RT-Preempt on J1900 NUC?

Would really like to upgrade my machine from a older Dell to a fanless NUC. I have a NUC with a J1900 Quad-Core Intel Celeron processor. I did the LinuxCNC realtime on an RT-Preempt install on the Dell but it was a long time ago and I don't remember a lot. I don't know much about kernels either, to be honest. I'll keep the control box and Mesa 7i92 card.

Is there any reason this won't work?

I was looking up what I need to do, do I just need to go to linuxcnc.org/downloads/ and do the LinuxCNC 2.8.0 Debian 10 Buster PREEMPT-RT ISO?

OK, I had the install screen with options "Live (amd64)", "Install", or "Graphical Install". So there was an amd-specific option, but nothing intel-specific.

I just used Install, removed the USB, and got it booting. I do see a message when booting about "Loading Linux (something something) rt-amd64". It goes by fast. This is an Intel, is there anything wrong here?

I can launch LinuxCNC, but there's no hardware on this machine yet, copying the config and setting up to work with the Mesa card may take some time so I don't know yet.

I can run Latency Test of course. I don't have anything installed to stress the machine though. I get Servo 1,030K jitter 30K, Base 59K jitter 34K. Is this ok? It's a stepper machine, no servo encoders or anything

Thanks, I had not heard this. I was worried about this being my only options and thought I had the wrong thing.

... but, sounds pretty dumb to explicitly specify "amd64" if you don't mean AMD. I do recall an era where AMD required a different install of stuff, or there would be an AMD-optimized option that worked better on AMD but would not run on Intel machines.

AMD64 is a standardized CPU instruction set. Invented by AMD but used by Intel as well. Also known as x86-64 and other things.

Does not refer to AMD CPUs, just the instruction set.

Partly got named this because of competing 64 bit architectures at the time. AMDs won and became the standard.

OK- so, when I put this machine together years ago, I took a .c file for a VFD, modified a few lines for my particular VFD, compiled it, and integrated it in. Simple as that should be, I do recall it was a significant hassle to get it to actually DO the compile. I don't remember why, just going through some frustrating back-and-forth of "use this line". "didn't work, gave me this error". "well, it's easy fix this sort of thing when I'm at the machine". "ok great what exactly do I need to do, exactly?". "just keep trying you're close".

Do I need to recompile it for the new machine, or is it ok to just copy the compiled result over and save the trouble?