Re: physmem: a new program to report memory information

[Pádraig Brady]

On 08/30/2012 04:35 PM, Assaf Gordon wrote:
> Hello,
> 
> Related to the previous sort+memory envvar usage thread: 
> http://thread.gmane.org/gmane.comp.gnu.coreutils.general/3028/focus=3090 .
> 
> Attached is a suggestion for a tiny command-line program "physmem", that 
> similarly to "nproc", exposes the gnulib functions "physmem_total()" and 
> "physmem_available()".
> 
> The code is closely modeled after nproc, and the "recommended" memory usage 
> is calculated using sort's default_sort_size() .
> 
> The program works like this:
> ===
> $ ./src/physmem --help
> Usage: ./src/physmem [OPTION]...
> Prints information about physical memory.
> 
>   -t, --total           print the total physical memory.
>   -a, --available       print the available physical memory.
>   -r, --recommended     print a safe recommended amount of useable memory.
>   -h, --human-readable  print sizes in human readable format (e.g., 1K 234M 
> 2G)
>       --si              like -h, but use powers of 1000 not 1024
>       --help     display this help and exit
>       --version  output version information and exit
> 
> Report physmem bugs to address@hidden
> GNU coreutils home page: <http://www.gnu.org/software/coreutils/>
> General help using GNU software: <http://www.gnu.org/gethelp/>
> Report physmem translation bugs to <http://translationproject.org/team/>
> For complete documentation, run: info coreutils 'physmem invocation'
> ===
> 
> The actual working code (at the bottom of physmem.c) is:
> ===
>   switch(memory_report_type)
>     {
>     case total:
>       memory = physmem_total();
>       break;
> 
>     case available:
>       memory = physmem_available();
>       break;
> 
>     case recommended:
>       memory = default_sort_size();
>       break;
>     }
> 
>   char buf[LONGEST_HUMAN_READABLE + 1];
>   fputs (human_readable (memory, buf, human_output_opts,1,1),stdout);
>   fputs("\n", stdout);
> ===
> 
> So it's very simple, and rely on existing coreutils code.
> 
> Please let me know if this is something you'd be willing to include in 
> coreutils.
> 
> Thanks,
>  -gordon


Thanks for taking the time to do a patch.
However this seems more appropriate for maybe util-linux.
Also it might be better as an option to the existing free(1) program?
Also RAM is just one aspect of the memory hierarchy.
For a separate tool, it should probably be more general
and allow inspection of any level. This already exists
in the hwloc utils for example:

$ hwloc-info -p -v --only machine
Machine (2860MB)
$ hwloc-info -p -v --only cache
L3 (3072KB)
L2 (256KB)
L1 (32KB)

So I wouldn't be that enthusiastic for such a util,
in coreutils at least.

cheers,
Pádraig.