Original Link: https://www.anandtech.com/show/11085/the-adata-ultimate-su800-ssd-review-128gb-256gb-512gb
The ADATA Ultimate SU800 SSD Review (128GB, 256GB, 512GB)
by Billy Tallis on February 1, 2017 12:01 PM EST
ADATA's Ultimate SU800 is their first SSD to use 3D NAND and the first 3D NAND SSD from a company that doesn't manufacture their own NAND flash memory. The SU800 pairs Micron's 3D TLC NAND with Silicon Motion's SM2258 controller to produce an entry-level consumer SATA SSD, available in both 2.5" and M.2 form factors.
We've previously reviewed the Silicon Motion SM2258 controller in the Intel SSD 540s, which uses SK Hynix 16nm TLC NAND. We also had the opportunity to test an SM2258 engineering sample/reference design from Silicon Motion that used the same Micron 3D TLC but earlier firmware and different overprovisioning settings. Silicon Motion's controllers are known for low cost and low power, and it's worth noting that their SATA SSD controllers are one of the few remaining single-core designs.
Micron's 3D TLC is the first 3D NAND to hit the open market in high volume and with reasonable cost. For the moment, Micron's 3D NAND is the only viable option for most SSD manufacturers looking to move beyond planar NAND. Samsung is keeping almost all of their 3D NAND for their own very successful SSDs. Toshiba and SK Hynix are technically shipping 3D NAND but in limited supply and the poor cost effectiveness have prevented any third parties from adopting it for commercial SSDs. Toshiba and SK Hynix each have at least one or two token models in the supply chain using their 3D NAND, but still no mass-market consumer or retail availability. (Toshiba and SK Hynix have seen more success in the smartphone and memory card markets, where performance requirements are lower and the large die sizes of Micron's 3D NAND are a problem.)
Micron's only consumer SSD with 3D NAND so far is the Crucial MX300, and Intel has released their SSD 600p, both using the Intel/Micron 32-layer 384Gb 3D TLC NAND flash. The Crucial MX300 has been quite successful, with acceptable mainstream performance, great pricing, and great power efficiency. The Intel 600p is the most affordable NVMe SSD on the market and it offers real-world performance that exceeds any SATA SSD. That leaves plenty of room in the market for other models with 3D NAND, and plenty of interest from consumers seeking the benefits of 3D NAND.
Micron's 32-layer 3D TLC NAND has a capacity of 384Gb (48GB) per die. This odd capacity means that while the SU800 has typical usable capacities like 128GB and 512GB, the die counts and raw flash capacities aren't powers of two. The 128GB SU800 has just three packages with one die each for a raw capacity of 144GiB. The 256GB and 512GB SU800s use three and six dual-die flash packages, respectively. Since the SM2258 controller has four channels for communicating with NAND, the smallest SU800 cannot make full use of the controller's capability and has reduced performance as a result. The larger capacity SU800 variants are rated to mostly saturate SATA on sequential transfers.
| ADATA Ultimate SU800 Specifications | ||||
| Capacity | 128 GB | 256 GB | 512 GB | 1 TB |
| Controller | Silicon Motion SM2258 | |||
| NAND Flash | Micron 32-layer 384Gb 3D TLC NAND | |||
| Form Factor | 2.5" 7mm or M.2 2280 | |||
| Sequential Read | 560 MB/s | |||
| Sequential Write | 300 MB/s | 520 MB/s | ||
| Random Read IOPS | Up to 90K IOPS | |||
| Random Write IOPS | Up to 80K IOPS | |||
| TCG Opal Encryption | No | |||
| Power Management | DevSleep | |||
| Warranty | 3 years | |||
| MTBF | 2,000,000 hours | |||
| Initial MSRP | $59.99 | $79.99 | $139.99 | $269.99 |
In this review, the SU800 will be compared against the following drives. Click on the names to see their subsequent reviews.
- Micron's Crucial MX300, using the same 3D TLC NAND but Marvell's 88SS1074 controller
- Intel SSD 540s, using the same Silicon Motion SM2258 controller but SK Hynix 16nm planar TLC NAND
- ADATA Premier SP550, using the earlier Silicon Motion SM2256 controller and SK Hynix 16nm planar TLC NAND
- PNY CS1311, one of many models using the Phison S11 controller and Toshiba 15nm planar TLC NAND
- Samsung 750 EVO, Samsung's return to planar TLC for the sake of cost-cutting and to avoid the limitations of low NAND die count on low capacity SSDs due to the high per-die capacity of 3D NAND
- ADATA XPG SX930 , an older mainstream SSD using Micron 16nm MLC and JMicron JMF670H controller
As always, other comparisons can be made using our Bench database.
| AnandTech 2015 SSD Test System | |
| CPU | Intel Core i7-4770K running at 3.5GHz (Turbo & EIST enabled, C-states disabled) |
| Motherboard | ASUS Z97 Pro (BIOS 2701) |
| Chipset | Intel Z97 |
| Memory | Corsair Vengeance DDR3-1866 2x8GB (9-10-9-27 2T) |
| Graphics | Intel HD Graphics 4600 |
| Desktop Resolution | 1920 x 1200 |
| OS | Windows 8.1 x64 |
- Thanks to Intel for the Core i7-4770K CPU
- Thanks to ASUS for the Z97 Deluxe motherboard
- Thanks to Corsair for the Vengeance 16GB DDR3-1866 DRAM kit, RM750 power supply, Carbide 200R case, and Hydro H60 CPU cooler
Performance Consistency
Our performance consistency test explores the extent to which a drive can reliably sustain performance during a long-duration random write test. Specifications for consumer drives typically list peak performance numbers only attainable in ideal conditions. The performance in a worst-case scenario can be drastically different as over the course of a long test drives can run out of spare area, have to start performing garbage collection, and sometimes even reach power or thermal limits.
In addition to an overall decline in performance, a long test can show patterns in how performance varies on shorter timescales. Some drives will exhibit very little variance in performance from second to second, while others will show massive drops in performance during each garbage collection cycle but otherwise maintain good performance, and others show constantly wide variance. If a drive periodically slows to hard drive levels of performance, it may feel slow to use even if its overall average performance is very high.
To maximally stress the drive's controller and force it to perform garbage collection and wear leveling, this test conducts 4kB random writes with a queue depth of 32. The drive is filled before the start of the test, and the test duration is one hour. Any spare area will be exhausted early in the test and by the end of the hour even the largest drives with the most overprovisioning will have reached a steady state. We use the last 400 seconds of the test to score the drive both on steady-state average writes per second and on its performance divided by the standard deviation.
The 512GB ADATA SU800 delivers lower steady-state random write performance than the 525GB Crucial MX300 despite the latter having less built-in overprovisioning. The 256GB SU800 isn't the slowest in its class, but the 128GB is in last place with barely half the performance of the 120GB Samsung 750 EVO. The Silicon Motion engineering sample with more overprovisioning outperformed the SU800 with twice the capacity.
The performance consistency scores from the SU800 are typical for a low-end SSD, but there's clearly room for improvement. Samsung's 750 EVO scores far higher than any of the other budget SSDs, and the Crucial MX300 has the next highest score.
 |
|||||||||
| Default | |||||||||
| 25% Over-Provisioning | |||||||||
For the first few seconds of the test, the SU800 delivers close to the advertised IOPS, then it drops into a fairly typical pattern of widely variable performance above 5k IOPS as it burns through the spare area. Once the SLC cache and spare area are full, the SU800 spends most of its time slowed down by garbage collection, with regular short bursts of higher performance.
 |
|||||||||
| Default | |||||||||
| 25% Over-Provisioning | |||||||||
The garbage collection cycles on the SU800 appear to last about a minute each, separated by several seconds of faster performance. Overall performance and variability are higher for the larger capacities.
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 and unlike our Iometer tests, 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.
We quantify performance on this test by reporting the drive's average data throughput, a few data points about its latency, and the total energy used by the drive over the course of the test.
The average data rates delivered by the ADATA SU800 are slower than most of its competition, but it is improved over the earlier engineering sample from Silicon Motion and the Intel 540s that paired the SM2258 controller with SK Hynix 16nm TLC.
Latency is a weak point for the SU800, with average service times that are higher than almost all of its competition. The 128GB SU800 suffers the most, with an average over 21ms.
Aside from the smallest 128GB capacity, the SU800 doesn't have too many operations that take more than 100ms to complete, but at the 10ms threshold all three capacities rank poorly against the competition.
With so much time spent bogged down by garbage collection, it is no surprise that the smaller capacities of the SU800 use somewhat more power than most of the competition. The 512GB SU800 compares favorably against the planar TLC competition, but doesn't come close to the efficiency of the Crucial MX300.
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.
The ADATA SU800 shows a huge drop in performance when the Heavy test is run starting with a full drive, but when starting with an empty drive the 512GB SU800 performs quite well for a budget SSD and the smaller capacities are only moderately behind their competition.
The average service time metric highlights the discrepancy between full drive and empty drive performance. The SU800s and the similar Silicon Motion engineering sample show the largest difference by far, followed by the Crucial MX300 and the Samsung 750 EVO.
Very few budget SSDs can keep latency below 10ms through almost all of the Heavy test, but the 512GB SU800 manages it when the test is run starting on an empty drive. When operating on a full drive or when considering the smaller capacities, high latency is far more common.
The power consumption of the 512GB SU800 trails slightly behind the Crucial MX300, but is still good for a budget TLC drive. The smaller capacities rank last due to taking longer to complete the test, and the energy usage is substantially higher when the test is run on a full drive.
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.
On the Light test, all three capacities of the ADATA SU800 perform quite well when the test is run on an empty drive. The performance penalty from running the test on a full drive is still huge, though the 512GB SU800 doesn't fall behind all of its competition.
The smaller capacities of the SU800 have trouble keeping latency under control when full, but otherwise the average service times are normal for budget SSDs.
The percentage of operations that require more than 10ms to complete shows us that the high average service times when full are due to a very high number of outliers, rather than just a uniform slowing of operations.
Thanks to their good performance, the SU800s use less energy than most of the competition over the course of the light test, though the slightly slower Silicon Motion engineering sample with more overprovisioning and earlier firmware was a bit more efficient.
Random Read Performance
The random read test requests 4kB blocks and tests queue depths ranging from 1 to 32. The queue depth is doubled every three minutes, for a total test duration of 18 minutes. The test spans the entire drive, which is filled before the test starts. The primary score we report is an average of performances at queue depths 1, 2 and 4, as client usage typically consists mostly of low queue depth operations.
Random read performance on the SU800 is typical for budget SSDs, with only the Samsung 750 EVO standing out from the crowd.
Power consumption during this test is a bit higher than the competition, making the SU800 a relatively inefficient drive, though it's not a severe problem.
 |
|||||||||
With only three flash chips, it's no surprise that the 128GB SU800 shows very little performance increase from higher queue depths in the second half of the test. The larger capacities show more normal scaling behavior.
Random Write Performance
The random write test writes 4kB blocks and tests queue depths ranging from 1 to 32. The queue depth is doubled every three minutes, for a total test duration of 18 minutes. The test is limited to a 16GB portion of the drive, and the drive is empty save for the 16GB test file. The primary score we report is an average of performances at queue depths 1, 2 and 4, as client usage typically consists mostly of low queue depth operations.
As with the long-running random write consistency test, the Samsung and Crucial drives show the best random write speeds, but the SU800 outperforms most of the planar TLC competition.
Unlike the random read test, the SU800's power consumption on the random write test is relatively good, though its efficiency is still well behind the Crucial MX300.
 |
|||||||||
The graphs of performance over the course of this test clearly show performance of the smaller two capacities dropping as the SLC cache is filled, and flat performance thereafter. The 512GB SU800's performance in the second half of the test is quite variable due to the garbage collection cycles but on average is decent and certainly better than the planar TLC competitors.
Sequential Read Performance
The sequential read test requests 128kB blocks and tests queue depths ranging from 1 to 32. The queue depth is doubled every three minutes, for a total test duration of 18 minutes. The test spans the entire drive, and the drive is filled before the test begins. The primary score we report is an average of performances at queue depths 1, 2 and 4, as client usage typically consists mostly of low queue depth operations.
The smallest capacity of the ADATA SU800 has the slowest sequential read speed of all the drives in this comparison, as it only has flash connected to three of the controller's four channels. The larger SU800s saturate the SATA connection.
The SU800's power consumption is middle of the road for TLC drives, giving the larger two capacities slightly better than average efficiency. The ADATA XPG SX930 demonstrates that MLC NAND flash still has a significant efficiency advantage.
 |
|||||||||
The larger two capacities of the SU800 show modest performance improvement across the first half of the test, while the 128GB model exhibits almost no scaling.
Sequential Write Performance
The sequential write test writes 128kB blocks and tests queue depths ranging from 1 to 32. The queue depth is doubled every three minutes, for a total test duration of 18 minutes. The test spans the entire drive, and the drive is filled before the test begins. The primary score we report is an average of performances at queue depths 1, 2 and 4, as client usage typically consists mostly of low queue depth operations.
The sustained sequential write speeds of the SU800 are quite poor, but this is no surprise after seeing how much a full drive hurt performance on the ATSB Heavy and Light tests.
Power consumption is at least better than average, so the SU800 isn't too far behind the average for efficiency.
 |
|||||||||
Aside from the initial effects of filling up the SLC cache, performance and power consumption during sequential writes don't scale with queue depth on the SU800.
Mixed Random Read/Write Performance
The mixed random I/O benchmark starts with a pure read test and gradually increases the proportion of writes, finishing with pure writes. The queue depth is 3 for the entire test and each subtest lasts for 3 minutes, for a total test duration of 18 minutes. As with the pure random write test, this test is restricted to a 16GB span of the drive, which is empty save for the 16GB test file.
All three capacities of the ADATA SU800 deliver good performance for their class on this test. The Crucial MX300 and the earlier Silicon Motion engineering sample offer slightly higher performance.
Power consumption is a little higher than average, but given the good performance, the SU800's efficiency is not bad.
 |
|||||||||
The 128GB SU800 runs out of SLC cache before the end of this test, leading to poor performance during the final phase of the test where the workload is 100% writes. The larger SU800s show increasing performance over the course of the test as more writes can be cached and combined.
Mixed Sequential Read/Write Performance
The mixed sequential access test covers the entire span of the drive and uses a queue depth of one. It starts with a pure read test and gradually increases the proportion of writes, finishing with pure writes. Each subtest lasts for 3 minutes, for a total test duration of 18 minutes. The drive is filled before the test starts.
Capacity is a big factor in performance on this test. The 512GB SU800 performs quite well for a budget drive, while the 128GB is clearly slower than its competitors.
The power consumption of the SU800 on this test is only slightly higher than average.
 |
|||||||||
For the SU800, performance over the course of this test is largely a question of how soon the SLC cache fills up. The 128GB runs out very early while the 512GB doesn't bottom out until the second half of the test.
ATTO
ATTO's Disk Benchmark is a quick and easy freeware tool to measure drive performance across various transfer sizes.
 |
|||||||||
The inferior write speed of the smallest SU800 is quite clear from the ATTO plots, but there are no other major issues revealed by this test.
AS-SSD
AS-SSD is another quick and free benchmark tool. It uses incompressible data for all of its tests, making it an easy way to keep an eye on which drives are relying on transparent data compression. The short duration of the test makes it a decent indicator of peak drive performance.
All three capacities of the SU800 deliver great sequential read speeds. The low write speed of the 128GB SU800 stands out even on this short test, but the larger capacities are only slightly behind their competition.
Idle Power Consumption
Since the ATSB tests based on real-world usage cut idle times short to 25ms, their power consumption scores paint an inaccurate picture of the relative suitability of drives for mobile use. During real-world client use, a solid state drive will spend far more time idle than actively processing commands. Our testbed doesn't support the deepest DevSlp power saving mode that SATA drives can implement, but we can measure the power usage in the intermediate slumber state where both the host and device ends of the SATA link enter a low-power state and the drive is free to engage its internal power savings measures.
We also report the drive's idle power consumption while the SATA link is active and not in any power saving state. Drives are required to be able to wake from the slumber state in under 10 milliseconds, but that still leaves plenty of room for them to add latency to a burst of I/O. Because of this, many desktops default to either not using SATA Aggressive Link Power Management (ALPM) at all or to only enable it partially without making use of the device-initiated power management (DIPM) capability. Additionally, SATA Hot-Swap is incompatible with the use of DIPM, so our SSD testbed usually has DIPM turned off during performance testing.
The ADATA SU800 achieves respectably low idle power when link power management is used. The active idle power consumption is a problem for all of the drives with Silicon Motion controllers, and worst for the SU800.
Final Words
Judged against other entry-level SATA SSD, the ADATA Ultimate SU800 does not offer many significant performance improvements from its use of Micron 3D TLC NAND. Even if Micron's 3D TLC were substantially faster than the planar TLC it is competing against, the capacity of 384Gb per die compared to 128Gb for the planar TLC means that the SU800 has to get by with far fewer NAND chips to stripe accesses across. This parallelism is very important to achieving high performance, but the smallest 128GB SU800 has only three NAND flash chips to work with.
As total NAND capacity per die has increased, the page and erase block sizes have increased also. This is likely a major contributor to garbage collection having a much larger performance impact on the SU800 and Crucial MX300 than earlier TLC drives. Silicon Motion's controller and firmware in the ADATA SU800 don't seem to have adjusted to this quite as well as Micron's firmware has for the Marvell controller in the Crucial MX300. Despite having more spare area due to offering less usable capacity from the same amount of NAND, the 512GB ADATA SU800 seldom outperforms the 525GB MX300.
With the Crucial MX300, Micron chose to offer capacities of 275GB and up, retiring the 128GB capacity class. They also chose not to sample us the 275GB MX300, likely because of its lower performance than the larger capacities. ADATA is emphasizing those smaller capacities with the SU800 and taking on the challenge of offering decent performance with less parallelism available.
For light workloads where SLC caching is highly effective, ADATA has succeeded in roughly matching the performance of the last generation of planar TLC budget SSDs. This has come at the apparent expense of performance on heavier workloads and when working with a full drive. The 512GB SU800 is the smallest capacity that performs well on our ATSB Heavy test, and at all capacities it is important to not let the SU800 fill up or operate without TRIM being used.
Back when most of the SSD market was still using MLC NAND, Silicon Motion established a reputation for offering one of the most power efficient platforms. This advantage has been reduced with the transition to TLC for mainstream drives, and with the SU800 it now seems to be completely gone. In terms of power consumption, the SU800 and every other budget and mainstream SSD are still overshadowed by the Crucial MX300's remarkable efficiency.
| 120-128GB | 240-275GB | 480-525GB | 960-1050GB | |
| ADATA SU800 | $52.99 (41¢/GB) | $81.99 (32¢/GB) | $147.07 (29¢/GB) | $269.99 (26¢/GB) |
| ADATA SP550 | $48.99 (41¢/GB) | $71.97 (30¢/GB) | $134.99 (28¢/GB) | $299.99 (31¢/GB) |
| PNY CS1311 | $49.99 (42¢/GB) | $59.99 (25¢/GB) | $129.99 (27¢/GB) | $269.99 (28¢/GB) |
| Samsung 750 EVO | $85.98 (72¢/GB) | $129.95 (52¢/GB) | $139.99 (28¢/GB) | |
| Samsung 850 EVO | $98.00 (39¢/GB) | $169.99 (34¢/GB) | $319.99 (32¢/GB) | |
| Crucial MX300 | $89.99 (33¢/GB) | $146.42 (28¢/GB) | $259.99 (25¢/GB) |
The ADATA SU800 is priced as an entry-level SSD, but the entire market is heavily affected by an ongoing NAND shortage. There are a few older planar TLC SSDs that are still able to beat the SU800's prices by a few dollars, when they're in stock. But the bigger problem for the SU800 is that ADATA can't reliably beat Micron's pricing on the Crucial MX300. When taking into account the slightly higher usable capacity and better performance and efficiency, the MX300 is a better deal than the equivalent SU800.
That leaves the 128GB SU800 as the only member of the lineup that might make sense to buy at the moment. It has far fewer competitors as 120GB SSDs are disappearing from the market. With the caveat that the 128GB SU800 should only be used in scenarios where it is definitely larger than necessary and will be presented with light workloads, the SU800 is a fine alternative and a reasonable purchase if it's roughly tied for being the cheapest SSD in that capacity class.
Related SATA SSD Reading:
Facebook
Twitter
RSS
