Netac NV7000-T 2 TB

[W1zzard]

The Netac NV7000-T is available at unbelievable pricing of $80 for 2 TB. It still is one of the fastest SSD's we've ever tested, matching WD Black SN850, Samsung 980 Pro and Phison E18 drives. Netac even includes an optional heatsink with their solid-state-drive.

Show full review

 

[rlifeh]

The new legend nvme !

 

[HairyLobsters]

Don't know much about YMTC nand, how good is it?

 

[hurakura]

Have to be crazy to buy nvme with chinese controller and chinese nand. All their tech is stolen and all they can do is come up with half-baked product. Saving few dollars is not worth the headaches you gonna get from a product like that.

 

[b1k3rdude]

No bad, its up there with the Sabrent drive for slightly less money -

- https://www.amazon.co.uk/stores/Netac/NV7000-t_PCIe40SSD/page/BE6A6E1E-CF76-41EA-A484-77FCB6B6B835

But as captain paranoid pointed out its from China, so for the few extra ££ I would rather go with a known brand.

 

[R-T-B]

Have to be crazy to buy nvme with chinese controller and chinese nand. All their tech is stolen and all they can do is come up with half-baked product. Saving few dollars is not worth the headaches you gonna get from a product like that.

Half baked product that is up there with the best, doesn't really compute.

Still, there are valid reasons to avoid China but they aren't really up for discussion here. We talk about technical merits.

 

[cchi]

What does the green lines in the graph mean on page 8 (https://www.techpowerup.com/review/netac-nv7000-t-2-tb/8.html)?
Also why in the same graphs the read performance is around 4GB/s instead of closer to 7GB/s?

 

[cnlutong]

In fact, the 2TB PCIE 4.0 SSD, which combines Yangtze Memory Technologies' 232-layer TLC particles with Maxio's MAP1602 controller scheme, has been available in the Chinese market for some time now, with prices as low as 600 RMB (around 75 Euros). This package not only delivers top-tier PCIE 4.0 performance but also offers a warranty of 3600TBW. Many brands offer products based on this scheme. In Chinese testing articles and videos, many have extensively and thoroughly tested this SSD and have fully recognized its excellent performance, good quality, and affordable price.

And a 2TB PCIE 4.0 SSD that uses a similar scheme with 128-layer particles usually sells for 500 RMB (about 63 Euros).

I'm mentioning all this because I'm surprised that TechPowerUp, an excellent tech website, seems to be completely unaware of this. You might need some editors who are more attuned to Chinese language content and the Chinese market.

 

[bonehead123]

The new legend nvme !

"You're a legend in your own mind" - Dirty Harry

But as captain paranoid pointed out its from China, so for the few extra ££ I would rather go with a known brand.

^^THIS^^

Me too, I'd much rather give my $$ to a company I know I can & have trusted for over 8 years to make solid, reliable, backdoor-free products....like WD, Crucial, or Sammy

 

[Deleted member 229121]

It'll be fiiine. I bet it even comes with software that totally won't spy on you.

 

[W1zzard]

What does the green lines in the graph mean

The drive has two thermal sensors, sorry for not mentioning that

Also why in the same graphs the read performance is around 4GB/s instead of closer to 7GB/s?

Because I'm running QD1 sequential

It'll be fiiine. I bet it even comes with software that totally won't spy on you.

There is no software, it just works

 

[Tahagomizer]

The price seems to have jumped to 95USD. Personally I'd rather buy something much slower (like Kingston NV2) or slightly more expensive but from a more reputable manufacturer with higher chance of getting proper warranty. Either way, it's always good to see storage prices go down.

 

[hurakura]

Half baked product that is up there with the best, doesn't really compute.

Still, there are valid reasons to avoid China but they aren't really up for discussion here. We talk about technical merits.

Buy two then and store your most important data on them. I bet these will start having some firmware issues after a month or two.

 

[R-T-B]

I don't need to buy them to discuss their theoretical quality or merits. w1zard did that for us.

 

[Wirko]

@W1zzard

There's something weird going on with your synthetic random read tests, and has been going on consistently since they were introduced (the BX300 in 2017 was the earliest I found). I'm not saying the results are wrong but an explanation would be very welcome.

The 4KB QD1 random read IOPS numbers can't possibly be that high. For the Netac NV7000-T, it's ~65000 IOPS in the graph, which computes to 260 MB/s. (And for the old BX300, it's ~38000 IOPS, or 150 MB/s). No other site gets even half as much in their reviews; your own result in CDM is 89.42 MB/s.

Actually, no flash-based SSD reaches more than about 110 MB/s, with the possible exception of some exotic stuff like the Samsung Z-NAND SLC. Even the Phison representative here at TPU didn't mention faster speeds when asked specifically about that.

Random write IOPS seem more consistent with what CDM says, as well as with other reviews.

What software do you use to measure IOPS? What are the parameters, such as test area size?

 

[N/A]

The price seems to have jumped to 95USD. Personally I'd rather buy something much slower (like Kingston NV2)

VAT inclusive, 77,7666, so you have 23% VAT, but hey there is a 5 discount if you buy 2 or a 4GB

 

[OneMoar]

yea call me in 6 months when the swap to garbage nand and no-name controlers after good inital reviews

@W1zzard

There's something weird going on with your synthetic random read tests, and has been going on consistently since they were introduced (the BX300 in 2017 was the earliest I found). I'm not saying the results are wrong but an explanation would be very welcome.

The 4KB QD1 random read IOPS numbers can't possibly be that high. For the Netac NV7000-T, it's ~65000 IOPS in the graph, which computes to 260 MB/s. (And for the old BX300, it's ~38000 IOPS, or 150 MB/s). No other site gets even half as much in their reviews; your own result in CDM is 89.42 MB/s.

Actually, no flash-based SSD reaches more than about 110 MB/s, with the possible exception of some exotic stuff like the Samsung Z-NAND SLC. Even the Phison representative here at TPU didn't mention faster speeds when asked specifically about that.

Random write IOPS seem more consistent with what CDM says, as well as with other reviews.

What software do you use to measure IOPS? What are the parameters, such as test area size?

what 65000 isn't even that THAT fast we where getting 7-12 on sata drives 3 years ago

Enterprise SATA SSDs: Can Budget 2020 beat Top Line 2017?

www.anandtech.com

if you are seeing low results its because the person running the bench is an idiot and doesn't know what they are doing

 

[W1zzard]

@W1zzard

There's something weird going on with your synthetic random read tests, and has been going on consistently since they were introduced (the BX300 in 2017 was the earliest I found). I'm not saying the results are wrong but an explanation would be very welcome.

The 4KB QD1 random read IOPS numbers can't possibly be that high. For the Netac NV7000-T, it's ~65000 IOPS in the graph, which computes to 260 MB/s. (And for the old BX300, it's ~38000 IOPS, or 150 MB/s). No other site gets even half as much in their reviews; your own result in CDM is 89.42 MB/s.

Actually, no flash-based SSD reaches more than about 110 MB/s, with the possible exception of some exotic stuff like the Samsung Z-NAND SLC. Even the Phison representative here at TPU didn't mention faster speeds when asked specifically about that.

Random write IOPS seem more consistent with what CDM says, as well as with other reviews.

What software do you use to measure IOPS? What are the parameters, such as test area size?

Thanks for your feedback, I just investigated:

https://imgur.com/HFdEIKF

This suggests that my numbers are accurate and the load profile on the hardware matches what the test claims to do. The Queue Depth is 1 (actually only 0.7), the transfer size is 4K, and it shows IOPS/MB/s that are in-line with my reviews (the drive is Samsung 990 Pro, makes no difference in this case).

Maybe CDM is simply SUPER slow in the way it performs QD1 ?

 

[Assimilator]

In fact, the 2TB PCIE 4.0 SSD, which combines Yangtze Memory Technologies' 232-layer TLC particles with Maxio's MAP1602 controller scheme, has been available in the Chinese market for some time now, with prices as low as 600 RMB (around 75 Euros). This package not only delivers top-tier PCIE 4.0 performance but also offers a warranty of 3600TBW. Many brands offer products based on this scheme. In Chinese testing articles and videos, many have extensively and thoroughly tested this SSD and have fully recognized its excellent performance, good quality, and affordable price.

And a 2TB PCIE 4.0 SSD that uses a similar scheme with 128-layer particles usually sells for 500 RMB (about 63 Euros).

I'm mentioning all this because I'm surprised that TechPowerUp, an excellent tech website, seems to be completely unaware of this. You might need some editors who are more attuned to Chinese language content and the Chinese market.

I'm confused, is this a different model you're referring to? Or are you saying that there are variants of the NV7000-T with different NAND and TBW?

 

[Wirko]

Maybe CDM is simply SUPER slow in the way it performs QD1 ?

AS SSD results correspond with CDM results, it never gets higher than about 115 MB/s (luckily there's a TPU thread for everything). It would be even better if these programs reported IOPS instead of MB/s for random but it is what it is.

To me it looks like there are different levels of randomness in "random" reads/writes. No idea how to prove or disprove that. However, your recent results are consistent with your older results, probably all the way back to the ancient BX300, so I don't consider it a big issue.

 

[W1zzard]

AS SSD results correspond with CDM results, it never gets higher than about 115 MB/s (luckily there's a TPU thread for everything). It would be even better if these programs reported IOPS instead of MB/s for random but it is what it is.

To me it looks like there are different levels of randomness in "random" reads/writes. No idea how to prove or disprove that. However, your recent results are consistent with your older results, probably all the way back to the ancient BX300, so I don't consider it a big issue.

https://imgur.com/4PG4fCd

that's CDM .. how interesting .. 4K size confirmed, and queue depth is close to 1, too, yet IOPS are much lower

 

[Kaleid]

Have to be crazy to buy nvme with chinese controller and chinese nand. All their tech is stolen and all they can do is come up with half-baked product. Saving few dollars is not worth the headaches you gonna get from a product like that.

This is always fun to read, as if they don't do any work on their own, no research either.

 

[Sabishii Hito]

The "Semizoku" (Cicada family) NVMe drives finally hit the US market I see. They got that name in Japan as they're all variants of HikSemi CC700

 

[mechtech]

Don't know much about YMTC nand, how good is it?

I believe Apple was using that NAND in their devices until the US Gov said nope.

 

[chrcoluk]

Sadly due to recent SN850X price drop, the price actually isnt anything special in my opinion.

SN850X trusted brand with DRAM UK price £114.98
Notac NV7000 2TB no DRAM UK price £116.99

The goal posts got moved with the recent price drops of higher end drives, these dramless drives now need to be at least 20% cheaper than what they are so in USD about $60-65 for 2TB in my opinion.