Is an NVME ssd worth it?

[chalky16]

Is the 960 evo going to be worth the money? For booting up windows, some applications and some smaller games will it make that big of a difference over something like the 850 evo?

 

[Supahos]

Depends. If you're dropping it into a $850 hell no, if you're dropping it into a $1500+ build probably. On a cheaper build the $30-40 premium could actually help performance a good bit.

 

[geofelt]

Worth is something only YOU can determine.
The 960 EVO primary benefit will come from fast sequential speeds.
A Virus scan, for interest will go noticeably faster.
I would imagine game level loads would do better also.
Not really a big advantage in boot times which have a heavy cpu component to them.
On everyday activities, both will be quick because of low latency.

What to do?

My suggestion is to go ahead and buy the 960 EVO.
The cost difference is not great.
If you do not, you will forever wonder if you should have.

 

[feelinfroggy777]

I like mine and they are not that much more than a SATA SSD. As stated above, I can run a virus scan seconds and not minutes.

 

[Solandri]

Generally they're not that much faster than a SATA 3 SSD, but their price is often close enough that you might as well pay for it. The reason is that MB/s is the inverse of how we perceive drive speed. We don't think a drive is fast because it has high MB/s, we think it's fast because we don't have to wait as long for it to finish tasks - sec/MB.

If you compare these drives in sec/MB, you'll find that the bigger MB/s becomes, the less difference it makes. Say you're copying 1 GB of sequential files.

125 MB/s HDD = 8 sec
250 MB/s SATA 2 SSD = 4 sec
500 MB/s SATA 3 SSD = 2 sec
1000 MB/s NVM SSD = 1 sec
2000 MB/s NVM SSD = 0.5 sec

Notice how every time MB/s doubles, the additional time saved is halved? So even though the jump from an early 1000 MB/s NVM SSD to a newer 2000 MB/s NVM SSD sounds like a big deal (+1000 MB/s), it only shaves 0.5 seconds off this hypothetical copy. Whereas what looks like a small jump (SATA 2 SSD to SATA 3 SSD, +250 MB/s) saves 4x as much time - 2 seconds.

Or put another way, compared to the speedup from a HDD to a 2000 MB/s NVM SSD (7.5 seconds):

A 250 MB/s SATA 2 SSD gives you 53% of the speedup
A 500 MB/s SATA 3 SSD gives you 80% of the speedup
A 1000 MB/s NVM SSD gives you 93% of the speedup
A 2000 MB/s NVM SSD gives you 100% of the speedup

The bulk of the speedup is at the lower MB/s speeds. Since the SATA 3 SSD gives you 80% of the speedup (wait time reduction) of a 2000 MB/s NVM SSD, most people aren't going to notice the difference between the two in normal everyday use. On top of that, the small file speeds on even newer NVM SSDs are still down around 30-70 MB/s - the same as for SATA 3 SSDs, so the speed difference there is even smaller. The exception is if you're working with lots of large files (sequential data). Then that extra 20% can start to become noticeable. But about the only real-world task which works with large amounts of sequential data is real-time video editing.

(MPG for cars suffers from the same problem. Fuel efficiency is the inverse - GPM. So the bigger MPG gets, the less gas you're saving. The Prius with its 50 MPG sounds impressive, but it's really not. Someone switching from a 25 MPG sedan to a 50 MPG Prius only saves half the fuel of someone switching from a 12.5 MPG SUV to a 25 MPG sedan. If the country were serious about reducing oil imports, we'd be working hard to improve truck economy and strongly encouraging people to stop driving SUVs. But instead we're sidetracked with hybrid econoboxes because people assume the 25 MPG improvement for getting a Prius is better than the 10-15 MPG improvement from switching away from a SUV, when it's actually the other way around.)

 

[saravanakmr789]

Generally they're not that much faster than a SATA 3 SSD, but their price is often close enough that you might as well pay for it. The reason is that MB/s is the inverse of how we perceive drive speed. We don't think a drive is fast because it has high MB/s, we think it's fast because we don't have to wait as long for it to finish tasks - sec/MB.

If you compare these drives in sec/MB, you'll find that the bigger MB/s becomes, the less difference it makes. Say you're copying 1 GB of sequential files.

125 MB/s HDD = 8 sec
250 MB/s SATA 2 SSD = 4 sec
500 MB/s SATA 3 SSD = 2 sec
1000 MB/s NVM SSD = 1 sec
2000 MB/s NVM SSD = 0.5 sec

Notice how every time MB/s doubles, the additional time saved is halved? So even though the jump from an early 1000 MB/s NVM SSD to a newer 2000 MB/s NVM SSD sounds like a big deal (+1000 MB/s), it only shaves 0.5 seconds off this hypothetical copy. Whereas what looks like a small jump (SATA 2 SSD to SATA 3 SSD, +250 MB/s) saves 4x as much time - 2 seconds.

Or put another way, compared to the speedup from a HDD to a 2000 MB/s NVM SSD (7.5 seconds):

A 250 MB/s SATA 2 SSD gives you 53% of the speedup
A 500 MB/s SATA 3 SSD gives you 80% of the speedup
A 1000 MB/s NVM SSD gives you 93% of the speedup
A 2000 MB/s NVM SSD gives you 100% of the speedup

The bulk of the speedup is at the lower MB/s speeds. Since the SATA 3 SSD gives you 80% of the speedup (wait time reduction) of a 2000 MB/s NVM SSD, most people aren't going to notice the difference between the two in normal everyday use. On top of that, the small file speeds on even newer NVM SSDs are still down around 30-70 MB/s - the same as for SATA 3 SSDs, so the speed difference there is even smaller. The exception is if you're working with lots of large files (sequential data). Then that extra 20% can start to become noticeable. But about the only real-world task which works with large amounts of sequential data is real-time video editing.

(MPG for cars suffers from the same problem. Fuel efficiency is the inverse - GPM. So the bigger MPG gets, the less gas you're saving. The Prius with its 50 MPG sounds impressive, but it's really not. Someone switching from a 25 MPG sedan to a 50 MPG Prius only saves half the fuel of someone switching from a 12.5 MPG SUV to a 25 MPG sedan. If the country were serious about reducing oil imports, we'd be working hard to improve truck economy and strongly encouraging people to stop driving SUVs. But instead we're sidetracked with hybrid econoboxes because people assume the 25 MPG improvement for getting a Prius is better than the 10-15 MPG improvement from switching away from a SUV, when it's actually the other way around.)

Still very helpful in 2019. Thank you!