Closed malonzo47 closed 4 years ago
After reinstallation of gdal following this tutorial it still fails to allocate memory.
Also I tried to check BIGTIFF support but the installation I have does not.
C:\Users\Piao David>gdalinfo --format BIGTIFF ERROR 1: --format option given with format 'BIGTIFF', but that format not recognised. Use the --formats option to get a list of available formats, and use the short code (i.e. GTiff or HFA) as the format identifier.
I am exploring the solution to support BIGTIFF.
Well, the python WHL package file I use for Windows from here include complete GDAL. And checking BigTIFF support from Python using the code below from Stackoverflow, it seems to be fine. Do you know how to check if my installation is fine with BIGTIFF format?
(About the memory allocation issue, I still have it and exploring)
The memory allocation fails at X block offset 1, Y block offset 2. Do you have any idea from the image below?
Please let me know your python version. I am using x32 and some articles say that x64 version would not have this memory allocation problem.
I have 64 bit python on 64 bit windows.
OK, so I believe the memory issue is from that. With 32bit python, it's limited to allocate 2GB memory at a time and it causes this error.
I've never seen the memory allocation issue. I can't even tell if this is related to BigTiff or not. What about allocating more RAM like here:
export GDAL_CACHEMAX=4000
no idea if that's useful or not... (source: https://github.com/tum-gis/cesium-terrain-builder-docker/issues/3)
I believe this is related to the 32bit/64bit. I tried GDAL_CACHEMAX option and it does not help.
So I want to proceed this way.
Since we never actually encountered the BigTiff issue, I will close. The issue was indeed related to 32 bit python and limited memory allocation for large image stacks.
If this is not the case (i.e., with newest version of gdal), then the program will crash when trying to load a large image stack. What's odd though, is that I think this image stack can cause issues even when it's less than 2 GB or 4 GB or whatever the BigTiff threshold is.
Question: Is the relevant gdal version the one that's within python (e.g., Anaconda) or the standalone gdal installed along with OGR?