Switch from 2GB memory space to 3/4 of physical memory

[ChrisJefferson]

The current 2GB memory limit was set in 2012, and memory has grown since then.

On 64-bit OSes, this switches to a default of 3/4 of physical memory, or 2GB, whichever is larger. 32-bit OSes are left at their previous 1GB (just because I don't want to think about issues of filling up the memory space on 32-bit OSes, and most people won't be using them any more).

Tested on linux, mac and cygwin.

For release notes:

• Changed the pre-set memory limit default from 2GB to 3/4 of physical memory. Use the '-o' option if you want to change this limit.

[codecov]

Codecov Report

Merging #3483 into master will decrease coverage by 0.06%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #3483      +/-   ##
==========================================
- Coverage   85.43%   85.37%   -0.07%     
==========================================
  Files         699      698       -1     
  Lines      346759   344604    -2155     
==========================================
- Hits       296253   294204    -2049     
+ Misses      50506    50400     -106

Impacted Files	Coverage Δ
src/system.c	76.1% <100%> (+2.32%)	⬆️
src/modules.h	33.33% <0%> (-66.67%)	⬇️
src/cyclotom.h	50% <0%> (-50%)	⬇️
src/records.h	50% <0%> (-50%)	⬇️
src/gapstate.h	42.85% <0%> (-42.86%)	⬇️
src/vec8bit.h	51.78% <0%> (-42.86%)	⬇️
src/macfloat.h	57.14% <0%> (-42.86%)	⬇️
lib/gprdpc.gi	45.74% <0%> (-42.2%)	⬇️
src/code.h	57.14% <0%> (-40.16%)	⬇️
src/gvars.h	66.66% <0%> (-33.34%)	⬇️
... and 115 more

[coveralls]

Coverage increased (+0.0003%) to 85.259% when pulling cf43990 on ChrisJefferson:physical-mem into a7e6a89 on gap-system:master.

[wilfwilson]

Do you think it would be a good idea to document (perhaps with the command line option for -o) what the defaults are? Not a requirement.

[hulpke]

I think it is fantastic to be able to get the available memory -- in fact I would love in the library to have the abilikty to poll the available memory to check wehther certain calculations are possible.

Setting the default to 3/4 available memoryt could be an issue on multi-user systems (and that was the initial reason for setting a limit): A careless beginner can easily start allocating enormous amounts of memory, by asking for the elements of S_50 or so. How about maintaining a safety level at what would be above a single-user system, but could be on a multi-user server -- somewhere around 16GB?