Merge pull request nugraph#3 from wkliao/scalars

adjust for scalar variables

numl/core/out.py

@@ -31,7 +31,10 @@ def __init__(self, fname):
   def save(self, obj, name):
     for key, val in obj:
       # set chunk sizes to val shape, so there is only one chunk per dataset
-      self.f.create_dataset(f"/{name}/{key}", data=val, chunks=val.shape, compression="gzip")
+      if (isinstance(val, torch.Tensor)) :
+        self.f.create_dataset(f"/{name}/{key}", data=val, chunks=val.shape, compression="gzip")
+      else:
+        self.f.create_dataset(f"/{name}/{key}", data=val)
       # below is to disable data compression (and chunking)
       # self.f.create_dataset(f"/{name}/{key}", data=val)
 

numl/process/hitgraph.py

@@ -7,6 +7,8 @@
 import numpy as np
 import sys
 
+edep1_t = 0.0
+edep2_t = 0.0
 hit_merge_t = 0.0
 torch_t = 0.0
 plane_t = 0.0
@@ -312,8 +314,12 @@ def process_file(out, fname, g=process_event, l=standard.semantic_label,
         if profiling:
           grp_size = 0
           for key, val in data:
-            # calculate size in bytes of val, a pytorch tensor
-            grp_size += val.element_size() * val.nelement()
+            # calculate size in bytes of val
+            if (isinstance(val, torch.Tensor)):
+              # val is a pytorch tensor
+              grp_size += val.element_size() * val.nelement()
+            else:
+              grp_size += sys.getsizeof(val)
           num_grps += 1
           grp_size_sum += grp_size
           if grp_size > grp_size_max : grp_size_max = grp_size
@@ -406,4 +412,4 @@ def process_file(out, fname, g=process_event, l=standard.semantic_label,
       print("graph creation  time MAX=%8.2f  MIN=%8.2f" % (max_total_t[2], min_total_t[2]))
       print("write to files  time MAX=%8.2f  MIN=%8.2f" % (max_total_t[3], min_total_t[3]))
       print("total           time MAX=%8.2f  MIN=%8.2f" % (max_total_t[4], min_total_t[4]))
-      print("(MAX and MIN timings are among %d processes)" % nprocs)
+      print("(MAX and MIN timings are among %d processes)" % nprocs)