Re: [gdal-dev] GDAL Warp - NUM_THREADS=ALL_CPUS detrimental?

2020-10-06 Thread Daniel Evans
Thanks Even. From a bit of testing, I can see the time scaling linearly with the requested number of threads despite inputting only 9 pixels, so I suspect that feature is newer than 3.0.4. I'll have a look in the changelogs and keep it in mind. Dr Daniel Evans Software Developer From: gdal-dev

Re: [gdal-dev] GDAL Warp - NUM_THREADS=ALL_CPUS detrimental?

2020-10-06 Thread Even Rouault
On mardi 6 octobre 2020 13:54:30 CEST Daniel Evans wrote: > From reading the documentation more closely, I notice that NUM_THREADS is a > separate bit of functionality from -multithread (which only refers to IO), > so any thoughts of load balancing weren't correct. > > I suspect GDAL is simply doi

Re: [gdal-dev] GDAL Warp - NUM_THREADS=ALL_CPUS detrimental?

2020-10-06 Thread Daniel Evans
>From reading the documentation more closely, I notice that NUM_THREADS is a >separate bit of functionality from -multithread (which only refers to IO), so >any thoughts of load balancing weren't correct. I suspect GDAL is simply doing exactly as instructed - spawning X threads, even if it won'

[gdal-dev] Reading a DODS file from GrADS Data Server

2020-10-06 Thread Kodi Arfer
I currently have GDAL 2.2.3. `gdalinfo --formats` includes "DODS -raster- (ro): DAP 3.x servers". However, if I download a small MERRA-2 DODS file from NASA with this OPeNDAP URL wget 'http://opendap.nccs.nasa.gov/dods/merra2_gmi/tavg1_2d_aer_Nx.dods?pm25[1:3][1:3][1:3]' -O out.dods then `gd

[gdal-dev] GDAL Warp - NUM_THREADS=ALL_CPUS detrimental?

2020-10-06 Thread Daniel Evans
Hello, I've found today that when calling gdal.Warp (from Python, though I doubt it matters) with multithreading enabled, setting the option NUM_THREADS=ALL_CPUS results in increased runtime. I find there to be a minimum runtime of ~0.3-0.4s with NUM_THREADS set like this, compared to a runtime

Re: [gdal-dev] Handling non-standard character encoding

2020-10-06 Thread Even Rouault
On mardi 6 octobre 2020 14:36:44 CEST Brad Hards wrote: > While looking at some NITF test data, I’ve found a problem in how we handle > ECS-A (and ECS) characters. BCS-A (essentially ASCII) is fine. However > MIL-STD-2500C CN2 allows some header fields to be ECS-A, and the test data > does this. >