PR 56919 Improve SYSTEM_CLOCK intrinsic on Windows.

frontend ChangeLog:

2013-04-15  Janne Blomqvist  <jb@gcc.gnu.org>

	PR fortran/56919
	* intrinsics.texi (SYSTEM_CLOCK): Update documentation.


libgfortran ChangeLog:

2013-04-15  Janne Blomqvist  <jb@gcc.gnu.org>

	PR fortran/56919
	* intrinsics/time_1.h: Check __CYGWIN__ in addition to
	__MINGW32__.
	* intrinsics/system_clock.c (GF_CLOCK_MONOTONIC): Check
	_POSIX_MONOTONIC_CLOCK as well.
	(system_clock_4): Use GetTickCount on Windows.
	(system_clock_8): Use QueryPerformanceCounter and
	QueryPerformanceCounterFrequency on Windows.



git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@197968 138bc75d-0d04-0410-961f-82ee72b054a4

1 files changed, 30 insertions, 19 deletions

diff --git a/gcc/fortran/intrinsic.texi b/gcc/fortran/intrinsic.texi
index 8c0edc712a0..d5ff9a048dc 100644
--- a/gcc/fortran/intrinsic.texi
+++ b/gcc/fortran/intrinsic.texi
@@ -12038,27 +12038,38 @@ and should considered in new code for future portability.
 @item @emph{Description}:
 Determines the @var{COUNT} of a processor clock since an unspecified
 time in the past modulo @var{COUNT_MAX}, @var{COUNT_RATE} determines
-the number of clock ticks per second.  If the platform supports a high
-resolution monotonic clock, that clock is used and can provide up to
-nanosecond resolution.  If a high resolution monotonic clock is not
-available, the implementation falls back to a potentially lower
-resolution realtime clock.
-
-@var{COUNT_RATE} is system dependent and can vary depending on the kind of the
-arguments. For @var{kind=4} arguments, @var{COUNT} usually represents
-milliseconds, while for @var{kind=8} arguments, @var{COUNT} typically
-represents micro- or nanoseconds. @var{COUNT_MAX} usually equals
-@code{HUGE(COUNT_MAX)}.
-
-If there is no clock, @var{COUNT} is set to @code{-HUGE(COUNT)}, and
-@var{COUNT_RATE} and @var{COUNT_MAX} are set to zero.
-
-When running on a platform using the GNU C library (glibc), or a
-derivative thereof, the high resolution monotonic clock is available
-only when linking with the @var{rt} library.  This can be done
-explicitly by adding the @code{-lrt} flag when linking the
+the number of clock ticks per second.  If the platform supports a
+monotonic clock, that clock is used and can, depending on the platform
+clock implementation, provide up to nanosecond resolution.  If a
+monotonic clock is not available, the implementation falls back to a
+realtime clock.
+
+@var{COUNT_RATE} is system dependent and can vary depending on the
+kind of the arguments. For @var{kind=4} arguments, @var{COUNT}
+represents milliseconds, while for @var{kind=8} arguments, @var{COUNT}
+typically represents micro- or nanoseconds depending on resolution of
+the underlying platform clock. @var{COUNT_MAX} usually equals
+@code{HUGE(COUNT_MAX)}. Note that the millisecond resolution of the
+@var{kind=4} version implies that the @var{COUNT} will wrap around in
+roughly 25 days. In order to avoid issues with the wrap around and for
+more precise timing, please use the @var{kind=4} version.
+
+If there is no clock, or querying the clock fails, @var{COUNT} is set
+to @code{-HUGE(COUNT)}, and @var{COUNT_RATE} and @var{COUNT_MAX} are
+set to zero.
+
+When running on a platform using the GNU C library (glibc) version
+2.16 or older, or a derivative thereof, the high resolution monotonic
+clock is available only when linking with the @var{rt} library.  This
+can be done explicitly by adding the @code{-lrt} flag when linking the
 application, but is also done implicitly when using OpenMP.
 
+On the Windows platform, the version with @var{kind=4} arguments uses
+the @code{GetTickCount} function, whereas the @var{kind=8} version
+uses @code{QueryPerformanceCounter} and
+@code{QueryPerformanceCounterFrequency}. For more information, and
+potential caveats, please see the platform documentation.
+
 @item @emph{Standard}:
 Fortran 95 and later