Merged revisions 148-395 via svnmerge from
https://software.sandia.gov/svn/public/exact/exact/trunk
……..
r272 | wehart | 2006-12-23 12:44:11 -0700 (Sat, 23 Dec 2006) | 5 lines
Fixed up management of the EXACT_DRIVER environmental variable,
to ensure that it's reset after an experiment is performed.
Updated the TODO in the README file.
……..
r273 | wehart | 2006-12-23 14:08:48 -0700 (Sat, 23 Dec 2006) | 2 lines
Misc extension of testing semantics. Can use 'leq' for less-than-or-equal-to, and 'geq' for greater-than-or-equal-to.
……..
r274 | wehart | 2006-12-23 14:43:11 -0700 (Sat, 23 Dec 2006) | 3 lines
Simplified the EXACT valgrind driver.
Copy this driver in the Acro bin directory.
……..
r311 | wehart | 2007-01-12 14:09:12 -0700 (Fri, 12 Jan 2007) | 3 lines
Adding feature to preserve old experimental results when
rerunning an experiment.
……..
r312 | schake | 2007-01-12 17:04:06 -0700 (Fri, 12 Jan 2007) | 2 lines
memmon - monitor a linux executable's memory
……..
r313 | wehart | 2007-01-13 14:25:42 -0700 (Sat, 13 Jan 2007) | 2 lines
Moving src/memmon directory to src/exec.
……..
r314 | wehart | 2007-01-13 16:33:55 -0700 (Sat, 13 Jan 2007) | 15 lines
Rework of EXACT's executable path to point to the exact/src
directory when testing exact itself.
Rework of memmon.c testing tool. This used to compute the mapped
memory, but that does not correspond to the virtual memory size, which
this now grabs from proc$pid/status.
Added an EXACT test study that uses memmon and verifies that it's getting
a nonzero memory.
Revised the testfn.cpp command to sleep if a flag is passed. This ensures
that memmon can assess it's peak memory usage.
……..
r315 | slbrow | 2007-01-13 16:58:14 -0700 (Sat, 13 Jan 2007) | 2 lines
Adding the exact_valgrind to the EXTRA_DIST list to fix 'make distcheck'.
……..
r352 | wehart | 2007-01-18 09:41:07 -0700 (Thu, 18 Jan 2007) | 5 lines
Misc change to EXACT: Guessing path info when exact is a subdirectory.
Misc bug fix in experiment.py … for when solver info is not provided
as an option.
……..
r360 | wehart | 2007-01-20 10:36:06 -0700 (Sat, 20 Jan 2007) | 3 lines
Adding a conditional compile to creat a dummy memmon executable
when building with MINGW.
……..
r363 | wehart | 2007-01-20 12:39:12 -0700 (Sat, 20 Jan 2007) | 2 lines
Portability fix for MINGW.
……..
r378 | wehart | 2007-01-23 08:54:44 -0700 (Tue, 23 Jan 2007) | 2 lines
Adding logic to test if Valgrind exists.
……..
r379 | schake | 2007-01-23 11:14:51 -0700 (Tue, 23 Jan 2007) | 7 lines
Added interval timer (ms) to print memory usage periodically
-i 2000 will print every 2 sec
Added logfile
-o logfile will print non-error messages to logfile
……..
r380 | schake | 2007-01-23 17:56:31 -0700 (Tue, 23 Jan 2007) | 4 lines
code cleanup with getopt
major bugfix with gettimeofday
……..
r381 | schake | 2007-01-24 10:19:13 -0700 (Wed, 24 Jan 2007) | 2 lines
bugfix
……..
r382 | schake | 2007-01-24 16:50:47 -0700 (Wed, 24 Jan 2007) | 2 lines
type fixes for 64bit
……..
r383 | wehart | 2007-01-24 20:35:59 -0700 (Wed, 24 Jan 2007) | 3 lines
Misc fix for valgrind. Tests to see if valgrind exists, and if not
it prints "dummy statistics".
……..
r385 | wehart | 2007-01-24 22:02:48 -0700 (Wed, 24 Jan 2007) | 2 lines
Portability fix.
……..
r388 | wehart | 2007-01-26 07:50:40 -0700 (Fri, 26 Jan 2007) | 4 lines
Reworked the parsing of Valgrind information.
Setup example5 to verify that there is a memory leak in the
testfn.cpp command.
……..
r389 | wehart | 2007-01-26 07:56:58 -0700 (Fri, 26 Jan 2007) | 4 lines
Cleaned up the exact_valgrind script.
Setup the testfn.script to print dummy valgrind information on
platforms where valgrind is not available.
……..
r390 | wehart | 2007-01-26 07:58:27 -0700 (Fri, 26 Jan 2007) | 2 lines
Commented out debugging print statement.
……..
r391 | wehart | 2007-01-26 08:45:53 -0700 (Fri, 26 Jan 2007) | 8 lines
Rework of how a study file processes results.
Enabled a data option that can specify that only a part of
a study's results are read in, and from a specific file.
This allows for simpler specification of baseline results files
for comparisons.
See example19, which tests this new capability.
……..
r392 | wehart | 2007-01-26 10:39:59 -0700 (Fri, 26 Jan 2007) | 5 lines
Adding exact_timing driver, which print timing statistics in a canonical
format. This is useful, for example, for timing third-party solvers.
Fixed misc bug in exact_valgrind. Needed to print leak check info.
……..
r393 | wehart | 2007-01-26 11:01:44 -0700 (Fri, 26 Jan 2007) | 2 lines
Adding exact_pexec driver. This needs more work…
……..
r394 | wehart | 2007-01-26 12:54:12 -0700 (Fri, 26 Jan 2007) | 3 lines
Extending EXACT semantics to allow python tests for
filtering with python.
……..
