performance analysis script with atomic limit data

Author: aeantipov

scripts/gen_data_atomic.py

@@ -0,0 +1,82 @@
+import os
+import numpy as np
+import subprocess
+import time
+import distutils
+from distutils.spawn import find_executable
+import sys
+import shutil
+import itertools
+import argparse
+import collections
+
+#nbos_array = [1, 5, 16]
+nbos_array = [1,2,3,4,5,6,8,12,16]
+#nbos_array = [8,12,16]
+nbos_array = [3]
+#nbos_array = [1,2,3,4,5,8,16]
+#nbos_array = [6,5,4,3,2,1][::-1]
+#nferm_array = [16,32,48,64]
+nferm_array = [4,8,16,24,32,48,64,96,128,160]
+#nferm_array=[16,24,32,48,64]
+#nferm_array=[4,8,12]#16,24,32,48,64]
+#nferm_array = [48]
+#kpts_array = [4, 8, 16, 32]#, 32]
+#kpts_array = [4,8,16,24,32,48,64]
+kpts_array =[16]
+
+U=20.0
+beta = 1.0
+
+dry_run = False # True
+
+inp_section="atomic"
+data_dir="data_atomic"
+resume=False
+plaintext=0
+df_sc_iter = 150
+df_sc_cutoff = 1e-8
+
+def main():
+    counter = 0
+    origdir = os.path.abspath(os.getcwd())
+    for nbos in nbos_array:
+            for nferm in nferm_array:
+                for kpts in kpts_array:
+                    exec_dir=origdir + os.path.sep+data_dir+os.path.sep + "b" + str(nbos) + os.path.sep + "f" + str(nferm) + os.path.sep + "k" + str(kpts)
+                    os.makedirs(exec_dir) if not os.path.exists(exec_dir) else None
+                    os.chdir(exec_dir)
+                    
+                    prepare_string = ["generate_atomic", 
+                        "--nfermionic", str(nferm),
+                        "--nbosonic", str(nbos),
+                        "--U", str(U),
+                        "--beta", str(beta),
+                        "--output", "atomic.h5",
+                        "--plaintext", str(int(plaintext))
+                    ] 
+                    df_string = ["hub_df_cubic2d", 
+                        "--input", "atomic.h5", 
+                        "--output", "output.h5", 
+                        "--inp_section", inp_section,
+                        "--df_sc_iter", str(df_sc_iter), 
+                        "--df_sc_cutoff", str(df_sc_cutoff),
+                        "--nbosonic", str(nbos),
+                        "--resume", str(int(resume)),
+                        "--kpts", str(kpts),
+                        "--plaintext", str(int(plaintext))
+                    ]
+                    if not dry_run:
+                        print " ".join(prepare_string)
+                        print subprocess.call(" ".join(prepare_string), shell=True)
+                        print " ".join(df_string)
+                        print subprocess.call(" ".join(df_string), shell=True)
+                        
+                    counter = counter + 1
+
+                    os.chdir(origdir)
+
+    print counter, "calcs ran"
+
+if __name__ == "__main__":
+    main()