How many GBs of data does a light desktop user write a day?

[RayLopez99]

Assume no heavy databases, light gaming, word processing, compiling some code on Visual Studio, Internet surfing, movie watching (which is mostly readnot write operations). For my quest to answer the below.

RL

Intel's owns statements about their X25-M SSD, which AnandTech analyzed in 2008 here: http://www.anandtech.com/show/2614/4 And the conclusion: fiveyears of use under heavy usage. "Note: by the way, some manufacturers give the total amount of data written to the drive as one of the drive lifetime indicators. For example, Intel guarantees that the total of about 37 TB of data will be written to X25-M drives (20 GB per day for 5 years: “The drive will have a minimum of 5 years of useful life under typical client workloads with up to 20 GB host writes per day.”)

 

[Paul]

Assume no heavy databases, light gaming, word processing, compiling some code on Visual Studio, Internet surfing, movie watching (which is mostly read not write operations). For my quest to answer the below.

RL

Intel's owns statements about their X25-M SSD, which AnandTech analyzed in 2008 here: http://www.anandtech.com/show/2614/4 And the conclusion: five years of use under heavy usage. "Note: by the way, some manufacturers give the total amount of data written to the drive as one of the drive lifetime indicators. For example, Intel guarantees that the total of about 37 TB of data will be written to X25-M drives (20 GB per day for 5 years: “The drive will have a minimum of 5 years of useful life under typical client workloads with up to 20 GB host writes per day.”)

The simple answer might be:

fsutil fsinfo statistics C:

which gives

File System Type : NTFS
UserFileReads : 26260631
UserFileReadBytes : 2517118976 <---
UserDiskReads : 26260658
UserFileWrites : 7632546
UserFileWriteBytes : 1891827712 <---
UserDiskWrites : 7632547
MetaDataReads : 32099
MetaDataReadBytes : 131477504
MetaDataDiskReads : 32108
MetaDataWrites : 44450
MetaDataWriteBytes : 200445952
MetaDataDiskWrites : 54390
...

You might say "oh, how wonderful", until you
actually test it.

I read a 77GB file from the partition, recording
the fsutil output before and after. The number
did not say "77GB". It wasn't Original_number+77GB.

In my testing, it seems to indicate "UserFileReadBytes"
uses a 32 bit counter. If I take (original_number+77GB) mod 2^32
I get much closer to the right answer.

So that route appears useless for the task. Braindead.

Maybe, if there's such a thing as "fsutil64", and
the OS is better than my copy of WinXP, you'll get
the right answer.

I think my point is, if testing that util, do a read op
longer than 4GB, to see whether the counter "rolls over"
or not. Maybe on a 64 bit system, the clever designers
at Microsoft can use a *64 bit* counter. (The counter
is likely in the file system, and not FSUTIL, so I'm
not necessarily blaming the author of FSUTIL for this.
You'd think NTFS would be using 64 bits for things
measuring sizes, in keeping with its gargantuan capacity
for things.)

I don't see a reason, why the WMI subsystem would do
any better. If the brain-dead-ness is in the NTFS
design, then no amount of different paths to get
to it, will change the result. Hopefully Vista/Win7/Win8
have fixed this. Maybe the next time I'm in Win8, I'll
test.

*******

If you want to do a dummy read, try this.

http://www.chrysocome.net/dd

http://www.chrysocome.net/downloads/dd-0.6beta3.zip (Windows)

That version has /dev/null support, like this.

dd if=F:\77gbfile.dd of=/dev/null bs=129024

My block size chosen, is divisible into the test
file size of 77,934,495,744 bytes. It's a read block
size of 252 sectors.

The read took around 13 minutes or so, after which
I ran fsutil, to get the new "UserFileReadBytes".
And noticed the counter seems to be mod 2^32. My
OS used for this test was WinXP Pro SP3 x32.

Paul

 

[Robin Bignall]

Assume no heavy databases, light gaming, word processing, compiling somecode on Visual Studio, Internet surfing, movie watching (which is mostlyread not write operations). For my quest to answer the below.

RL

Intel's owns statements about their X25-M SSD, which AnandTech analyzed in 2008 here: http://www.anandtech.com/show/2614/4 And the conclusion: five years of use under heavy usage. "Note: by the way, some manufacturers give the total amount of data written to the drive as one of the drive lifetime indicators. For example, Intel guarantees that the total of about 37 TB of data will be written to X25-M drives (20 GB per day for 5 years: “The drive will have a minimum of 5 years of useful life under typical client workloads with up to 20 GB host writes per day.â€)

The pro version of SSDlife has a read / write counter, FOR THOSE SSDs
THAT PROVIDE DATA. Unfortunately my Crucial M4 does not support this
function.

 

[Flasherly]

For my quest to answer the below.

I've a nifty utility which analyzes net thruput writes and data
uploads (not reads) into logs up to monthly quota. Even it wasn't
commercial, (within any IT maintenance of data collection routines),
how few might find results spanning back a year & a half interesting.
37Tbytes actually sounds more encouraging than a discrepancy for
abuses, such as time frames a computer is cycled on and off within any
overall indication of quality for the PS unit at times affects, per
se, were one such weakest grade component perhaps latency within a
malignancy for factoring failure by stages respective of components
from a MB's chronic capacitors.

 

[RayLopez99]

I've a nifty utility which analyzes net thruput writes and data

uploads (not reads) into logs up to monthly quota. Even it wasn't

commercial, (within any IT maintenance of data collection routines),

how few might find results spanning back a year & a half interesting.

37Tbytes actually sounds more encouraging than a discrepancy for

abuses, such as time frames a computer is cycled on and off within any

overall indication of quality for the PS unit at times affects, per

se, were one such weakest grade component perhaps latency within a

malignancy for factoring failure by stages respective of components

from a MB's chronic capacitors.

I think that's a "YES, FIVE YEARS OR MORE" as an answer. Thanks!

RL

 

[Yousuf Khan]

Assume no heavy databases, light gaming, word processing, compiling some code on Visual Studio, Internet surfing, movie watching (which is mostly read not write operations). For my quest to answer the below.

RL

Intel's owns statements about their X25-M SSD, which AnandTech analyzed in 2008 here: http://www.anandtech.com/show/2614/4 And the conclusion: five years of use under heavy usage. "Note: by the way, some manufacturers give the total amount of data written to the drive as one of the drive lifetime indicators. For example, Intel guarantees that the total of about 37 TB of data will be written to X25-M drives (20 GB per day for 5 years: “The drive will have a minimum of 5 years of useful life under typical client workloads with up to 20 GB host writes per day.”)

Okay, I can give you an average figure for me. My boot SSD (Corsair
Force 3-240GB), has been in use for 133 days, and so far it's been
written upto 2.66TB, so that works out to exactly 20GB/day. And that's
just the boot drive, I have other drives in the machine used for storing
movies and videos which get used way more during downloads. I got the
figures from HD Sentinel, which runs immediately upon boot-up.

Yousuf Khan

 

[Robin Bignall]

Okay, I can give you an average figure for me. My boot SSD (Corsair
Force 3-240GB), has been in use for 133 days, and so far it's been
written upto 2.66TB, so that works out to exactly 20GB/day. And that's
just the boot drive, I have other drives in the machine used for storing
movies and videos which get used way more during downloads. I got the
figures from HD Sentinel, which runs immediately upon boot-up.

Whereabouts in Sentinel? If it's performance, that section is blank
because Crucial M4s don't make that data available.

 

[RayLopez99]

Okay, I can give you an average figure for me. My boot SSD (Corsair

Force 3-240GB), has been in use for 133 days, and so far it's been

written upto 2.66TB, so that works out to exactly 20GB/day. And that's

just the boot drive, I have other drives in the machine used for storing

movies and videos which get used way more during downloads.

Thanks. I am assuming you are a consumer and not an enterprise client. So 20 GB for a power user but not for a commercial user. Interesting. Almost too much of a coincidence (that it works out to 20 GB/day).

So you should expect your HD to output up to 37 TB data in five years, and with whatever write cycle thresholds you have and the size of the SSD, perhaps your drive may or may not fill up completely in five years.

RL

 

[Robin Bignall]

Whereabouts in Sentinel? If it's performance, that section is blank
because Crucial M4s don't make that data available.

Scrub that. I just found in Sentinel's FAQ how to enable the counters.

 

[Yousuf Khan]

Thanks. I am assuming you are a consumer and not an enterprise
client. So 20 GB for a power user but not for a commercial user.
Interesting. Almost too much of a coincidence (that it works out to
20 GB/day).

Yeah, it was a bit of a surprise to me as well that it worked out so
perfectly to that specific number, which was also exactly IBM's estimate
for average daily disk usage. But then again, the question also seemed
to be asked at an exactly coincidental time when my drive was exactly
133 days old.

So you should expect your HD to output up to 37 TB data in five
years, and with whatever write cycle thresholds you have and the size
of the SSD, perhaps your drive may or may not fill up completely in
five years.

Not necessarily, remember writes also include modifications to existing
files on the disk, as well as erases are considered a write. So things
like log files and paging/swap files will never really grow (or they get
trimmed continuously), but they will keep adding to the write count. My
boot drive stays relatively stable in capacity these days because I do
all of my data storage on my secondary hard drives.

Yousuf Khan

 

[Yousuf Khan]

Whereabouts in Sentinel? If it's performance, that section is blank
because Crucial M4s don't make that data available.

No, not in the performance tab, although it's available there too. I
took the info right out of the SSD's own SMART data, which has a field
called "Total SSD Writes". The field in the performance tab will get
that information out of the Windows performance counters, but the SMART
data is maintained by the SSD itself. I'm not sure if they will see any
discrepancy, as I don't have any other operating systems running like
Linux.

Yousuf Khan

 

[Robin Bignall]

No, not in the performance tab, although it's available there too. I
took the info right out of the SSD's own SMART data, which has a field
called "Total SSD Writes". The field in the performance tab will get
that information out of the Windows performance counters, but the SMART
data is maintained by the SSD itself. I'm not sure if they will see any
discrepancy, as I don't have any other operating systems running like
Linux.

Thanks, Yousef. The SMART fields displayed in HDS don't seem to include
that field, but performance does. Perf Dsk can't see my SSD's SMART
data, although it's there for HDDs. SSDLife Pro says that the Crucial
M4 does not make performance data available.
My average writes / day to SSD are 5.43 GB. Not a heavy user.

 

[Yousuf Khan]

Thanks, Yousef. The SMART fields displayed in HDS don't seem to include
that field, but performance does. Perf Dsk can't see my SSD's SMART
data, although it's there for HDDs. SSDLife Pro says that the Crucial
M4 does not make performance data available.
My average writes / day to SSD are 5.43 GB. Not a heavy user.

It's probably all very vendor-specific. Mine is a Corsair Force 3 SSD,
so it has these fields, but other SSD's may not, even if they use the
same SSD chipset. Each one probably implements different aspects of the
chipset's data.

Yousuf Khan