One Design, Two Products: The SanDisk Ultra 3D (1TB) and WD Blue 3D (1TB) SSD Reviews, with BiCS 3D NAND
by Billy Tallis on September 14, 2017 9:00 AM ESTMixed 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%.
Orange is for the new drives, Blue is for the previous generation models
The mixed random I/O performance of the WD Blue 3D NAND and the SanDisk Ultra 3D is significantly better than the previous generation, but the improvement is not enough to match any of the other 3D NAND SSDs. The Samsung, Intel and Micron drives with 3D NAND are all still much faster overall on this test.
The power efficiency of the WD/SanDisk drives has improved to a larger degree than performance, but is still slightly worse than Samsung's 850 EVO and 850 PRO, and is far lower than the Crucial MX300 and the Intel 545s.
 |
|||||||||
Both generations of WD/SanDisk drives show fairly flat performance across most of the mixed random I/O test, with only a modest spike at the end when the workload shifts to pure random writes. Most of the other drives either have steady performance increases across the test as the proportion of writes grows, or a much larger spike in performance at the very end.
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 mixed sequential I/O performance of the WD Blue 3D NAND and SanDisk Ultra 3D is comparable to the Intel and Micron 3D NAND drives, and only slightly behind Samsung's 850 PRO and 850 EVO.
The new WD/SanDisk drives have good power efficiency on the mixed sequential I/O test, but the Intel/Micron 3D NAND SSDs are still slightly better, and the OCZ VX500 is still on top.
 |
|||||||||
The performance of the WD Blue 3D NAND and SanDisk Ultra 3D across the mixed sequential I/O test is mostly flat, with the occasional blip. The Samsung drives achieve their standout performance by maintaining better performance through the read-heavy first half of the test, and only experiencing a relatively short and shallow drop in performance during the second half of the test. The Crucial MX300 achieves a very similar overall average to the WD/SanDisk drives but with relatively lower performance during the read-heavy portions of the test and clearly higher performance on the write-heavy half.
Facebook
Twitter
RSS
52 Comments
View All Comments
MrSpadge - Thursday, September 14, 2017 - link
The production cost is pretty much the same, be it SATA or PCIe. So "SATA being tapped" doesn't help price at all, except for the fact that manufacturers can't bill you for extra performance. But that was always the case with the slower SSDs.nathanddrews - Friday, September 15, 2017 - link
I know it's not realistic, hence "it would just be nice if".CheapSushi - Sunday, September 17, 2017 - link
V-NAND QLC will make that happen. I think for bulk storage, QLC SATA drives will be perfect for that duty and will decrease price per GB.Magichands8 - Thursday, September 14, 2017 - link
Unfortunately, it's still about 3 times more expensive than it should be for it to be viable. Still wouldn't buy either as they're both crippled by the SATA interface but hey, at least they got the form factor right by offering them in 2.5".MajGenRelativity - Thursday, September 14, 2017 - link
I'm not aware of any SSDs that are 1/3 the price, and there certainly aren't any that are 1/3 the price and have competitive performance. The SATA interface will not be going away for a while, and most people don't need the performance afforded by PCIeDanNeely - Thursday, September 14, 2017 - link
I assume he's sulking because it's still about 5-6x as expensive as spinning rust. ($50 for the 1TB blue at 5400 RPM on amazon). I haven't seen any more recent projections but as of a a year ago the crossover in price per TB was predicted to occur in the mid 2020's; so we've still got a way to go.MajGenRelativity - Thursday, September 14, 2017 - link
Fair point, but SSDs are still viable without a price drop, mass-market adoption is what requires the price dropMagichands8 - Friday, September 15, 2017 - link
Oh I don't mind paying a premium for the SSD tech but I do mind the ridiculously inflated prices and performance bottlenecks that we've had to put up with for years and years. From the other posts here it's obvious that there are a lot of people comfortable with that though and willing if not eager to pay very high prices for low capacity and low performance drives even while manufacturers have had years to differentiate their products. Even when said people must know of the supply shortages and the impending lower prices only a matter of months away. Like I've said before, drives like these might be real last ditch options for people in a crunch who absolutely need a replacement drive immediately or perhaps some other niche reason. But otherwise it just doesn't make much sense.CheapSushi - Sunday, September 17, 2017 - link
Are you saying this because you want to have ONE drive in your system to function as a performance panacea? I can see why someone would advocate for that particular setup if JUST a gamer with a mini-ITX system. But with ATX systems, there's nothing wrong with multiple drives; fom NVMe Optane, to NVMe PCIe to AHCI SATA, each have a place.Magichands8 - Sunday, September 17, 2017 - link
EVERYONE should advocate for that setup. You're obviously very accustomed to think it natural for someone to have 3 or 4 different kinds of storage to achieve their goal(s). Are you telling me that if I offered you a single drive and interface that satisfied all of those rolls you would reject it? Are you actually advocating that computer users should be FORCED to compromise at every step of the way when they use their system?