 PRO, while the 500GB model is well behind the MLC-based drives of similar capacity.

Samsung 970 EVO 1TBIntel SSD 760p 512GBSamsung 950 PRO 512GBSamsung 960 EVO 1TBSamsung 960 EVO 250GBSamsung 960 PRO 1TBSamsung 960 PRO 2TBSamsung 960 PRO 512GBSamsung 970 EVO 500GBSamsung PM981 1TBSamsung PM981 512GBOCZ RD400 1TBOCZ RD400 512GBWestern Digital WD Black 1TB (3D NAND)Toshiba XG5 1TBSamsung 860 PRO 512GBCrucial MX500 1TBCrucial MX500 500GBIntel Optane SSD 900P 280GB

The 1TB 970 EVO starts off with a much higher QD1 performance on the sequential write test than the PM981 offers, and at higher queue depths it maintains a slight lead. At 500GB, the 970 EVO's performance oscillates as only some portions of the test are hitting the SLC cache.

Mixed Random Performance

Our test of mixed random reads and writes covers mixes varying from pure reads to pure writes at 10% increments. Each mix is tested for up to 1 minute or 32GB of data transferred. The test is conducted with a queue depth of 4, and is limited to a 64GB span of the drive. In between each mix, the drive is given idle time of up to one minute so that the overall duty cycle is 50%.

The Samsung 970 EVO is slightly slower than the OEM PM981 on the mixed random I/O test, but that still leaves the 1TB model very near the top of the chart, and the 500GB 970 EVO is only slightly behind the MLC-based 960 PRO.

Power Efficiency in MB/s/W	Average Power in W

The power efficiency of the Samsung 970 EVO trails the PM981 by a larger margin than performance alone did. The efficiency of the best MLC drives seems almost out of reach for TLC drives, except that the WD Black is in third place overall with 26% better efficiency than the 970 EVO.

Samsung 970 EVO 1TBIntel SSD 760p 512GBSamsung 950 PRO 512GBSamsung 960 EVO 1TBSamsung 960 EVO 250GBSamsung 960 PRO 1TBSamsung 960 PRO 2TBSamsung 960 PRO 512GBSamsung 970 EVO 500GBSamsung PM981 1TBSamsung PM981 512GBOCZ RD400 1TBOCZ RD400 512GBWestern Digital WD Black 1TB (3D NAND)Toshiba XG5 1TBSamsung 860 PRO 512GBCrucial MX500 1TBCrucial MX500 500GBIntel Optane SSD 900P 280GB

The Samsung 970 EVO's performance barely drops when writes are first added to the mix, and it grows at an accelerating rate through the rest of the test. The PM981 pulls ahead in the final phases with higher random write performance than the 970 EVO. The 960 EVO showed very flat performance until fairly late in the test, leaving it well behind the 970 EVO for overall performance despite offering similar performance at either end of the test.

Mixed Sequential Performance

Our test of mixed sequential reads and writes differs from the mixed random I/O test by performing 128kB sequential accesses rather than 4kB accesses at random locations, and the sequential test is conducted at queue depth 1. The range of mixes tested is the same, and the timing and limits on data transfers are also the same as above.

The Samsung 970 EVO sets new records on the mixed sequential I/O test, with the 1TB model beating the Intel Optane SSD and the WD Black. The 500GB model is significantly slower, but still performs well for its capacity. Both models are much faster than the PM981.

Power Efficiency in MB/s/W	Average Power in W

The 1TB Samsung 970 EVO is essentially tied for second place in power efficiency on the mixed sequential I/O test, but the first place WD Black has a large lead. The improved performance of the 970 EVO over the PM981 is match by improved efficiency, but in absolute terms the 970 EVO is drawing more power than almost any flash-based SSD on this test.

Samsung 970 EVO 1TBIntel SSD 760p 512GBSamsung 950 PRO 512GBSamsung 960 EVO 1TBSamsung 960 EVO 250GBSamsung 960 PRO 1TBSamsung 960 PRO 2TBSamsung 960 PRO 512GBSamsung 970 EVO 500GBSamsung PM981 1TBSamsung PM981 512GBOCZ RD400 1TBOCZ RD400 512GBWestern Digital WD Black 1TB (3D NAND)Toshiba XG5 1TBSamsung 860 PRO 512GBCrucial MX500 1TBCrucial MX500 500GBIntel Optane SSD 900P 280GB

The performance of the Samsung 970 EVO on the mixed sequential I/O test wobbles around with an unusual pattern that mirrors that of the PM981 but little else. The 1TB 970 EVO shows the typical spike in performance at the end when the workload shifts to fully cacheable writes, but the test fulls the SLC write cache on the 500GB model and prevents it from getting that boost.

Power Management Features

Real-world client storage workloads leave SSDs idle most of the time, so the active power measurements presented earlier in this review only account for a small part of what determines a drive's suitability for battery-powered use. Especially under light use, the power efficiency of a SSD is determined mostly be how well it can save power when idle.

For many NVMe SSDs, the closely related matter of thermal management can also be important. M.2 SSDs can concentrate a lot of power in a very small space. They may also be used in locations with high ambient temperatures and poor cooling, such as tucked under a GPU on a desktop motherboard, or in a poorly-ventilated notebook.

Samsung 970 EVO NVMe Power and Thermal Management Features
Controller	Samsung Phoenix
Firmware	1B2QEXE7
NVMe Version	Feature	Status
1.0	Number of operational (active) power states	3
1.1	Number of non-operational (idle) power states	2
	Autonomous Power State Transition (APST)	Supported
1.2	Warning Temperature	85°C
	Critical Temperature	85°C
1.3	Host Controlled Thermal Management	Supported
	Non-Operational Power State Permissive Mode	Not Supported

The Samsung 970 EVO bumps the supported NVMe spec version to 1.3, compared to the 1.2 feature set supported by the PM981 and 960 series. The 970 EVO implements the Host Controlled Thermal Management feature, allowing operating systems to configure the drive to throttle at a lower temperature than it normally would. The (optional) non-operational power state permissive mode feature is not included, so the 970 EVO is not supposed to do background tasks like garbage collection when it is in idle power states (unless they can be done within the power constraints of the idle states, which is unrealistic).

Samsung 970 EVO NVMe Power States
Controller	Samsung Phoenix
Firmware	1B2QEXE7
Power State	Maximum Power	Active/Idle	Entry Latency	Exit Latency
PS 0	6.2 W	Active	-	-
PS 1	4.3 W	Active	-	-
PS 2	2.1 W	Active	-	-
PS 3	0.04 W	Idle	0.21 ms	1.2 ms
PS 4	0.005 W	Idle	2 ms	8 ms

Note that the above tables reflect only the information provided by the drive to the OS. The power and latency numbers are often very conservative estimates, but they are what the OS uses to determine which idle states to use and how long to wait before dropping to a deeper idle state.

Idle Power Measurement

SATA SSDs are tested with SATA link power management disabled to measure their active idle power draw, and with it enabled for the deeper idle power consumption score and the idle wake-up latency test. Our testbed, like any ordinary desktop system, cannot trigger the deepest DevSleep idle state.

Idle power management for NVMe SSDs is far more complicated than for SATA SSDs. NVMe SSDs can support several different idle power states, and through the Autonomous Power State Transition (APST) feature the operating system can set a drive's policy for when to drop down to a lower power state. There is typically a tradeoff in that lower-power states take longer to enter and wake up from, so the choice about what power states to use may differ for desktop and notebooks.

We report two idle power measurements. Active idle is representative of a typical desktop, where none of the advanced PCIe link or NVMe power saving features are enabled and the drive is immediately ready to process new commands. The idle power consumption metric is measured with PCIe Active State Power Management L1.2 state enabled and NVMe APST enabled if supported.

Active idle power draw of the 970 EVO seems to be about 20% higher than the preceding generation of Samsung drives, but the low-power idle we measured is about the same as most other high-end NVMe drives.

The idle wake-up latency of the 970 EVO is more than twice that of its predecessors and also significantly higher than that of the Samsung PM981. This ~14ms latency exceeds the 8ms that the drive itself claims as its latency to wake up from its deepest sleep state.

Conclusion

The biggest difference between the Samsung 970 EVO and Samsung's earlier retail NVMe drives isn't anything to do with the drive's performance or power consumption: It's the fact that the 970 EVO has real competition in drives like the second-generation Western Digital WD Black SSD. This year, Samsung is faced not only with the task of improving over their already top-notch NVMe SSDs, but also with maintaining their lead over competitors who are rapidly catching up. Samsung's 3D NAND is no longer a unique competitive advantage, and we're finally seeing other NVMe controllers that can reach these performance levels when paired with suitable NAND.

The Samsung 970 EVO is set to compete against the latest top consumer SSDs from other brands. Unlike years past, Samsung won't always be on top of the performance charts, but we don't expect any upcoming products for this year to thoroughly outclass the 970 EVO. Most of the cases where the performance of the 970 EVO disappointed us are simply because we measured the PM981 OEM SSD as being a few percent faster. On the other hand, there are a few tests where the firmware tweaks in the 970 EVO have allowed it to be significantly faster than the PM981.

The Intel SSD 750 was the first to bring the large performance benefits of NVMe to the consumer market. It was soon eclipsed by the Samsung 950 PRO, which offered much better real-world performance thanks to better optimization for consumer workloads - the Intel SSD 750's enterprise roots were still quite apparent. When Samsung introduced the 960 PRO and 960 EVO generation, performance jumped again thanks in large part to their much improved second-generation NVMe controller. The 970 EVO brings another generation of new controllers and NAND flash, but huge performance jumps aren't as easy to come by. We're closing in on the limits of PCIe 3 x4 for sequential read speeds, and there's not much low-hanging fruit for optimization left in the NVMe controllers and how they manage flash memory. Samsung's 3D NAND is still increasing in density, but we're not seeing much improvement in performance or power efficiency from it.

That leaves Samsung having to make tradeoffs with the 970 EVO, sacrificing power efficiency in many places for slight performance gains. Since almost any consumer would find the 960 PRO and 960 EVO to already be plenty fast enough, this means the 970 EVO is not at all a compelling upgrade over its predecessors. It simply doesn't bring much new to the market. The new Western Digital WD Black isn't always as fast as the Samsung drives, but its great power efficiency is a unique advantage that distinguishes it in a high-end market segment that now has multiple viable contenders.

NVMe SSD Price Comparison
	240-256GB	400-512GB	960-1024GB	2TB
Samsung 970 PRO (shipping May 7)		$329.99 (64¢/GB)	$629.99 (62¢/GB)
Samsung 970 EVO (shipping May 7)	$119.99 (48¢/GB)	$229.99 (46¢/GB)	$449.99 (45¢/GB)	$849.99 (42¢/GB)
Samsung 960 PRO		$324.99 (63¢/GB)	$608.58 (59¢/GB)	$1299.90 (63¢/GB)
Samsung 960 EVO	$119.99 (48¢/GB)	$199.99 (40¢/GB)	$449.97 (45¢/GB)
WD Black aka SanDisk Extreme PRO	$119.99 (48¢/GB)	$229.99 (46¢/GB)	$449.99 (45¢/GB)
Intel SSD 760p	$99.99 (39¢/GB)	$215.99 (42¢/GB)	$399.99 (39¢/GB)
Plextor M9Pe	$119.99 (47¢/GB)	$209.19 (41¢/GB)	$408.26 (40¢/GB)
HP EX920		$199.99 (39¢/GB)	$369.99 (36¢/GB)
MyDigitalSSD SBX	$94.99 (37¢/GB)	$159.99 (31¢/GB)	$339.99 (33¢/GB)
Crucial MX500	$74.99 (30¢/GB)	$124.99 (25¢/GB)	$249.99 (25¢/GB)	$499.99 (25¢/GB)

With the performance of high-end consumer SSDs no longer growing by leaps and bounds, consumers should stop and consider whether a high-end drive is the right choice for them. The NVMe SSD market no longer consists of just Samsung's drives and everybody else's failed attempts at a high-end drive. There's a real low-end segment to the NVMe SSD market, with drives that are much closer in price to SATA SSDs but still offering much better performance. Since faster storage brings diminishing returns to overall real-world application performance, many consumers can find drives cheaper than the 970 EVO that are still fast enough, and drives like the 970 PRO are impossible to justify on the basis of performance alone.

There aren't a lot of 2TB NVMe options yet, and the 970 EVO for $849 sounds a lot better than the 960 PRO for $1299. For the other capacities, there are a lot of current-generation options to consider. At the moment, it looks like most of them are at or below the MSRPs for the 970 EVO, confirming that the 970 EVO could easily have been a flagship SSD if Samsung were not also launching the 970 PRO.

The Western Digital WD Black is currently matching the pricing on the 970 EVO, and it will be interesting to see if Western Digital drops the price at any point. They have the clear winner for mobile use, but the 970 EVO is usually faster. Moving down the price scale, NVMe drives start to come with caveats such as performance weak spots where they are no better than a SATA drive. Drives like the Intel 760p are still great performers overall, and even the low-end NVMe MyDigitalSSD SBX that's 20-30% cheaper than the 970 EVO is still a clear performance upgrade over any SATA drive. Today's market gives consumers a broad spectrum of NVMe options, with the Samsung 970 EVO well-positioned near the top.

Read our Best SSDs: Q2 2018 Guide Here