[Bug c/50994] New: wanted: interface for querying cache size

[Bug c/50994] New: wanted: interface for querying cache size

[bkoz at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50994

             Bug #: 50994
           Summary: wanted: interface for querying cache size
    Classification: Unclassified
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: enhancement
          Priority: P3
         Component: c
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: bkoz@gcc.gnu.org


GCC is missing a way to query cache line size. This could be provided as a
compile-time constant, in the form of predefined macros in a form similar to
the existing sizeof macros via something like __SIZEOF_L1_CACHE_LINE__. Or, it
could be via a __builtin, like __builtin_l1_cache_line_size(). 

Results would be consistent with existing --param flags, such that

--param l1-cache-line-size=64

meant that

__builtin_l1_cache_line_size() == 64.

I notice that intel provides the following interface, via:
http://software.intel.com/sites/products/documentation/studio/composer/en-us/2011Update/compiler_c/index.htm


unsigned int __cacheSize(unsigned int cacheLevel)

"__cacheSize(n) returns the size in kilobytes of the cache at level n. 1
represents the first-level cache. 0 is returned for a non-existent cache level.
For example, an application may query the cache size and use it to select block
sizes in algorithms that operate on matrices."

ding ding ding!!

[Bug c/50994] wanted: interface for querying cache size

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50994

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> 2011-11-04 23:12:51 UTC ---
The cache block/line size could be different on two different machines so it
cannot be constant.  Also there is no portable way of figuring out the cache
size.

[Bug middle-end/50994] wanted: interface for querying cache size

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=50994

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=88466

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I think this has the same issues as std::hardware_destructive_interference_size
has really. In that they can't be a constant as it might be different between
runtime env.