Fix the by_open performance calculation so the close timestamp must be

greater than the open timestamp. (occurs if file is never closed)
Fix performance calculation so there's no divide by zero.
Fix handling of --total for ACCESS and STRIDE counters.


git-svn-id: https://svn.mcs.anl.gov/repos/darshan/trunk@415 3b7491f3-a168-0410-bf4b-c445ed680a29

darshan-parser.c

@@ -543,15 +543,26 @@ void accum_file(struct darshan_file *dfile,
             break;
         case CP_MAX_BYTE_READ:
         case CP_MAX_BYTE_WRITTEN:
-            if (hfile->counters[i] == -1)
-            {
-                hfile->counters[i] = dfile->counters[i];
-            }
             if (hfile->counters[i] < dfile->counters[i])
             {
                 hfile->counters[i] = dfile->counters[i];
             }
             break;
+
+        case CP_STRIDE1_STRIDE:
+        case CP_STRIDE2_STRIDE:
+        case CP_STRIDE3_STRIDE:
+        case CP_STRIDE4_STRIDE:
+        case CP_ACCESS1_ACCESS:
+        case CP_ACCESS2_ACCESS:
+        case CP_ACCESS3_ACCESS:
+        case CP_ACCESS4_ACCESS:
+           /*
+            * do nothing here because these will be stored
+            * when the _COUNT is accessed.
+            */
+           break;
+ 
         case CP_STRIDE1_COUNT:
         case CP_STRIDE2_COUNT:
         case CP_STRIDE3_COUNT:
@@ -560,10 +571,6 @@ void accum_file(struct darshan_file *dfile,
         case CP_ACCESS2_COUNT:
         case CP_ACCESS3_COUNT:
         case CP_ACCESS4_COUNT:
-            if (hfile->counters[i] == -1)
-            {
-                hfile->counters[i] = dfile->counters[i];
-            }
             if (hfile->counters[i] < dfile->counters[i])
             {
                 hfile->counters[i]   = dfile->counters[i];
@@ -726,9 +733,13 @@ void accum_perf(struct darshan_file *dfile,
     if (dfile->rank == -1)
     {
         /* by_open (same for MPI or POSIX) */
-        pdata->shared_time_by_open +=
-            dfile->fcounters[CP_F_CLOSE_TIMESTAMP] -
-            dfile->fcounters[CP_F_OPEN_TIMESTAMP];
+        if (dfile->fcounters[CP_F_CLOSE_TIMESTAMP] >
+            dfile->fcounters[CP_F_OPEN_TIMESTAMP])
+        {
+            pdata->shared_time_by_open +=
+                dfile->fcounters[CP_F_CLOSE_TIMESTAMP] -
+                dfile->fcounters[CP_F_OPEN_TIMESTAMP];
+        }
 
         /* by_open_lastio (same for MPI or POSIX) */
         if (dfile->fcounters[CP_F_READ_END_TIMESTAMP] >
@@ -841,19 +852,25 @@ void calc_perf(struct darshan_job *djob,
         }
     }
 
+    if (pdata->slowest_rank_time + pdata->shared_time_by_cumul)
     pdata->agg_perf_by_cumul = ((double)pdata->total_bytes / 1048576.0) /
-                               (pdata->slowest_rank_time +
-                                pdata->shared_time_by_cumul);
+                                  (pdata->slowest_rank_time +
+                                   pdata->shared_time_by_cumul);
+
+    if (pdata->slowest_rank_time + pdata->shared_time_by_open)
     pdata->agg_perf_by_open  = ((double)pdata->total_bytes / 1048576.0) / 
-                               (pdata->slowest_rank_time +
-                                pdata->shared_time_by_open);
+                                   (pdata->slowest_rank_time +
+                                    pdata->shared_time_by_open);
+
+    if (pdata->slowest_rank_time + pdata->shared_time_by_open_lastio)
     pdata->agg_perf_by_open_lastio = ((double)pdata->total_bytes / 1048576.0) /
                                      (pdata->slowest_rank_time +
                                       pdata->shared_time_by_open_lastio);
+
     if (pdata->slowest_rank_time + pdata->shared_time_by_slowest)
     pdata->agg_perf_by_slowest = ((double)pdata->total_bytes / 1048576.0) /
-                                 (pdata->slowest_rank_time +
-                                  pdata->shared_time_by_slowest);
+                                     (pdata->slowest_rank_time +
+                                      pdata->shared_time_by_slowest);
 
     return;
 }