Model Builder Optimizations For Large Scale Models

[TahaEzzy]

Does anyone know ways to optimize model builder to run faster? So far I have got parallel processing on, temp layers going to memory, and overall processes optimized to reduce the number of layers being created. 

Was wondering if there were any other nifty tricks to speed up the process. 

Thanks 

[Ed_]

other than what you have mentioned, faster hardware also helps. For example an SATA SSD is much slower than NVMe4 SSDs. Similarly, running the analysis on a network will be slower than running it on a local drive. Also if the tools in the model have the option to utilize a GPU rather than the CPU, then distributing some of the load to the GPU can also help. If using machine learning, then always use an Nvidia GPU, as ArcGIS Pro is optimized on it. 

[TahaEzzy]

I'm currently using a Gen 4 NVMe, but thank you for the recommendation. The network thing I'm not really sure about as I have the project stored locally along with layers being copied but I utilize One-Drive, which does create backups.  I'm guessing that takes a lot of time, but the only solution to that would be disabling it for the particular project. 

I'm also looking into achieving true parallel processing by turning the tool into a script and using python's in-built parallel processing to manually organize the process, rather than relying on the tools in ArcGIS to potentially have parallel capability built into them. 

[Ed_]

yes, OneDrive does slow down things