util: add a script for testing checkpointing

See comments in util/checkpoint-tester.py for details.
This commit is contained in:
Steve Reinhardt 2010-07-05 21:39:38 -07:00
parent f98cce5771
commit cbd65e95db
2 changed files with 145 additions and 0 deletions

View file

@ -292,6 +292,16 @@ def run(options, root, testsys, cpu_class):
num_checkpoints = 0
exit_cause = ''
# If we're taking and restoring checkpoints, use checkpoint_dir
# option only for finding the checkpoints to restore from. This
# lets us test checkpointing by restoring from one set of
# checkpoints, generating a second set, and then comparing them.
if options.take_checkpoints and options.checkpoint_restore:
if m5.options.outdir:
cptdir = m5.options.outdir
else:
cptdir = getcwd()
# Checkpoints being taken via the command line at <when> and at
# subsequent periods of <period>. Checkpoint instructions
# received from the benchmark running are ignored and skipped in

135
util/checkpoint-tester.py Executable file
View file

@ -0,0 +1,135 @@
#! /usr/bin/env python
# Copyright (c) 2010 Advanced Micro Devices, Inc.
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met: redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer;
# redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution;
# neither the name of the copyright holders nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
# Author: Steve Reinhardt
#
# Basic test script for checkpointing.
#
# Given an M5 command and an interval (in ticks), this script will:
# 1. Run the command, dumping periodic checkpoints at the given interval.
# 2. Rerun the command for each pair of adjacent checkpoints:
# a. Restore from checkpoint N
# b. Run until the timestamp of checkpoint N+1
# c. Dump a checkpoint and end the simulation
# d. Diff the new checkpoint with the original checkpoint N+1
#
# Note that '--' must be used to separate the script options from the
# M5 command line.
#
# Caveats:
#
# - This script relies on the checkpoint options implemented in
# configs/common/Simulation.py, so it works with commands based on
# the se.py and fs.py scripts in configs/example, but does not work
# directly with the existing regression tests.
# - Interleaving simulator and program output can cause discrepancies
# in the file position checkpoint information since different runs
# have different amount of simulator output.
# - Probably lots more issues we don't even know about yet.
#
# Examples:
#
# util/checkpoint-tester.py -i 400000 -- build/ALPHA_SE/m5.opt \
# configs/example/se.py -c tests/test-progs/hello/bin/alpha/tru64/hello \
# --output=progout --errout=progerr
#
# util/checkpoint-tester.py -i 200000000000 -- build/ALPHA_FS/m5.opt \
# configs/example/fs.py --script tests/halt.sh
#
import os, sys, re
import subprocess
import optparse
parser = optparse.OptionParser()
parser.add_option('-i', '--interval', type='int')
parser.add_option('-d', '--directory', default='checkpoint-test')
(options, args) = parser.parse_args()
interval = options.interval
if os.path.exists(options.directory):
print 'Error: test directory', options.directory, 'exists'
print ' Tester needs to create directory from scratch'
sys.exit(1)
top_dir = options.directory
os.mkdir(top_dir)
cmd_echo = open(os.path.join(top_dir, 'command'), 'w')
print >>cmd_echo, ' '.join(sys.argv)
cmd_echo.close()
m5_binary = args[0]
options = args[1:]
initial_args = ['--take-checkpoints', '%d,%d' % (interval, interval)]
cptdir = os.path.join(top_dir, 'm5out')
print '===> Running initial simulation.'
subprocess.call([m5_binary] + ['-red', cptdir] + options + initial_args)
dirs = os.listdir(cptdir)
expr = re.compile('cpt\.([0-9]*)')
cpts = []
for dir in dirs:
match = expr.match(dir)
if match:
cpts.append(int(match.group(1)))
cpts.sort()
# We test by loading checkpoint N, simulating to (and dumping at)
# checkpoint N+1, then comparing the resulting checkpoint with the
# original checkpoint N+1. Thus the number of tests we can run is one
# less than tha number of checkpoints.
for i in range(1, len(cpts)):
print '===> Running test %d of %d.' % (i, len(cpts)-1)
mydir = os.path.join(top_dir, 'test.%d' % i)
subprocess.call([m5_binary] + ['-red', mydir] + options + initial_args +
['--max-checkpoints' , '1', '--checkpoint-dir', cptdir,
'--checkpoint-restore', str(i)])
cpt_name = 'cpt.%d' % cpts[i]
diff_name = os.path.join(mydir, 'diffout')
diffout = open(diff_name, 'w')
subprocess.call(['diff', '-ru', '-I', '^##.*',
'%s/%s' % (cptdir, cpt_name),
'%s/%s' % (mydir, cpt_name)], stdout=diffout)
diffout.close()
# print out the diff
diffout = open(diff_name)
print diffout.read(),
diffout.close()