The ADATA XPG GAMMIX S10 (512GB) SSD Review: Entry-Level NVMe With Style
by Billy Tallis on October 25, 2017 8:30 AM ESTSequential 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%.
While the ADATA XPG GAMMIX S10 offers good burst random read performance, its burst sequential read performance is very close to last place among this collection of drives. Most mainstream SATA SSDs are quicker on this test, and the Samsung 960 EVO is more than seven times faster.
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 GAMMIX S10 comes in last place, tied with the WD Black and just a hair behind the Crucial MX300 and Intel 600p. The PCIe interface isn't helping at all here; Micron's 3D NAND is just too slow.
The power efficiency of the GAMMIX S10 is also in last place alongside the Intel 600p, and on this metric the WD Black has a small advantage. The other TLC-based NVMe SSDs deliver more than four times the performance per Watt.
 |
|||||||||
The sequential read performance of the GAMMIX S10 doesn't saturate until QD4, and that top speed is unimpressive even by SATA standards. With power consumption only reaching about 4W, thermal throttling is unlikely to be the problem. Instead, this is mostly a limitation of the flash itself, though the controller isn't making the situation better.
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 speed of the ADATA XPG GAMMIX S10 is much higher than the Intel 600p and more than twice as fast as a SATA SSD. The GAMMIX S10 still falls on the low end by NVMe standards.
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 GAMMIX S10's performance falls apart. It is barely faster than the last-place Intel 600p, and can't keep up with the typical mainstream SATA SSD.
The GAMMIX S10's power efficiency during the sustained sequential write test is bad, but at least it isn't alone at the bottom of the chart: both the Intel 600p and 750 rank slightly below the GAMMIX S10, and the WD Black's efficiency is comparable. Among the rest of the field, there isn't a clear separation between NVMe and SATA SSDs when it comes to efficiency; the fast NVMe drives pay for that performance with higher power consumption.
 |
|||||||||
The sequential write performance of the ADATA XPG GAMMIX S10 doesn't increase with higher queue depths, and instead the performance drops and varies a bit as the SLC cache is filled and background garbage collection adds some overhead.
Facebook
Twitter
RSS
27 Comments
View All Comments
rrinker - Wednesday, October 25, 2017 - link
For some reason I thought the black insert was plastic, but it actually doesn't say either way. If plastic, well... You pretty much said what I was skirting around, it's all cosmetic like so many of the weird and funky heat sinks/heat spreaders you see on "enthusiast" gear. There to look cool, doesn't actually do much of anything, either because it's not really needed, the device with the heat sink doesn't actually get all that hot, or it's no more effective than a plain old simple heatsink.Wwhat - Sunday, November 12, 2017 - link
I am absolutely perplexed that anandtech doesn't specify the material of that insert. I mean that is so goddamn relevant.And I'm actually wondering what adata's comment on that insert would be. They should have asked them.
As for needed, all tests have shown a good heatsink on the M2 SSD's does indeed do a world of good. And tests have also shown that half the companies in the world that supply such, either bundled or third party, can't figure out such a simple thing as a piece of aluminium on a freaking small rectangle. I mean how complex is it? But again it's been shown that if you put a good one on it it's always noticeably beneficial.
jjj - Wednesday, October 25, 2017 - link
Don't understand how one imagines a product with that seq read perf.trparky - Wednesday, October 25, 2017 - link
I went to go look at what you were talking about and my first reaction when I saw the graphs was "Ouch! WTF!". It barely matches the performance of a SATA SSD, putting it on NVMe is just a waste.r3loaded - Wednesday, October 25, 2017 - link
If your product is going to get curbstomped in performance by the 960 Evo, at least try to price it at a significant discount vs the 960 Evo.ddriver - Wednesday, October 25, 2017 - link
It is very hard to undersell samsung, they make the memory, they make the controller, they make the pcb. Vertical integration.Now imagine that you have to buy each of those components from someone else, pay profit margins on the components, then put work into making a product out of it, and sell that product with a profit margin for yourself.
It is a pointless endeavor. In a better world, they won't even bother to compete with a vertically integrated company like samsung. But in the world as it is, there are plenty of idiots who will buy this, one way or the other, either for the gimmicky but nice looking heatsink, or it will get it shoved down their throats by an OEM who has a deal with the manufacturer to bundle the product.
The negative aspect of this is that after so many years of domination, samsung will significantly cut on the purchase value of their products, because it still has tremendous lead and could let some of it melt in order to materialize as profits. Upcoming SSDs from samsung, even if a tad faster, will be disappointing compared to what we used to get in the past. Get ready for a slump of TLC and QLC, endurance and warranty period cuts.
Ratman6161 - Wednesday, October 25, 2017 - link
People aren't idiots. They are just uninformed. There is a big difference. For my uninformed friends who either can't or don't want to do the research, my recommendation is always "just buy Samsung and don't worry about it". Or if they want reasonably priced but larger storage with "OK" performance then it's "just buy Crucial and don't worry about it.If you know specifically what you are looking for and know what trade offs you are willing to make there are other choices that fit certain niches. But even knowing what I wanted/needed I ended up with a 512 GB 960 EVO and a 750 GB MX300 (got on prime day last year for $139 :)).
ddriver - Wednesday, October 25, 2017 - link
"People aren't idiots. They are just uninformed."Tomayto, tomahto... If you make a purchase uninformed, that's idiocy right there.
IMO everyone is an idiot, including me. It is not a matter of idiots or non-idiots, at this stage of evolution we are all idiots, it's just that some are more idiotic than others. And the first step to overcoming idiocy is to acknowledge it :) The gradation of idiocy is acknowledgement of it, ignorance of it, mistaking it for something else and finally taking pride in it.
Lolimaster - Wednesday, October 25, 2017 - link
For sata the WD/Sandisk 3D SSD's seems a like a better option than the Crucials.KAlmquist - Friday, October 27, 2017 - link
A vertically integrated company like Samsung still wants to make a profit on its investments into R&D, so the fact that all of the R&D is done by one company doesn't directly help costs. It does remove some of the costs associated with developing and maintaining relationships between companies. For example, Silicon Motion has to market its products to SDD manufacturers, which is an overhead that Samsung's controller development group doesn't have. Another possible cost savings for Samsung is that it knows it's controllers will never be used with anything other than Samsung flash, so it doesn't have to develop a controller that is compatible with multiple brands of flash.