Seeking Hardware recommendations for Georeferencing

[JasperRomero]

Hello all,

I am looking for input or suggestions from anyone who has experience with hardware-intensive georeferencing workflows in ArcGIS Pro.

I have several sets of historic imagery (hundreds of images) that I am georeferencing to a 3-inch 2022 aerial survey. I work for a city, so my hardware has been limited to (initially) a laptop with very wimpy specs (no dedicated graphics) and (subsequently) a server running remote application of ArcGIS Pro with lots of RAM and CPU (but no dedicated graphics). The reference aerials are on the city Enterprise Connection geodatabase - this may be a throttle on rendering speed, but the remote app was expected to resolve that issue as the server it runs on is proximal to the Enterprise GDB.

Throughout the process, the workflow has been bottlenecked by very slow rendering. Given the necessity to consistently pan, zoom, and toggle different imagery layers on/off rapidly, this has made my progress extremely slow (~50% of time on this project is spent just waiting for Pro's drawing to catch up with the last pan, toggle, zoom etc.).

I suspect that, more so than virtually any other Pro-based workflow, georeferencing demands GPU in addition to RAM and CPU - is this reasonable? I am talking with IT folks and my supervisor about trying to improve hardware to improve workflow efficiency, but I want to have some idea of what kind of hardware I would need to see a significant improvement.

If anyone has experience with this kind of project, could you please weigh in with:

1) what specs your workstation has, 
2) if those specs result in a smooth, efficient process, and
3) if you have been able to improve rendering speed and smoothness for a workflow like this without needing new hardware?

Thanks in advance for any insights or suggestions!

[VenkataKondepati]

I’ve had to make these choices a few times, and one thing I’ve learned is that ArcGIS Pro benefits from a balanced configuration rather than just maxing out one component.

• CPU: Look for fewer cores with higher clock speeds, Pro is still more sensitive to GHz than sheer core count.

• GPU: A dedicated card (NVIDIA RTX/Quadro) really helps with 3D visualization, large basemaps, and smooth navigation.

• Memory: 32 GB is the sweet spot if you’re working with big datasets or running geoprocessing tasks. I’ve noticed performance drops with only 16 GB when handling complex layers.

• Storage: NVMe SSDs make a noticeable difference for project load times and handling temp files.

It also depends on your work type, if you’re more into 2D cartography, CPU/RAM matter most; if you’re doing 3D scenes or heavy symbology, prioritize the GPU.

That’s what’s worked well in my projects, hope it helps you narrow down the choices!

[JasperRomero]

Thank you for weighing in, I will keep this in mind if we end up looking for new hardware.

[tomESRI]

I noticed that you are on an Enterprise Deployment. Something to consider is an Image Server to help you distribute some of the processing, storage etc. However, if you are a relatively small team, it might not be so feasible: https://enterprise.arcgis.com/en/image/latest/get-started/windows/what-is-arcgis-image-server-.htm

[JasperRomero]

Thanks for the suggestion - is it fair to assume this would limit demand on local graphics in terms of the drawing workload for panning, zooming, toggling?