def execute(argv):
"""Generate random sequences for each benchmark"""
del argv
FLAGS = flags.FLAGS
# The benchmarks
benchmarks = IO.load_yaml(FLAGS.benchmarks_filename)
if not benchmarks:
logging.fatal('There are no benchmarks to process')
# Create the results directory
try:
os.makedirs(FLAGS.results_directory)
except FileExistsError:
pass
# Process each benchmark
for benchmark in tqdm(benchmarks, desc='Processing'):
index = benchmark.find('.')
bench_dir = benchmark[:index]
bench_name = benchmark[index + 1:]
bench_in_dir = os.path.join(FLAGS.benchmarks_directory, bench_dir,
bench_name)
if not os.path.isdir(bench_in_dir):
continue
bench_out_dir = os.path.join(FLAGS.results_directory, bench_dir)
# Create the results directory for the suite
try:
os.makedirs(bench_out_dir)
except FileExistsError:
pass
filename = '{}/{}.yaml'.format(bench_out_dir, bench_name)
if FLAGS.verify_report and os.path.isfile(filename):
continue
results = {}
for compiler in ['opt', 'llvm']:
for level in FLAGS.levels:
goal_value = Engine.evaluate(
Goals.prepare_goals(FLAGS.goals,
FLAGS.weights), '-{}'.format(level),
compiler, bench_in_dir, FLAGS.working_set, FLAGS.times,
FLAGS.tool, FLAGS.verify_output)
compiler_name = 'clang' if compiler == 'llvm' else 'opt'
if compiler_name not in results:
results[compiler_name] = {}
results[compiler_name][level] = {
'goal': goal_value,
'seq': ['-{}'.format(level)]
}
IO.dump_yaml(results, filename)
def run(self, benchmark):
"""Random algorithm.
Argument
---------
benchmark: str
"""
self.__results = {}
counter = 0
stop = False
while True:
# Create N sequences
sequences = Sequence.create_random_sequences(
self.__flags.nof_sequences, self.__flags.minimum,
self.__flags.maximum, self.__flags.factor, self.__flags.ssa,
self.__flags.shuffle, self.__flags.update,
self.__flags.repetition, self.__flags.original,
self.__flags.passes_filename)
# For each sequence
for _, data in sequences.items():
sequence = data['seq']
# Calculate the fitness
if Sequence.exist(sequence, self.__results):
continue
index = benchmark.find('.')
goal_value = Engine.evaluate(
self.__flags.goals, Sequence.name_pass_to_string(sequence),
self.__flags.compiler,
os.path.join(self.__flags.benchmarks_directory,
benchmark[:index], benchmark[index + 1:]),
self.__flags.working_set, self.__flags.times,
self.__flags.tool, self.__flags.verify_output)
if goal_value == float('inf'):
continue
# Store the results
self.__results[counter] = {'seq': sequence, 'goal': goal_value}
counter += 1
if counter == self.__flags.nof_sequences:
stop = True
break
if stop:
break
def fitness(self,
sequence):
"""Calculate and return the fitness."""
sequence = Sequence.fix_index(list(sequence))
sequence = Sequence.sanitize(sequence)
sequence = Sequence.index_pass_to_list(sequence,
self.passes_dict)
goal_value = Engine.evaluate(self.goal,
Sequence.name_pass_to_string(
sequence
),
self.compiler,
self.benchmark_directory,
self.working_set,
self.times,
self.tool,
self.verify_output)
return [goal_value]
def execute(argv):
"""Evaluate N sequences"""
del argv
FLAGS = flags.FLAGS
# The benchmarks
benchmarks = IO.load_yaml_or_fail(FLAGS.benchmarks_filename)
if not benchmarks:
logging.error('There are no benchmarks to process')
sys.exit(1)
# Create the results directory
try:
os.makedirs(FLAGS.results_directory)
except FileExistsError:
pass
# Process each benchmark
for benchmark in tqdm(benchmarks, desc='Processing'):
# The benchmark
index = benchmark.find('.')
suite = benchmark[:index]
bench_name = benchmark[index + 1:]
# Create the results directory for the suite
results_dir = os.path.join(FLAGS.results_directory, suite)
try:
os.makedirs(results_dir)
except FileExistsError:
pass
# Verify report
if FLAGS.suffix:
output_filename = '{}/{}_{}.yaml'.format(results_dir, bench_name,
FLAGS.suffix)
else:
output_filename = '{}/{}.yaml'.format(results_dir, bench_name)
if FLAGS.verify_report and os.path.isfile(output_filename):
continue
# Benchmark directory
bench_dir = os.path.join(FLAGS.benchmarks_directory, suite, bench_name)
if not os.path.isdir(bench_dir):
logging.error('Benchmark {} does not exist.'.format(benchmark))
sys.exit(1)
# The training data
training_dir = os.path.join(FLAGS.training_directory, suite)
filename = '{}/{}.yaml'.format(training_dir, bench_name)
sequences = IO.load_yaml_or_fail(filename)
if not sequences:
logging.error('There are no sequences to process')
sys.exit(1)
best_sequence = Sequence.get_the_best(sequences)
# Verify if the best sequence is better than the baseline
baseline_dir = os.path.join(FLAGS.baseline_directory, suite)
filename = '{}/{}.yaml'.format(baseline_dir, bench_name)
baseline_data = IO.load_yaml_or_fail(filename)
if not baseline_data:
logging.error('There are no baseline data')
sys.exit(1)
baseline_goal = baseline_data[FLAGS.compiler][FLAGS.baseline]['goal']
for _, data in best_sequence.items():
best_sequence_goal = data['goal']
if not (best_sequence_goal < baseline_goal):
continue
sequences = split_sequence(best_sequence)
# Process the sequences
results = {}
for key, data in sequences.items():
goal_value = Engine.evaluate(
Goals.prepare_goals(FLAGS.goals, FLAGS.weights),
Sequence.name_pass_to_string(data['seq']), 'opt', bench_dir,
FLAGS.working_set, FLAGS.times, FLAGS.tool,
FLAGS.verify_output)
results[key] = {'seq': data['seq'], 'goal': goal_value}
# Store the results
IO.dump_yaml(results, output_filename)
def run(self,
benchmark,
sequence):
"""This method invokes C-Reduce to create a small benchmark.
Argument
----------
benchmark : str
sequence : list
Return
------
ret : bool
True if reduction ok, otherwise False.
"""
index = benchmark.find('.')
bench_in_dir = os.path.join(self.__flags.benchmarks_directory,
benchmark[:index],
benchmark[index+1:])
if not os.path.isdir(bench_in_dir):
lg.error('The directory {} does not exist.'.format(bench_in_dir))
sys.exit(1)
# 1. Calculate the difference (gain)
# 1.1. Baseline's goal value
baseline_goal_value = Engine.evaluate(self.__flags.goals,
self.__flags.baseline,
self.__flags.compiler,
bench_in_dir)
# 1.2. Sequence's goal value
sequence_goal_value = Engine.evaluate(self.__flags.goals,
Sequence.name_pass_to_string(
sequence
),
self.__flags.compiler,
bench_in_dir)
diff = int(baseline_goal_value - sequence_goal_value)
if diff <= 0:
lg.info(
'Warning: It is not possible to reduce the code ({}).'.format(
diff
)
)
return False
# 2. Prepare the benchmark
bench_out_dir = os.path.join(self.__flags.output_directory,
benchmark[:index],
benchmark[index+1:])
try:
os.makedirs(bench_out_dir)
except FileExistsError:
pass
cmdline = 'cp {0}/*.c {1}'.format(bench_in_dir,
bench_out_dir)
try:
subprocess.run(cmdline,
shell=True,
check=True,
capture_output=False)
except subprocess.CalledProcessError:
lg.fatal('Prepare the benchmark')
# 3. Create the C-Reduce script
filename = '{}/test.sh'.format(bench_out_dir)
fout = open(filename, 'w')
fout.write('#!/bin/bash\n')
fout.write('DIFF={}\n'.format(diff))
fout.write('LIB=libMilepostStaticFeatures.so\n')
fout.write('PASSES="{}"\n'.format(
Sequence.name_pass_to_string(sequence))
)
fout.write('clang -Xclang -disable-O0-optnone -S -w -c -emit-llvm *.c\n')
fout.write('llvm-link *.ll -S -o creduce.ll\n')
fout.write('opt --disable-output -load $LIB {} \
--msf creduce.ll 2> msf.txt\n'.format(self.__flags.baseline))
fout.write('size_Oz=`grep f25 msf.txt | \
awk \'{total += $NF} END { print total }\'`\n')
fout.write('rm -f msf.txt\n')
fout.write('opt --disable-output -load $LIB \
$PASSES --msf creduce.ll 2> msf.txt\n')
fout.write('size_PASSES=`grep f25 msf.txt | \
awk \'{total += $NF} END { print total }\'`\n')
fout.write('rm -f msf.txt\n')
fout.write('diff=$(($size_Oz-$size_PASSES))\n')
fout.write('if [[ $diff -eq $DIFF ]]; then\n')
fout.write(' exit 0\n')
fout.write('else\n')
fout.write(' exit 1\n')
fout.write('fi\n')
fout.close()
cmdline = 'chmod +x {}/test.sh'.format(bench_out_dir)
try:
subprocess.run(cmdline,
shell=True,
check=True,
capture_output=False)
except subprocess.CalledProcessError:
lg.fatal('Writing C-Reduce script')
# 4. Invoke C-Reduce
prefix = '{}/'.format(bench_out_dir)
sources = glob.glob('{}*.c'.format(prefix))
sources = [source.replace(prefix, '').replace('.c', '')
for source in sources]
sources = ''.join(sources)
cmdline = 'cd {0} ; creduce ./test.sh \
{1}.c > creduce.log 2> creduce.log'.format(bench_out_dir,
sources)
try:
subprocess.run(cmdline,
shell=True,
check=True,
capture_output=False)
except subprocess.CalledProcessError:
lg.fatal('Invoke C-Reduce')
# 5. Create the new benchmark directory
cmdline = 'cp {0}/Makefile.* {0}/compile.sh {1}'.format(bench_in_dir,
bench_out_dir)
try:
subprocess.run(cmdline,
shell=True,
check=True,
capture_output=False)
except subprocess.CalledProcessError:
lg.fatal('Invoke C-Reduce - Move the small file')
return True
def run(self, sequence, benchmark):
"""Sequence Reduction algorithm.
Suresh Purini and Lakshya Jain.
Finding Good Optimization Sequences Covering Program Space.
TACO.
2013
Argument
--------
sequence : list
benchmark : str
"""
# Calculate the initial value of the goal.
index = benchmark.find('.')
bench_dir = os.path.join(self.__flags.benchmarks_directory,
benchmark[:index], benchmark[index + 1:])
goal_value = Engine.evaluate(self.__flags.goals,
Sequence.name_pass_to_string(sequence),
self.__flags.compiler, bench_dir,
self.__flags.working_set,
self.__flags.times, self.__flags.tool,
self.__flags.verify_output)
# Store the initial value of the goal.
self.__results[0] = {'seq': sequence, 'goal': goal_value}
# Sequence Reduction algorithm
lst_best_sequence = sequence.copy()
best_goal_value = goal_value
change = True
while change:
change = False
bestseqlen = len(lst_best_sequence)
for i in range(bestseqlen):
vector = [1 for i in range(bestseqlen)]
vector[i] = 0
lst_new_sequence = Sequence.remove_passes(
lst_best_sequence, vector)
goal_value = Engine.evaluate(
self.__flags.goals,
Sequence.name_pass_to_string(lst_new_sequence),
self.__flags.compiler, bench_dir, self.__flags.working_set,
self.__flags.times, self.__flags.tool,
self.__flags.verify_output)
if goal_value <= best_goal_value:
best_goal_value = goal_value
lst_best_sequence = lst_new_sequence[:]
change = True
break
goal_value = Engine.evaluate(
self.__flags.goals,
Sequence.name_pass_to_string(lst_best_sequence),
self.__flags.compiler, bench_dir, self.__flags.working_set,
self.__flags.times, self.__flags.tool, self.__flags.verify_output)
# Store the final value of the goal.
self.__results[1] = {'seq': lst_best_sequence, 'goal': goal_value}
def execute(argv):
"""Evaluate N sequences"""
del argv
FLAGS = flags.FLAGS
# The benchmarks
benchmarks = IO.load_yaml_or_fail(FLAGS.benchmarks_filename)
if not benchmarks:
logging.error('There are no benchmarks to process')
sys.exit(1)
# The sequences
sequences = IO.load_yaml_or_fail(FLAGS.sequences_filename)
if not sequences:
logging.error('There are no benchmarks to process')
sys.exit(1)
# Create the results directory
try:
os.makedirs(FLAGS.results_directory)
except FileExistsError:
pass
# Process each benchmark
for benchmark in tqdm(benchmarks, desc='Processing'):
index = benchmark.find('.')
bench_dir = benchmark[:index]
bench_name = benchmark[index+1:]
bench_in_dir = os.path.join(FLAGS.benchmarks_directory,
bench_dir,
bench_name)
if not os.path.isdir(bench_in_dir):
continue
bench_out_dir = os.path.join(FLAGS.results_directory,
bench_dir)
# Create the results directory for the suite
try:
os.makedirs(bench_out_dir)
except FileExistsError:
pass
# Verify report
if FLAGS.suffix:
filename = '{}/{}_{}.yaml'.format(
bench_out_dir,
bench_name,
FLAGS.suffix
)
else:
filename = '{}/{}.yaml'.format(bench_out_dir, bench_name)
if FLAGS.verify_report and os.path.isfile(filename):
continue
# Process the sequences
results = {}
for key, data in sequences.items():
goal_value = Engine.evaluate(
Goals.prepare_goals(FLAGS.goals, FLAGS.weights),
Sequence.name_pass_to_string(data['seq']),
'opt',
bench_in_dir,
FLAGS.working_set,
FLAGS.times,
FLAGS.tool,
FLAGS.verify_output
)
results[key] = {'seq': data['seq'], 'goal': goal_value}
# Store the results
IO.dump_yaml(results, filename)
def execute(argv):
"""Generate random sequences for each benchmark"""
FLAGS = flags.FLAGS
results_directory = FLAGS.results_directory
# Test benchmarks
test_benchmarks = IO.load_yaml_or_fail(FLAGS.test_benchs_filename)
if not test_benchmarks:
logging.fatal('There are no test benchmarks to process')
# Training benchmarks
training_benchmarks = IO.load_yaml_or_fail(FLAGS.training_benchs_filename)
if not training_benchmarks:
logging.fatal('There are no training benchmarks to process')
# Create the results directory
try:
os.makedirs(FLAGS.results_directory)
except FileExistsError:
pass
# Extract the representation for test programs
print(bold('1. EXTRACTING THE REPRESENTATION'))
FLAGS.results_directory = os.path.join(results_directory, 'representation')
FLAGS.benchmarks_filename = FLAGS.test_benchs_filename
representation.execute(argv)
# Distance: test --> training
print(bold('2. MEASURING THE DISTANCE'))
distance_results_directory = os.path.join(results_directory, 'distance')
FLAGS.results_directory = distance_results_directory
FLAGS.test_representation_directory = os.path.join(results_directory,
'representation')
distance.execute(argv)
# Process test benchmarks
print(bold('3. PROCESSING THE BENCHMARKS'))
for nof_sequences in tqdm(FLAGS.nof_sequences, desc='Processing'):
for test_benchmark in test_benchmarks:
index = test_benchmark.find('.')
suite_name = test_benchmark[:index]
bench_name = test_benchmark[index + 1:]
# Find the best N sequences
training_suite, sequences = find_sequences(
test_benchmark, training_benchmarks,
distance_results_directory, FLAGS.training_data_directory,
int(nof_sequences))
# Goal_name
if len(FLAGS.goals) > 1:
goal_name = '_'.join(FLAGS.goals)
else:
goal_name = FLAGS.goals[0]
# Create the results directory for the suite
results_dir = os.path.join(results_directory,
'predictive_compilation',
training_suite, goal_name)
try:
os.makedirs(results_dir)
except FileExistsError:
pass
filename = '{}/{}_j{}.yaml'.format(results_dir, bench_name,
nof_sequences)
if FLAGS.verify_report and os.path.isfile(filename):
continue
results = {}
for key, data in sequences.items():
goal_value = Engine.evaluate(
Goals.prepare_goals(FLAGS.goals, FLAGS.weights),
Sequence.name_pass_to_string(data['seq']), 'opt',
os.path.join(FLAGS.benchmarks_directory, suite_name,
bench_name), FLAGS.working_set, FLAGS.times,
FLAGS.tool, FLAGS.verify_output)
results[key] = {'seq': data['seq'], 'goal': goal_value}
IO.dump_yaml(results, filename)