Closed GoogleCodeExporter closed 8 years ago
The documentation for Kingston SF-2000 based SSDs says the following:
"This Attribute returns a byte count, in units of Gigabytes
at an update resolution of 64 GBytes. The count
represents the number of bytes written. The Attribute
reads ‘0’ until the number of bytes written reaches
64GB; at 64GB the Attribute increments to a value of ‘64’
(decimal)."
This means the value does not need to be multiplied by 64, but increments in 64
steps only.
The documentation you have linked "number of bytes, in 64GB increments" is
unclear as to what this really means. The above quoted documentation is a lot
clearer about this. Now it could be that things have been different for SF-1200
drives, but I have no data that would support this. The OCZ Vertex 2 I use as
well gives reasonable results too.
Do you have any tests that would support your claim of too low values? Can you
attach a report of your system?
Original comment by moel.mich
on 26 Jul 2012 at 11:47
Attached the relevant parts of a report for my system.
My 250 GB SandForce-1200 based SSD (which is 60% full at the moment) has a
"Host Writes to Controller" value of 2944... The resulting write amplification
value seems a bit off as well, but let's take this one step at a time ;)
Original comment by thijsput...@gmail.com
on 26 Jul 2012 at 12:04
Attachments:
All three numbers 4352, 2944, 5120 are a multiples of 64, which agrees with the
documentation I quoted above.
Have you ever updated your firmware? I am not sure what counters could get
reset by a firmware update, but this is something to consider.
I still don't see why the value of 2944 should be wrong.
Original comment by moel.mich
on 26 Jul 2012 at 12:16
You're absolutely right, my bad :|
For some reason I was interpreting the numbers as being in MB instead of GB,
causing them to not make any sense... 2944 GB (instead of MB ;) makes complete
sense.
Anyway, sorry for the wasted effort and again thanks for a great tool!
Original comment by thijsput...@gmail.com
on 10 Aug 2012 at 4:26
Original comment by moel.mich
on 11 Aug 2012 at 10:03
Original issue reported on code.google.com by
thijsput...@gmail.com
on 26 Jul 2012 at 11:29