Fix PME load balancing skipping a setup

During stage 0 of the PME tuning we only time every second setup.
In stage 1 only a reduced range of setups is considered. If the tpr
setup (setup index 0) was not part of this reduced range, the first
setup in the range would never be timed. This could lead to choosing
a non-bonded cut-off that was longer than optimal.

Fixes #1880

Also fixed two conditionals (but this had no effect on the settings).

Change-Id: Id9f361d23e3294e41553b39ae8e8f37ceb80bbe5

diff --git a/src/gromacs/ewald/pme-load-balancing.cpp b/src/gromacs/ewald/pme-load-balancing.cpp
index 3863b9bd8b930fc5112294bf00ed94b881165ccc..ac7d0af4c3ba07748f8d46726b86f4ed270d8f8d 100644 (file)
--- a/src/gromacs/ewald/pme-load-balancing.cpp
+++ b/src/gromacs/ewald/pme-load-balancing.cpp
@@ -1,7 +1,7 @@
 /*
  * This file is part of the GROMACS molecular simulation package.
  *
- * Copyright (c) 2012,2013,2014,2015, by the GROMACS development team, led by
+ * Copyright (c) 2012,2013,2014,2015,2016, by the GROMACS development team, led by
  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
  * and including many others, as listed in the AUTHORS file in the
  * top-level source directory and at http://www.gromacs.org.
@@ -504,6 +504,9 @@ static void print_loadbal_limited(FILE *fp_err, FILE *fp_log,
  * In this stage, only reasonably fast setups are run again. */
 static void switch_to_stage1(pme_load_balancing_t *pme_lb)
 {
+    /* Increase start until we find a setup that is not slower than
+     * maxRelativeSlowdownAccepted times the fastest setup.
+     */
     pme_lb->start = pme_lb->lower_limit;
     while (pme_lb->start + 1 < pme_lb->n &&
            (pme_lb->setup[pme_lb->start].count == 0 ||
@@ -512,11 +515,18 @@ static void switch_to_stage1(pme_load_balancing_t *pme_lb)
     {
         pme_lb->start++;
     }
-    while (pme_lb->start > 0 && pme_lb->setup[pme_lb->start - 1].cycles == 0)
+    /* While increasing start, we might have skipped setups that we did not
+     * time during stage 0. We want to extend the range for stage 1 to include
+     * any skipped setups that lie between setups that were measured to be
+     * acceptably fast and too slow.
+     */
+    while (pme_lb->start > pme_lb->lower_limit &&
+           pme_lb->setup[pme_lb->start - 1].count == 0)
     {
         pme_lb->start--;
     }
 
+    /* Decrease end only with setups that we timed and that are slow. */
     pme_lb->end = pme_lb->n;
     if (pme_lb->setup[pme_lb->end - 1].count > 0 &&
         pme_lb->setup[pme_lb->end - 1].cycles >
@@ -528,7 +538,7 @@ static void switch_to_stage1(pme_load_balancing_t *pme_lb)
     pme_lb->stage = 1;
 
     /* Next we want to choose setup pme_lb->end-1, but as we will decrease
-     * pme_ln->cur by one right after returning, we set cur to end.
+     * pme_lb->cur by one right after returning, we set cur to end.
      */
     pme_lb->cur = pme_lb->end;
 }
@@ -736,8 +746,11 @@ pme_load_balance(pme_load_balancing_t      *pme_lb,
          */
         do
         {
-            pme_lb->cur--;
-            if (pme_lb->cur == pme_lb->start)
+            if (pme_lb->cur > pme_lb->start)
+            {
+                pme_lb->cur--;
+            }
+            else
             {
                 pme_lb->stage++;