Desktop vs. Cloud Comparison

I have posted about this in different threads and wanted to create one place to consolidate information on the effect of different hardware on time and results.

My desktop is not “average” and my projects are 3D models, not survey work so your individual experience will likely be different…hopefully this gives you a hint of the testing you should do to validate your work.

The project was created in the Desktop version and then uploaded to Pix4D’s cloud and ran on my desktop computer with identical settings.


My rig: 

CPU - Intel Core i9-7900X overclocked 39%
CPU Cooler - Liquid Cooler by Cooler Master
GPU - DUAL SLI GeForce GTX 1080 Ti Xtreme Edition 11GB
RAM Amount - 128GB in (8) 32GB chips
RAM Speed - DDR4 DRAM 3466 (PC4-27700) C16
Hard Drive or SSD or M.2 - Pair of M.2 1TB NVMe PCI-Express 3.0 x4 in RAID 0
Motherboard Model - LGA 2066 Intel X299 SATA 6Gb/s USB 3.1 ATX
Operating system - Windows 10 Home


The cloud server is Pix4D’s and I don’t have all the specifications but I figure it is better than average.


First is the Desktop version:

And now the Pix4D Enterprise cloud version:

In case these images are not readable, I have put them in this folder on my Dropbox: - please note there are also some other cloud server comparisons of a different project in that folder.

Could you let me know the overclocking process you chose for the i9?

Philip, it is much easier than 20 years ago so it just depends on your hardware and the tools that come with it…I didn’t do anything special to overclock the CPU, just the usual as with any recent Intel chip.

Adam, I have the same CPU as you, the i9 7900X and a Titan X Pascal GPU.

Having major issues with step 3 on a 200 image Jon. Hours and hours. Have been in touch with support but something is very very wrong for this to take a whole day. On my less spec Xeon rig, it flies.

Philip, what GPU is on your Xeon rig? I am testing a variety of video cards (and CPUs) in Pix4D, and have found that the Titan and Quadro cards perform much worse than even mid-range GeForce cards. They should be every bit as fast, or faster, but there appears to be something that isn’t playing nicely between Pix4D and those families of video cards. If you have a GeForce instead on the Xeon system, that probably explains the performance difference you have observed.