new Sabrent Rocket Q not performing as expected in all testing

[Ricky T]

I downloaded crystaldiskmark and anvils storage utilities to test out my drive and it seems lacking compared to reviews. My scores are on the right and you can see in some tests I am not even close. I have looked at several reviews and they are all pretty consistent and its just some of my scores that are way off. And no it is not running hot as it is 29 C at idle and never even hit 50 C under testing load.

 

[LFaWolf]

I think it might be because you have it as the boot drive and Windows could be reading and writing data to the drive while the tests were run. Your scores are better than my scores and I have the Samsung 970 EVO Plus 2TB also running as boot drive and a Corsair MP600

 

[Ricky T]

Wow did not even think of that but I did look at disc usage and had every single thing closed before running these tests and I consistently get the low scores.

 

[LFaWolf]

Even with all the programs closed, Windows still has a ton of processes running in the background and could be reading/writing to the Windows (C:\ drive) I don't know what is your usage pattern but if I get an 8TB NVMe drive I would simply use for fast storage, and get a fast 2TB drive for boot. I mean, how many programs are you going to install? Use the C:\ drive for programs/install and use the larger drive for storage.

 

[noko]

Turn off virus protection, see if that changes the values.

 

[Ricky T]

That makes no difference.

 

[kirbyrj]

You did notice like LFaWolf says that in the reviews it is the "E:" drive while in your system it is the "C:" drive right? You're still going to have some overhead from using it as a boot drive vs. just an additional disk in the system.

 

[Ricky T]

Even with all the programs closed, Windows still has a ton of processes running in the background and could be reading/writing to the Windows (C:\ drive) I don't know what is your usage pattern but if I get an 8TB NVMe drive I would simply use for fast storage, and get a fast 2TB drive for boot. I mean, how many programs are you going to install? Use the C:\ drive for programs/install and use the larger drive for storage.

You did notice like LFaWolf says that in the reviews it is the "E:" drive while in your system it is the "C:" drive right? You're still going to have some overhead from using it as a boot drive vs. just an additional disk in the system.

Just so you know it made no difference using it as my main drive as it did as a secondary drive just like a I figured as again nothing else was running. In fact no matter what drive I test in this system it comes up well under what it should be getting in several tests. That goes for nvme or normal sata drives like my Samsung 850 or 860.

 

[sinisterDei]

OK, there's two things.

First as others have mentioned you are booting off your drive, where the benchmarks you are comparing to the drive is not being used as the boot drive.

As a comparison point, when I benchmarked the Rocket Q 8TB drive, I got results more similar to yours when testing mine as a boot drive:

Second, your drive is different than mine or the other benchmark you have chosen to use to compare it with.

Take a look at the drive names in Anvil - mine and your other benchmark comparison point list a Sabrent Rocket Q 7681GB where yours shows up as Rocket Q 8001GB. This is not just a naming difference either, look at the actual amount of disk space that shows up in Windows - my drive has 7154GB while yours shows 7451. Your variant of the drive does not have the 320GB of reserved NAND area that mine does. This results in your drive having all 8TB of flash space actually available for user consumption, while my drive is in reality a 7.68TB drive. On the other hand, my drive has 320GB of reserved NAND to use for spare area and pseudo-SLC cache that might have an impact in some performance situations (like when the drive is full). The point is, apples are being compared with oranges, despite the drives supposedly being "the same"

 

[Maxx]

7154GiB vs 7451GiB is indeed just a matter of overprovisioning. Many people complained about having "less space than advertised" with the 8TB Rocket Q - just check Amazon reviews - but it's not an unusual amount of OP. 7154GiB matches up with the typical 960GB per TiB (960 * 8 = 7680 GB = 7153GiB). 7451GiB matches up with 1000GB per TiB (1000 * 8 = 8000 GB = 7451GiB). One could in fact also have 1024GB per TiB and even more space available to the user. It doesn't impact SLC on this drive because it's full-drive dynamic SLC caching, although it could impact write performance in some situations due to less OP (as mentioned, being fuller is one such situation). In any case, Sabrent changed it to 8000GB due to complaints, or so I imagine.

(my reply with regards to how reviews show the former capacity vs. this one's)

 

[sinisterDei]

it's not an unusual amount of OP. 7154GiB matches up with the typical 960GB per TiB

While that is true on a technical level, it is not true on a marketing level. 960GB per TiB is 100% the norm for many enterprise class drives, but exactly *none* of those drives market themselves as 1000GB drives. Sabrent's sin was in the marketing department - their "8TB" drive was a 7.68TB drive. Now they have apparently released a new revision that is actually 8TB.

The point of my response, though, is that almost all reviews online - mine included - reference the original 7.68TB variant of the drive. And while it is possible the only thing that changed is the reserved space for overprovisioning, it is equally possible that other aspects of the drive were adjusted as well, and that they might result in different performance characteristics. Hence the potential that OP is comparing apples and oranges.

 

[Maxx]

Which is what I said. Jesus, why do I even post here anymore.

And actually I've spoken to multiple reviewers about this drive (at least two who had to return it for a "good" sample) and users (one who had to return it) and there were some teething issues. That said you can check the firmware revision.

 

[Maxx]

some of my scores that are way off.

Test in Safe Mode.