I’ve been reading through a huge amount of posts in regards to hardware, but haven’t got to the bottom of one question;
Will High Clock speed or Core Count have a bigger impact overall? From the documentation, Stage 1 seems to be about Clock Speed, the rest seem to be about Core count.
Which makes a bigger difference on very large projects? (15,000 photos+)
I have a project to process that needs every output including 3D model. It’s a very large site and the photo count is around 20,000 at 20mp (Phantom 4 Pro). Total folder size of photos is around 200GB.
The project has been flown in 6 sections at around 3000 photos each - North/South oblique, then East/West oblique.
I currently have a i7 4.5ghz 4 cores, 64gb ram, 1070 GTX 8GB which is processing each section to Stage 1 ok. I’d like to process all photos in one project if possible, but it’s crashing my current system (Breakpoint A Reached?)
I have the option of purchasing the following systems at a decent price:
Workstation: Dell T7610 with 2x Xeon E5-2680v2 20 Cores/40 Threads (2.6/3.4GHz), Quadro K6000 GPU with 2880 Cuda cores, 128gb Ram
Workstation: Dell T7610 with 2x Xeon E5-2650v2 16 Cores/32 Threads (2.6/3.4GHz), Quadro k4000 GPU with 768 Cuda cores (swap for 1070 GTX or multiple K10 Tesla), 128gb Ram
Workstation: Dell T7600 with 2x Xeon E5-2687w 16 Cores/32 Threads (3.1/3.8GHz), Quadro 6000 GPU with 448 Cuda cores (swap for 1070 GTX or multiple K10 Tesla), 128gb Ram
Server: HPC Rack wth 2x Xeon E5-2620 12 Cores/24 Threads (2.GHz), 4x Tesla K10 GPU with 12288 Cuda cores , 256gb Ram
Is it possible to process all 200gb of photos if there is 256GB Ram?
Will more cores or higher clock speed benefit me the most?
Many thanks for all your help