So finally moved over to the new desktop.
This one is an Ultra7 270k, 16gb ddr5 ram 2.5tb disks and an nvidia rtx5060.
Took a bit to get the nvidia GPU sorted on the present gen of ubunutu desktop. The installation app had been discontinued in ubunutu so had to adapt the installation scripts to include the new version of ‘auto-install’.
The nvidia stuff is all up and running including the NVIDIA visual studio tools within visual studio code.
We’ve got it setup as a virtual GPU server so the graphics are handled by the onboard Intel graphics whilst the whole graphics card is available to ai models and cuda enabled apps and python libraries.
Its fast, really snappy and brings up to date for intel OneAPI, and Nvidia Cuda Compiler.
Usual toolset, Blender, KiCAD, FreeCAD etc and well impressed with the system.
Still to do : install a variety of AI models and compile the latest python libraries.
As regards development Ubuntu and Nvidia is the industry standard, The underlying substrate handles the workflows topologically. Where as windows struggles with that even with windows subsystem for linux.
Going forward- great software AI development environment so will revisit the stuff on developer.nvidia and further explore the ecosystem, Issac sim for robots, omniverse, etc.
Will be installing and operating field of agents to see if we can get a bit of a boost in terms of productivity applications outputs.