Performance Comparisons – Client QD Weighted
These results attempt to simplify things by focusing on what really matters – the Queue Depths that folks actually see when using these products. A dimension is eliminated from the previous charts by applying a weighted average to those results. The weights were derived from trace recordings of moderate to heavy workloads, which still ended up running closer to QD=1-2 even on a slower SATA SSD. The intent here is to distill the results into something for those wanting 'just the facts' to grab and go when making their purchasing decisions. Don't be alarmed by the low figures. Remember, these are low queue depths – the place where these SSDs actually operate when in use by those not just running benchmarks all day!
Note the tight grouping of random read performance here. At lower QD's, we are mostly waiting for the flash to respond to the read requests, and the majority of SSDs (NVMe or SATA) turn in similar figures for this type of workload. The NX500 is outpaced by many others on random writes, but many of those are benefiting from an SLC cache.
Sequential access is a different story entirely. We see a lot of variance across products here, but for the most part, the NX500 is among the top of the pack, with the exception of the Samsung 960 EVO.
For those curious how these results pan out in comparison to older / other SSDs, here are a couple of rather large charts for your viewing pleasure:
This kills my Intel 750 400
This kills my Intel 750 400 GB in sequential, which matters not at all for my typical workloads. About the same or worse in random IOPS. Probably feels exactly the same in daily use. $320 price is not bad – Intel launched theirs at $400 and it’s still the same price today (luckily I got mine for $300 during a rare sale at Newegg).
I’d love to see a direct comparison review, but I’m sure these will sell better – cheaper and better looking. I’ll keep my Intel drive, because their reliability is legendary, and it just feels like it will last forever.
You can get 3 sm951 for a
You can get 3 sm951 for a little more…
as nice as those are, if I
as nice as those are, if I had that money to spend on storage then i’d rather just get more cheap sata ssd(s), like a 1TB samsung 850evo for about ~ $340.
Funny, at the conclusion page
Funny, at the conclusion page I remembered you were reviewing the Corsair NX500, I was much more interested in the details of the new testing method. Excellent work Allyn!
Request: Can we get more
Request: Can we get more reviews of gaming headsets?
gaming headsets are almost
gaming headsets are almost never good though. Just buy a hyperx cloud or sennheiser game zero/one
I had an old OCZ Z-Drive R4
I had an old OCZ Z-Drive R4 SSD with a bunch of unpopulated capacitor pads on the PCB too. Do you think they designed in some kind of power smoothing / filter stage or something and then figured the cost of adding tantalum caps to the BOM outweighed any noticeable benefit to the user?
Of course, that would be for
Of course, that would be for power loss protection on the enterprise version of the card, now that I read what Al wrote instead of just looking at the pretty pictures. That brings up another topic I find crazy, the UPS. Convert AC to DC to store it in a battery, then back to AC to feed it into the computer’s PSU, where it is converted again to DC to run all the circuits. Can’t make it any more better. Computers are solved, guys.
Yeah, it makes more sense to
Yeah, it makes more sense to just have a single version of the PCB, and add components as applicable for the enterprise version, etc.
Wow Allyn, that performance
Wow Allyn, that performance comparison history is legendary! Pulled out every SSD you could dig up in the office? You need to make that model list searchable so people can find this. A recent SSD review comparing sooooo many models is a rare find!
I kinda treat SSDs like
I kinda treat SSDs like Pokemon :). We definitely want to do better things with the data, but with this site design, we're limited to pics of charts.