I am about to begin a PhD where I will be processing a large amount of LiDAR point clouds and optical imagery collected through UAV's. I am looking to build a new PC to undertake this work, and I would like some advice on hardware.
I currently have a dual-boot system with Windows and Linux (64-bit) installed. I primarily use Linux (Python, GDAL, SPDLib, RSGISLib) for data processing, however I also use commercial software (ArcGIS & eCognition) for certain projects.
I would like some advice on the following:
- For GIS/Remote Sensing applications, is it much better to have a “true” multi-core CPU than a cheaper dual-core processor with hyper-threading support? Does GIS/Remote Sensing software fully exploit parallel processing?
- For faster processing, is a solid-state hard drive desirable? Are the speed gains worth the expense?
- Is the faster DDR4 memory a necessary expense?
- Should I purchase a dedicated PCIe 3.0 graphics card or are the integrated graphics sufficient (I suspect not)?
- Should I wait a few months until the new AMD Zen and Intel Kaby Lake processors become widely-available?
In summary; given a limited amount of time and money, on what components should I focus my attention to achieve the best performance for GIS/Remote Sensing tasks?