After 10 years at the helm, CAIA Association chief executive officer William (Bill) Kelly has announced his plans to retire. John Bowman, president at CAIA — which offers education and certifications ...
For -O0, whether -march=native or -march=
Another thing to consider: if PCL's CMake configuration (found by find_package(PCL)) is setting -march=native and other SSE flags, it suggests the PCL on your machine was built with those flags enabled. Compiling your own program with different flags also risks binary compatibility issues. The safest solution may be to recompile PCL with the appropriate flags for the machine you want to run ...
As I understand it, -march=native will detect the ISA and extensions to use from cpuid (which include model, family and stepping information). -march=xxx will use a baseline set of extensions and a baseline ISA. There are a lot of possible combinations of extensions, so only the most relevant were chosen (e.g. skylake-avx512 was added to reflect an important extension of some skylakes). -march ...
What are the differences and tradeoffs between -march=haswell, -march=core-avx2, and -mavx2 for compiling avx2 intrinsics? I know that -mavx2 is a flag and -march=haswell/core-avx2 are architectures which just translate to a bunch of flags. So -mavx2 is a subset of the other two. But beyond that, how do I choose the right one for my application?
-march=foo implies -mtune=foo unless you also specify a different -mtune. This is one reason why using -march is better than just enabling options like -mavx without doing anything about tuning. Caveat: -march=native on a CPU that GCC doesn't specifically recognize will still enable new instruction sets that GCC can detect, but will leave -mtune=generic. Use a new enough GCC that knows about ...