def test_run():
wf_names = [
'Montage_25', 'CyberShake_30', 'Inspiral_30', 'Sipht_30',
'Epigenomics_24'
]
BASE_PARAMS["ga_params"]["population"] = 5
BASE_PARAMS["ga_params"]["Kbest"] = 1
BASE_PARAMS["ga_params"]["generations"] = 10
seq = make_linear_sequence(
BASE_PARAMS, {
"transfer_time": [0, 10, 50, 100, 300, 500, 2000],
"ideal_flops": [1, 20, 50, 100, 500]
})
to_run = [
partial(do_exp, wf_name, **params) for wf_name in wf_names
for params in seq
]
results = [t() for t in to_run for _ in range(3)]
saver = UniqueNameSaver(TEMP_PATH, EXPERIMENT_NAME)
for result in results:
saver(result)
pass
def inherited_pop_run(exp,
wf_tasksids_mapping,
repeat_count,
base_params,
is_debug=False):
path = os.path.join(TEMP_PATH, "igaheft_series")
saver = UniqueNameSaver(path, base_params["experiment_name"])
to_run = []
for wf_name, ids in wf_tasksids_mapping.items():
for id in ids:
params = deepcopy(base_params)
params["executor_params"]["task_id_to_fail"] = id
func = partial(exp, saver=saver, wf_name=wf_name, **params)
to_run.append(func)
to_run = randomize_order(to_run)
if is_debug:
results = [t() for t in to_run for _ in range(repeat_count)]
else:
results = multi_repeat(repeat_count, to_run)
# path = save_path if save_path is not None else os.path.join(TEMP_PATH, "igaheft_series")
# saver = UniqueNameSaver(path, base_params["experiment_name"])
# for result in results:
# saver(result)
pass
def do_exp():
result, logbook = run_dcga(_wf, estimator, rm, heft_mapping, heft_ordering, **params)
saver = UniqueNameSaver("{0}/temp/dcga_exp".format(__root_path__))
data = {
"final_makespan": result,
"iterations": logbook
}
saver(data)
return result
def coeff_run():
"""
coefficient of compute/data intensivity
"""
# wf_names = ['Montage_25', 'CyberShake_30', 'Inspiral_30', 'Sipht_30', 'Epigenomics_24']
#all_coeffs = [1/100, 1/50, 1/10, 1/5, 1/2.766, 15, 20, 25, 30, 35, 40, 45, 75]
all_coeffs = [1 / 100, 1 / 50, 1 / 10, 1 / 5, 1 / 2.766, 1, 2.766] + list(
range(5, 101, 5))
# wf_names = [('Montage_25', [10]),
# ('CyberShake_30', [0.1] + list(range(10, 46, 1))),
# ('Inspiral_30', [10, 1, 1/2.766]),
# ('Sipht_30', [0.02] + list(range(30, 50, 1)) + list(range(50, 101, 5))),
# ('Epigenomics_24', list(range(5, 46, 1)) + list(range(45, 101, 5)))]
# wf_names = [('Montage_25', [2.766])]
wf_names = [ #('Montage_25', [10]),
('CyberShake_30', all_coeffs),
#('Inspiral_30', [10, 1, 1/2.766]),
#('Sipht_30', [0.02] + list(range(30, 50, 1)) + list(range(50, 101, 5))),
('Epigenomics_24', all_coeffs)
]
def transfer_time(max_runtime, c):
transfer = max_runtime * BASE_PARAMS["ideal_flops"] / c
return transfer
to_run = []
for wf_name, coeffs in wf_names:
_wf = wf(wf_name)
max_runtime = max(_wf.get_all_unique_tasks(),
key=lambda x: x.runtime).runtime
param_sets = [
copy_and_set(BASE_PARAMS,
transfer_time=transfer_time(max_runtime, c),
data_intensive_coeff=c) for c in coeffs
]
exps = [partial(do_exp, wf_name, **params) for params in param_sets]
to_run = to_run + exps
# m_repeat = lambda n, funcs: [f() for f in funcs for _ in range(n)]
#results = m_repeat(REPEAT_COUNT, to_run)
results = multi_repeat(REPEAT_COUNT, to_run)
saver = UniqueNameSaver(TEMP_PATH, EXPERIMENT_NAME)
for result in results:
saver(result)
# coeff_aggregate(saver.directory, "coeff.png")
pass
def changing_reliability_run(exp,
reliability,
individuals_counts,
repeat_count,
wf_names,
base_params,
is_debug=False):
path = os.path.join(TEMP_PATH, "gaheft_series")
saver = UniqueNameSaver(path, base_params["experiment_name"])
configs = []
for r in reliability:
for ind_count in individuals_counts:
params = deepcopy(base_params)
params["estimator_settings"]["reliability"] = r
params["alg_params"]["n"] = ind_count
params["alg_params"]["migrCount"] = int(0.1 * ind_count)
configs.append(params)
to_run = [
partial(exp, saver=saver, wf_name=wf_name, **params)
for wf_name in wf_names for params in configs
]
to_run = randomize_order(to_run)
# i = 0
# results = []
# for _ in range(repeat_count):
# for t in to_run:
# print("//////////////////////RUN NUMBER {0}=================".format(i))
# i += 1
# results.append(t())
if is_debug:
results = [t() for t in to_run for _ in range(repeat_count)]
else:
results = multi_repeat(repeat_count, to_run)
# path = save_path if save_path is not None else os.path.join(TEMP_PATH, "gaheft_series")
# saver = UniqueNameSaver(path, base_params["experiment_name"])
# for result in results:
# saver(result)
pass
def real_run():
wf_names = [
'Montage_25', 'CyberShake_30', 'Inspiral_30', 'Sipht_30',
'Epigenomics_24'
]
seq = make_linear_sequence(
BASE_PARAMS, {
"transfer_time": [0, 10, 50, 100, 300, 500, 2000],
"ideal_flops": [1, 20, 50, 100, 500]
})
to_run = [
partial(do_exp, wf_name, **params) for wf_name in wf_names
for params in seq
]
results = multi_repeat(REPEAT_COUNT, to_run)
saver = UniqueNameSaver(TEMP_PATH, EXPERIMENT_NAME)
for result in results:
saver(result)
pass
def test_run(exp, base_params):
configs = []
# reliability = [1.0, 0.95, 0.9]
# reliability = [1.0]
reliability = [0.95]
wf_name = "Montage_25"
for r in reliability:
params = deepcopy(base_params)
params["estimator_settings"]["reliability"] = r
configs.append(params)
to_run = [partial(exp, wf_name=wf_name, **params) for params in configs]
results = [t() for t in to_run]
# results = multi_repeat(REPEAT_COUNT, to_run)
saver = UniqueNameSaver(os.path.join(TEMP_PATH, "gaheft_series"),
base_params["experiment_name"])
for result in results:
saver(result)
pass
def do_exp_schedule(takeHeftSchedule=True):
saver = UniqueNameSaver("../../temp/ga_vs_heft_exp_heft_schedule")
ga_makespan, heft_makespan, ga_schedule, heft_schedule = run(wf_names[0])
## TODO: pure hack
schedule = heft_schedule if takeHeftSchedule else ga_schedule
mapping = [(item.job.id, node.flops)
for node, items in schedule.mapping.items() for item in items]
mapping = sorted(mapping, key=lambda x: x[0])
ordering = [(item.job.id, item.start_time)
for node, items in heft_schedule.mapping.items()
for item in items]
ordering = [t for t, time in sorted(ordering, key=lambda x: x[1])]
data = {"mapping": mapping, "ordering": ordering}
name = saver(data)
return ga_makespan, heft_makespan, ga_schedule, heft_schedule, name
def do_exp(arg):
config, name = tasks[arg]
saver = UniqueNameSaver(os.path.join(base_path, name))
return do_experiment(saver, config, _wf, rm, estimator)
"ideal_inds": {
MAPPING_SPECIE: ms_str_repr,
ORDERING_SPECIE: os_ideal_ind
},
"wf_name": _wf.name
},
"initial_pops": initial_pops,
"final_solution": solution,
"final_makespan": m,
"iterations": logbook
}
saver(data)
return m
saver = UniqueNameSaver(os.path.join(__root_path__, "temp/cga_exp"))
def do_exp():
## TODO: remove time measure
tstart = datetime.now()
res = do_experiment(saver, config, _wf, rm, estimator)
tend = datetime.now()
tres = tend - tstart
print("Time Result: " + str(tres.total_seconds()))
return res
if __name__ == "__main__":
res = repeat(do_exp, 1)
"best_components_itself":
best_components_itself(sols),
"best":
-1 * Utility.makespan(
build_schedule(_wf, estimator, rm,
max(sols, key=lambda x: x.fitness)))
},
"operators": {
# "choose": default_choose,
"build_solutions": default_build_solutions,
# "fitness": fitness_mapping_and_ordering,
"fitness": overhead_fitness_mapping_and_ordering,
# "assign_credits": default_assign_credits
# "assign_credits": max_assign_credits
"assign_credits": assign_from_transfer_overhead
}
}
saver = UniqueNameSaver("../../temp/cga_heft_mixin")
def do_exp():
return do_experiment(saver, config, _wf, rm, estimator)
if __name__ == "__main__":
res = repeat(do_exp, 1)
print("RESULTS: ")
print(res)
"ga_params": {
"Kbest": 5,
"population": 50,
"crossover_probability": 0.3, #0.8
"replacing_mutation_probability": 0.1, #0.5
"sweep_mutation_probability": 0.3, #0.4
"generations": 100
},
"nodes_conf": [10, 15, 25, 30],
"transfer_time": 100,
"heft_initial": False
}
run = functools.partial(MixRunner(), **PARAMS)
directory = "../../temp/ga_vs_heft_exp"
saver = UniqueNameSaver("../../temp/ga_vs_heft_exp")
# def do_exp():
# ga_makespan, heft_makespan, ga_schedule, heft_schedule = run(wf_names[0])
# saver(ga_makespan)
# return ga_makespan
def do_exp_schedule(takeHeftSchedule=True):
saver = UniqueNameSaver("../../temp/ga_vs_heft_exp_heft_schedule")
ga_makespan, heft_makespan, ga_schedule, heft_schedule = run(wf_names[0])
## TODO: pure hack
schedule = heft_schedule if takeHeftSchedule else ga_schedule
"assign_credits": max_assign_credits
}
}
def do_experiment(saver, config, _wf, rm, estimator):
islands = [CoevolutionGA(**config), CoevolutionGA(**config),
CoevolutionGA(**config), CoevolutionGA(**config)]
# islands = [CoevolutionGA(**config), CoevolutionGA(**config)]
migration = partial(equal_social_migration_scheme, k=3, selection=best_selection)
best, islands = run_island_ga(islands, migration, 100, 20)
return best.fitness
saver = UniqueNameSaver("{0}/temp/icga_exp".format(__root_path__))
def do_exp():
res = do_experiment(saver, config, _wf, rm, estimator)
return res
if __name__ == "__main__":
res = repeat(do_exp, 1)
print("RESULTS: ")
print(res)
_wf = wf("Montage_100")
rm = ExperimentResourceManager(rg.r([10, 15, 25, 30]))
estimator = ExperimentEstimator(None, ideal_flops=20, transfer_time=100)
selector = ArchivedSelector(5)(tourn)
ms_ideal_ind = build_ms_ideal_ind(_wf, rm)
os_ideal_ind = build_os_ideal_ind(_wf)
os_representative = extract_ordering_from_ga_file(
"{0}/temp/ga_schedule_full_439_tr100_m100.json".format(__root_path__))
heft_mapping = extract_mapping_from_ga_file(
"{0}/temp/heft_etalon_full_tr100_m100.json".format(__root_path__), rm)
saver = UniqueNameSaver("../../temp/cga_fixed_ordering")
def do_exp():
config = {
"interact_individuals_count":
100,
"generations":
300,
"env":
Env(_wf, rm, estimator),
"species": [
Specie(
name=MAPPING_SPECIE,
pop_size=50,
cxb=0.9,
def do_exp():
saver = UniqueNameSaver("{0}/temp/cga_init_pop_bad".format(__root_path__))
return do_experiment(saver, cfg, _wf, rm, estimator)
def __init__(self, dir_name, experiment_name):
if os.path.isabs(dir_name):
path = dir_name
else:
path = os.path.join(TEMP_PATH, dir_name)
self._saver = UniqueNameSaver(path, experiment_name)
ctx, pop, heft_ordering, 3),
stat=lambda pop:
{"hamming_distances": hamming_distances(pop, os_ideal_ind)})
],
"solstat":
lambda sols: {
"best_components":
hamming_for_best_components(sols, ms_ideal_ind, os_ideal_ind),
"best_components_itself":
best_components_itself(sols)
},
"operators": {
# "choose": default_choose,
"build_solutions": default_build_solutions,
# "fitness": fitness_mapping_and_ordering,
"fitness": overhead_fitness_mapping_and_ordering,
"assign_credits": default_assign_credits
}
}
saver = UniqueNameSaver("../../temp/cga_fixed_mapping")
def do_exp():
return do_experiment(saver, config, _wf, rm, estimator)
if __name__ == "__main__":
res = repeat(do_exp, 1)
print("RESULTS: ")
print(res)