Processing over 2000 images over Multiple systems

[arpit]

Hello!!

I have multiple systems of following specifications.
CPU - i9 11 Gen
GPU - NVIDIA 24GB
RAM - 129 GB

I want to process over 2000 images as quickly as possible.
with medium or high point cloud quality.

1. Can you please explain how split-merge feature can help me out in this and
2. what should I do to process these images simultaneously on both systems.?

Also,
3. What else upgradation I need in hardware for better and fast results?

[Johnny5]

In my testing using the GPU can reduce the processing time by about 40%, and processing 20MPx images with feature-quality=ultra required 7.3GB GPU RAM so you’ve probably got more than enough GPU RAM (depending on how many MPx your images have).
With default settings processing 2000 images took about 4.5 hours, using about 50GB of RAM and 100GB disk space on my system (Dual Xeon 20-core CPUs, 512GB RAM, 12GB GPU).

Do you need the quality report?
If yes, you have to process the 2000 images without splitting.
If no, you can reduce overall processing time by setting up a cluster of nodes and doing splitting. I don’t know how much it will speed up processing, but I expect it would be a substantial improvement.