Manodeep Sinha

Interesting! Looks like both `gcc5.5.0` and `gcc6.4.0` do not enable AVX512VL with `-march=native` ```sh [~ @farnarkle2] ml --force purge && ml gcc/5.5.0 [~ @farnarkle2] gcc -march=native -dM -E - <...

`gcc` is terrible when it comes to intrinsics. For instance, we already have a non-standard `AVX_ABS_FLOAT` in `avx_calls.h` [here](https://github.com/manodeep/Corrfunc/blob/master/utils/avx_calls.h#L91). It is trivial to do the same for `AVX512_ABS_FLOAT` or better...

After looking through gcc docs (https://gcc.gnu.org/onlinedocs/gcc/x86-Options.html#x86-Options), apparently `-march=skylake` does **NOT** add any AVX512 instructions. For that, we need `-march=skylake-avx512`; and that target architecture is not available on gcc5 (only gcc6...

This will require a bit of thinking so that the complexity does not go up (and that we are not implementing `configure` like capabilities into `Corrfunc`). The original reason to...

Thanks Tom! One of the goals of Corrfunc is to ensure that the code compiles on all systems. Supporting OpenMP automatically in a Mac is a complex task, and that's...

I can compile and run with multiple threads even with Apple clang with the following steps: ```sh $ conda install llvm-openmp Then adding the following to the `common.mk` file CC...

If we were thinking of supporting cases where `conda` is not available, then [this page](https://mac.r-project.org/openmp/) might be useful

Thanks for the report. Could this be related to #197 ?

Within the C code, everything is printed to stderr, *except* for the final result that might be written out from a command-line executable. That means nothing is printed out to...

The progressbar code should be okay (just looked through the code), but it has string handling - so who knows!