> On 2018-09-19, at 17:54, Jack <ostrof...@users.sourceforge.net> wrote: > > Unless I'm missing something, I'm surprised nobody has mentioned one > potential problem. Unless you are careful, code you compile for an AMD may > not run on an Intel. If you are just going to compile binary packages to > move and install, this might be OK (assuming you don't trip over trying to > run test during a build.) You should be able to come up with settings so > compiled code will run on either architecture, but I have no idea what the > side effects might be, such as larger and/or less efficient code. I almost > wonder if this doesn't count as cross-compiling, even though it is not a > totally incompatible architecture. Hopefully someone can either expand on > this, or tell me why I'm barking up the wrong tree.
I've dealt with this problem before. The key is flags like -mtune=generic and not using -march=native. Optimisations should still be okay. Andrew
