[time-nuts] Timelab bug (was: Framework for simulation of oscillators)

John Miles john at miles.io
Tue Mar 29 15:26:50 EDT 2016


> Here it takes slightly less 3 minutes, but stops at 10M samples.

> Memory consumption of Timelab stays below 350MB the whole time

> and drops to 98MB after it finished.

> 

> OS is windows XP pro

> 

> How many samples did you get? It should have been 14926518 (ie slightly

> less than 15M) 

 

D'oh, sorry, I misunderstood -- I thought you were saying it locked up.
Yes, the normal file-import process stops at 10M samples because that's the
arbitrary maximum value allowed for the Trace Duration property in the
file-import dialog.  I'll go ahead and raise the limit to 100M points for
the 64-bit version.

 

For now, you can work around the 10M-point import limit by using
Acquire->Acquire from live ASCII file instead of File->Import
phase/frequency data.  Most of the same code is used for both cases, but
unlike the static file-import version of the dialog, the live data importer
will let you specify the expected duration yourself.  So you can give it a
duration value that you know will be long enough to cover the whole data set
(10 days in this case.)  

 

To use the live-ASCII acquisition option on a static file, uncheck "Auto"
and check "Include existing file contents" before hitting the Start
Measurement button.  When the file finishes loading, it will sit there
waiting for more data to be appended to the file.  At that point you can hit
the space bar to stop the acquisition, and it will readjust the duration to
correspond to the actual number of points it read.

 



 

 

-- john, KE5FX

Miles Design LLC

 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 24743 bytes
Desc: not available
URL: <http://www.febo.com/pipermail/time-nuts/attachments/20160329/aa392705/attachment.png>


More information about the time-nuts mailing list