Skip to content

Commit

Permalink
Update hiphive.py
Browse files Browse the repository at this point in the history
  • Loading branch information
@hrushikesh-s
hrushikesh-s committed Oct 27, 2023
1 parent adfd241 commit 9f61b69
Showing 1 changed file with 82 additions and 81 deletions.
163 changes: 82 additions & 81 deletions src/atomate2/vasp/jobs/hiphive.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,7 @@
from itertools import product
from pathlib import Path
from typing import Any, Dict, List, Optional, Tuple, Union
from joblib import Parallel, delayed

import numpy as np
import phonopy as phpy
Expand Down Expand Up @@ -995,10 +996,10 @@ def fit_force_constants(
logger.info("We are starting Joblib_s parallellized jobs")

# With Joblib's parallellization
# cutoff_results = Parallel(n_jobs=-1, backend="multiprocessing")
# (delayed(_run_cutoffs)(i, cutoffs, n_cutoffs, parent_structure, structures,
# supercell_matrix, fit_method,disp_cut, imaginary_tol, fit_kwargs) for
# i, cutoffs in enumerate(all_cutoffs))
cutoff_results = Parallel(n_jobs=-1, backend="multiprocessing")
(delayed(_run_cutoffs)(i, cutoffs, n_cutoffs, parent_structure, structures,
supercell_matrix, fit_method,disp_cut, imaginary_tol, fit_kwargs) for
i, cutoffs in enumerate(all_cutoffs))
# cutoff_results = Parallel(
# n_jobs=min(os.cpu_count(), len(all_cutoffs)), backend="multiprocessing"
# )(
Expand All @@ -1017,90 +1018,90 @@ def fit_force_constants(
# for i, cutoffs in enumerate(all_cutoffs)
# )

# logger.info(f"CUTOFF RESULTS \n {cutoff_results}")
logger.info(f"CUTOFF RESULTS \n {cutoff_results}")

# for result in cutoff_results:
# if result is None:
# continue

# fitting_data["cutoffs"].append(result["cutoffs"])
# fitting_data["rmse_test"].append(result["rmse_test"])
# # fitting_data["n_imaginary"].append(result["n_imaginary"])
# # fitting_data["min_frequency"].append(result["min_frequency"])

# if (
# result["rmse_test"]
# < best_fit["rmse_test"]
# # and result["min_frequency"] > -np.abs(max_imaginary_freq)
# # and result["n_imaginary"] <= max_n_imaginary
# # and result["n_imaginary"] < best_fit["n_imaginary"]
# ):
# best_fit.update(result)
# fitting_data["best"] = result["cutoffs"]

# logger.info("Finished fitting force constants.")

# print(f'all_cutoffs={all_cutoffs}')

# Without Joblib's parallellization
for i, cutoffs in enumerate(all_cutoffs):
print(f"disp_cut={disp_cut}")
print(f"fit_method={fit_method}")
print(f"cuttoffs={cutoffs}")
print(f"imaginary_tol={imaginary_tol}")
print(f"n_cutoffs={n_cutoffs}")
print(f"parent_structure={parent_structure}")
print(f"structures={structures}")
print(f"supercell_matrix={supercell_matrix}")

start_time = time.time()
cutoff_results = _run_cutoffs(
i,
cutoffs,
n_cutoffs,
parent_structure,
structures,
supercell_matrix,
fit_method,
disp_cut,
imaginary_tol,
fit_kwargs,
)
time_taken = time.time() - start_time
logger.info(f"Time taken for cutoffs {cutoffs} is {time_taken} seconds")
for result in cutoff_results:
if result is None:
continue

fitting_data["cutoffs"].append(result["cutoffs"])
fitting_data["rmse_test"].append(result["rmse_test"])
# fitting_data["n_imaginary"].append(result["n_imaginary"])
# fitting_data["min_frequency"].append(result["min_frequency"])

if (
result["rmse_test"]
< best_fit["rmse_test"]
# and result["min_frequency"] > -np.abs(max_imaginary_freq)
# and result["n_imaginary"] <= max_n_imaginary
# and result["n_imaginary"] < best_fit["n_imaginary"]
):
best_fit.update(result)
fitting_data["best"] = result["cutoffs"]

logger.info("Finished fitting force constants.")

print(f'all_cutoffs={all_cutoffs}')

# # Without Joblib's parallellization
# for i, cutoffs in enumerate(all_cutoffs):
# print(f"disp_cut={disp_cut}")
# print(f"fit_method={fit_method}")
# print(f"cuttoffs={cutoffs}")
# print(f"imaginary_tol={imaginary_tol}")
# print(f"n_cutoffs={n_cutoffs}")
# print(f"parent_structure={parent_structure}")
# print(f"structures={structures}")
# print(f"supercell_matrix={supercell_matrix}")

# start_time = time.time()
# cutoff_results = _run_cutoffs(
# i,
# cutoffs,
# n_cutoffs,
# parent_structure,
# structures,
# supercell_matrix,
# fit_method,
# disp_cut,
# imaginary_tol,
# fit_kwargs,
# )
# time_taken = time.time() - start_time
# logger.info(f"Time taken for cutoffs {cutoffs} is {time_taken} seconds")

print(f"cutoff_results={cutoff_results}")
print(f"time taken={time_taken}")
print(f'cutoffs={cutoff_results["cutoffs"]}')
print(f'rmse_test={cutoff_results["rmse_test"]}')
print(f'parameters={cutoff_results["parameters"]}')
# print(f"cutoff_results={cutoff_results}")
# print(f"time taken={time_taken}")
# print(f'cutoffs={cutoff_results["cutoffs"]}')
# print(f'rmse_test={cutoff_results["rmse_test"]}')
# print(f'parameters={cutoff_results["parameters"]}')

with open(f"timings_{cutoffs}_{fit_method}.csv", mode="a", newline="") as file:
writer = csv.writer(file)
writer.writerow([cutoffs, time_taken, cutoff_results["rmse_test"]])
# with open(f"timings_{cutoffs}_{fit_method}.csv", mode="a", newline="") as file:
# writer = csv.writer(file)
# writer.writerow([cutoffs, time_taken, cutoff_results["rmse_test"]])

cutoff_results["force_constants_potential"].write(
f"force_constants_potential_{cutoffs}_{fit_method}.fcp"
)
cutoff_results["force_constants"].write(
f"force_constants_potential_{cutoffs}_{fit_method}.fcs"
)
# cutoff_results["force_constants_potential"].write(
# f"force_constants_potential_{cutoffs}_{fit_method}.fcp"
# )
# cutoff_results["force_constants"].write(
# f"force_constants_potential_{cutoffs}_{fit_method}.fcs"
# )

if cutoff_results["rmse_test"] < best_fit["rmse_test"]:
best_fit["rmse_test"] = cutoff_results["rmse_test"]
best_fit["cluster_space"] = cutoff_results["cluster_space"]
best_fit["force_constants"] = cutoff_results["force_constants"]
best_fit["parameters"] = cutoff_results["parameters"]
best_fit["cutoffs"] = cutoff_results["cutoffs"]
best_fit["force_constants_potential"] = cutoff_results[
"force_constants_potential"
]
fitting_data["best"] = cutoff_results["cutoffs"]
# if cutoff_results["rmse_test"] < best_fit["rmse_test"]:
# best_fit["rmse_test"] = cutoff_results["rmse_test"]
# best_fit["cluster_space"] = cutoff_results["cluster_space"]
# best_fit["force_constants"] = cutoff_results["force_constants"]
# best_fit["parameters"] = cutoff_results["parameters"]
# best_fit["cutoffs"] = cutoff_results["cutoffs"]
# best_fit["force_constants_potential"] = cutoff_results[
# "force_constants_potential"
# ]
# fitting_data["best"] = cutoff_results["cutoffs"]

fitting_data["cutoffs"].append(cutoff_results["cutoffs"])
fitting_data["rmse_test"].append(cutoff_results["rmse_test"])
# fitting_data["cutoffs"].append(cutoff_results["cutoffs"])
# fitting_data["rmse_test"].append(cutoff_results["rmse_test"])

logger.info(f"CUTOFF RESULTS \n {cutoff_results}")
# logger.info(f"CUTOFF RESULTS \n {cutoff_results}")

return (
best_fit["force_constants"],
Expand Down

0 comments on commit 9f61b69

Please sign in to comment.