Patches for Mac (including ARM NEON 128 bits)

[valassi]

This is a WIP comprehensive patch for Mac specific issues (#375).

It also includes, modifies and supersedes Stefan's PS #421. For more detailed comments why this was necessary, see #221.

[valassi]

This is still in WIP. Amngst the things that are missing:

• generate tput logs for mac
• presently the above fails in the fortran test (it crashes? note that the fortran/c++ does build however)
• backport to codegen
• regenerate all five processes auto and manual
• rerun all tputs on mac for these five processes

[valassi]

This is now complete (no longer WIP). Amongst the things that were added:

• generate tput logs for mac
• fixed the fortran test (I was missing a path in DYLD_LIBRARY_PATH for my compiler, and I had to bypass MacOS SIP, System Integrity Protection, in my scripts)
• backport to codegen
• regenerate all five processes auto and manual
• rerun all tputs on mac for these five processes
• added a summary table

This is a summary table from the tests of all five processes double/float on Mac M1 ARMv8 with and without NEON SIMD. One nicely sees factors close to 2x and 4x for double and float
https://github.com/madgraph5/madgraph4gpu/blob/b5ef53ca10a77bece14c0814e5d5ef0275dee53e/epochX/cudacpp/tput/summaryTable_macm1.txt

*** FPTYPE=d ******************************************************************

+++ cudacpp REVISION ea28661 +++
On mac-1T0-438.local [CPU: Apple M1] [GPU: none]:

[Apple clang 12.0.5] 
HELINL=0 HRDCOD=0
                eemumu          ggtt         ggttg        ggttgg       ggttggg
         [2048/256/12]  [2048/256/1]    [64/256/1]    [64/256/1]     [1/256/1]
CPP/none      3.51e+06      5.77e+05      7.02e+04      5.43e+03      2.24e+02
CPP/sse4      7.55e+06      9.79e+05      1.39e+05      1.11e+04      3.65e+02

*** FPTYPE=f ******************************************************************

+++ cudacpp REVISION ea28661 +++
On mac-1T0-438.local [CPU: Apple M1] [GPU: none]:

[Apple clang 12.0.5] 
HELINL=0 HRDCOD=0
                eemumu          ggtt         ggttg        ggttgg       ggttggg
         [2048/256/12]  [2048/256/1]    [64/256/1]    [64/256/1]     [1/256/1]
CPP/none      3.57e+06      5.76e+05      7.10e+04      5.59e+03      2.61e+02
CPP/sse4      1.50e+07      1.85e+06      2.48e+05      2.16e+04      7.21e+02

[valassi]

All checks have passed. I am self merging.

@roiser, I will need to merge this into your alphas patch, I hope this does not cause issues. (I actually started from there, I saw that the alphas patch includes the Mac NEON patch, so I wanted to check it and merge it standalone first).