[Gdal-dev] using GDALDataset* in "point" mode and memory allocation

[Pushkar Pradhan]

> -----Original Message-----
> From: gdal-dev-admin at remotesensing.org
> [mailto:gdal-dev-admin at remotesensing.org]On Behalf Of Gregory, Matthew
> I am working on an app which runs a model at point, list and
> window modes, which gets determined at run-time.  I have a class
> which has a GDALDataset* as a private member and this dataset is
> opened at run initialization through GDALOpen, eg.
>
>   _varLayer = (GDALDataset *) GDALOpen( _varFileName.c_str(),
> GA_ReadOnly );
>
> This class exists throughout the entire model run.  What I've
> noticed when running the point and list (set of points) runs is
> that I am ramping up memory pretty quickly through numerous calls
> to RasterIO, eg.s
>
>   GDALRasterBand* tempBand = _varLayer->GetRasterBand( 1 );
>   tempBand->RasterIO(GF_Read, col, row, 1, 1, ptr, 1, 1,
> GDT_Float64, 0, 0);
>
> I'm guessing that this is NOT the intended way to run RasterIO,
> ie. a pixel at a time.  It looks like each call to RasterIO is
> allocating memory within GetBlockRef() either for cache raster
> blocks or doing actual reads on the data (I'm a bit fuzzy on this).
>
> Note that everything cleans up well when my class goes out of
> scope, but I'm a bit worried that users may run out of dynamic
> memory if they run a huge set of points.
>
> 1.  Is there a better way to free memory after each point is run?
>  I'm imagining that means taking my GDALDataset* out of scope
> each time, which obviously wouldn't be beneficial in terms of speed.
>
> 2.  Is it more economical to front load a number of tiles at
> initialization, rather than potentially one at a time.

Greg,
I don't know why you try to read 1 pixel at a time, maybe your app requires
it but a better way is to read a whole tile:
  poBand2 = poDstDS->GetRasterBand(dwtBand);

  for(row = 0; row < imgYSize; row += blockYSize) {
    for(col = 0; col < imgXSize; col += blockXSize) {
      rc = poBand->RasterIO(GF_Read, col, row, blockXSize, blockYSize,
                            pData, blockXSize, blockYSize, GDT_Float32, 0,
0);
      if(rc == CE_Failure)
        exit(EXIT_FAILURE);

      rc = poBand2->RasterIO(GF_Write, col, row, blockXSize, blockYSize,
                            pData, blockXSize, blockYSize, GDT_Float32, 0,
0);
      if(rc == CE_Failure)
        exit(EXIT_FAILURE);
    }
  }

I don't think you'll run out of memory because memory gets allocated by the
RasterIO() call and I think it does this only the first time. Hope that
helps.
Pushkar