Memory Leak in v0.6.0.3

[GoogleCodeExporter]

What is the expected output? What do you see instead?
I Ran Performance Monitor to track down memory usage over the period of 1 day 
and the expected output is not to see memory usage increases with time.
Unfortunately, the memory usage was increasing.  

What version of the product are you using? On what operating system?
0.6.0.3. Windows 7 home Premium 64-Bit , Service Pack 1.

Please provide any additional information below.
Please see attachment for the Performance Monitor output.

Please attach a Report created with "File / Save Report...".

Original issue reported on code.google.com by husam.el...@gmail.com on 9 Aug 2013 at 3:34

Attachments:

• OpenHardwareMonitor.Report.txt
• OHM0.6.0.3_MemoryLeak.png

[GoogleCodeExporter]

I would say the expected output for the memory usage over the period of 1 day 
is an increasing memory usage. All the sensor values are recorded over the 
period of 1 day (which of course comsumes memory). The additional memory usage 
should be about

(12 bytes per sensor value) * (15 * 60 * 24 recorded values per day and sensor) 
* (19 sensors) = 4.7MB

There is some compression applied to the storage of values, but there is 
overhead from the dynamic memory structure as well. While this can explain some 
of the memory usage, it does not explain the increase in "# Bytes in all 
Heaps". If I understand the values in this plot right, then the largest portion 
would have to be on the LOH (large objects heap). Either this problem is caused 
by some nasty LOH fragmentation effect (or the like), or there is some other 
leaking going on somewhere.

To do anything on this issue the source of the leaking has to be identified (if 
there is any).

Original comment by moel.mich on 9 Aug 2013 at 4:16

• Changed state: NeedsInfo

[GoogleCodeExporter]

I will investigate it more . Thank you for your reply.

Original comment by husam.el...@gmail.com on 9 Aug 2013 at 5:00

[GoogleCodeExporter]

I just realized that the scale on the left is not MB but 0.1 MB (except 
"Private Bytes"). So the "# Bytes in all Heaps" (on my system this is actually 
Gen 1 + Gen 2 + LOH but without Gen 0) increases from about 2MB to 8MB. The 
"Private Bytes" from 32MB to 38MB. This would still be within the expected 
increase in memory usage due the recording of sensor values. Of course if it 
continues to increase with each day, then there is really a leaking issue (the 
OHM starts dropping recorded sensor values after 24h).

Original comment by moel.mich on 9 Aug 2013 at 6:39