[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[IDV #NJS-553517]: Point data intolerably slowwwww even hourly 1 point



> Full Name: brian mapes
> Email Address: address@hidden
> Organization: University of Miami
> Package Version: 5.0u1 build date:2014-07-21 07:05 UTC
> Operating System: Windows 7
> Hardware: Java: home: C:\Program Files\IDV_5.0u1\jre version: 1.7.0_51 
> j3d:1.6.0-pre9-daily-experimental daily
> Description of problem: Here is a Time Driver bundle with several gridded PW 
> fields, and one point display (from a file that has  2448 hourly values).
> The point data is valuable to me because its absolute value is better 
> measured than the satellite and analysis grids.
> 
> When there is no point display, toggling (checking and unchecking) several 
> grids to do intercomparisons is fast.
> But when this ONE point model is present, checking and unchecking grids gives 
> spinning weels,
> and the response is too intolerably slow to be useful.
> This is in a 5GB IDV session with no other jobs running on my dedicated "IDV 
> machine", a PC laptop.
> There is no excuse due to memory or processor availability -- it is the 
> software ineffficiency.
> 
> 
> All data in this bundle are online for your convenience.
> Downloading the point .csv file to local disk gives the same result.
> It seems to be graphical internal activity, not rereading the file.
> But if you want to try it, you could grab the .csv file at
> https://weather.rsmas.miami.edu/repository/entry/get/Gan.UV%2BPW_LWP.csv?entryid=1ae6a83e-6a5e-4776-8dbe-4517a534a8ee
> 
> When creating a point display, I never know what to do with the dialog
> " This file contains 2448 values. Do you want to show them all?"
> I have tried both Yes and No.
> It doesn't lead to any further subsetting dialogs or options like Use Time 
> Driver.
> 
> Thanks for any aid.
> Point data functionality is just useless to me at this speed.
> We would have to make a sparse grid, or (absurdly) a raster image of a 
> station model offline, built by hand, to treat as a grid or image display. 
> Please help.
> 
> 
> 
> 
Brian,
     When the point dataset has more than 1000 time steps, you will get the 
warning message: do you want to show them all..., If you select yes, then all 
the data loaded. If no, 
the IDV will enable the time declutter with the default timeDeclutterMinutes 
set to 1 minute. This 1 minute has no effect on your dataset and you still get 
2448 time steps.

      I just checked in changes to calculate the timeDeclutterMinutes based on 
each dataset and limit the time steps to 1000 if the user select no. I am not 
sure if this is good enough
for your case.

     The best approach is to open the dataset property panel and set the time 
binning for each dataset based on your requirement and other datasets. So, you 
will 
have more controls and not lost any useful data.

Yuan 

Ticket Details
===================
Ticket ID: NJS-553517
Department: Support IDV
Priority: Normal
Status: Open