Results: Samsung SSD 860 EVO 4TB

[ferrix]

Driv Size  Drive Identity     Location:    0      25%     50%     75%     100
---- ----- ---------------------------- ------- ------- ------- ------- -------
 81  4.0TB Samsung SSD 860 EVO 4TB       538.5   541.1   539.5   542.3   542.3

                  Benchmarked: Wednesday, 2020-12-30 at 14:06

(second run follows)

Driv Size  Drive Identity     Location:    0      25%     50%     75%     100
---- ----- ---------------------------- ------- ------- ------- ------- -------
 81  4.0TB Samsung SSD 860 EVO 4TB       538.2   540.6   539.2   541.7   541.8 

                  Benchmarked: Wednesday, 2020-12-30 at 14:07

 

[lovecakeboy]

Hi, thanks for posting!

Looks pretty good! How old is that drive?

The slight variation between runs is interesting too...

 

[ferrix]

Looks pretty good! How old is that drive?

Purchased November 9, 2019. It's been in continuous service since then in a developer workstation hosting VMware guests.

 

[lovecakeboy]

Thanks for the info!

Wow, it's pretty impressive how little reduction in read speed there is for over a year's use. On here we've seen a lot of SSDs where the start of a drive in particular has lost a lot of speed. I'm getting the impression (as yet anecdotal and not scientific!) that the Samsung drives, either the quality of the flash memory, or the way it is managed, are doing something right!

 

[ferrix]

Thanks for the info!

Wow, it's pretty impressive how little reduction in read speed there is for over a year's use. [...] the Samsung drives, either the quality of the flash memory, or the way it is managed, are doing something right!

Yeah I've had lots of these over recent years (I keep switching to new bigger ones when they come out like a data hermit crab) and never had too much of an issue. It's possible that my dev workload is somehow more beneficial to the drive's health than other more normal use cases?

 

[lovecakeboy]

Yeah I've had lots of these over recent years (I keep switching to new bigger ones when they come out like a data hermit crab)

"Data hermit crab", beautiful metaphor there! Nice!

Yes, dunno about the workload thing. I don't think I understand enough about what's going on to speculate.

 

[ferrix]

I will say, some of my vmware operations have seemed to get slower and slower as the drive has been in service longer. I often have 10-30sec drags on certain operations. I suspect the sectors that ReadSpeed chose to look at are just all part of the fast ones, and there's probably a handful of little problem areas. The drive's so big it would be reasonable for it to have missed those.

If the drive was <2TB and it could run Spinrite in a realistic amount of time I'd do that to try and spiff it back up. This is the first drive I've ever had that's just *not possible* to run spinrite no matter how I contort it. And even if it was possible it would take intractably long.

 

[Steve]

If the drive was <2TB and it could run Spinrite in a realistic amount of time I'd do that to try and spiff it back up. This is the first drive I've ever had that's just *not possible* to run spinrite no matter how I contort it. And even if it was possible it would take intractably long.

I'm happy to report that won't be true much longer, Greg! I'm up to my elbows in SpinRite's code!

 

[ferrix]

I'm happy to report that won't be true much longer, Greg! I'm up to my elbows in SpinRite's code!

Ayep, and that's why I'm here. Now it's just a "race" between your assembly coding vs. me deciding to go find my old DRevitalize license

But I'm in pretty good shape still, so inclined to wait. Because once SR is done I'd like to give it a nice juicy 4T SSD test case