From: Szilárd Páll <pall.szilard@gmail.com>
Date: Tue, 23 Mar 2021 13:44:45 +0000 (+0000)
Subject: Fix deviceIdsAssigned() to return a unique list of IDs
X-Git-Url: http://biod.pnpi.spb.ru/gitweb/?a=commitdiff_plain;h=af5010c5ff9b6ea9f5dad7836a9887c24a86afed;p=alexxy%2Fgromacs.git

Fix deviceIdsAssigned() to return a unique list of IDs

The method originally returned the rank's task-to-device mapping
rather than the unique list of devices that have been assigned tasks.
When #ranks > #devices this list contains duplicate device IDs.

Due to the changes made in e2a2fe80 the vector returned by
deviceIdsAssigned() was passed to setupGpuDevicePeerAccess()
which results in redundant attempts to repeatedly enable peer access for
the same device pairs.

Refs #3980
---

diff --git a/src/gromacs/taskassignment/taskassignment.cpp b/src/gromacs/taskassignment/taskassignment.cpp
index a9256ba27c..34acebbd26 100644
--- a/src/gromacs/taskassignment/taskassignment.cpp
+++ b/src/gromacs/taskassignment/taskassignment.cpp
@@ -265,7 +265,7 @@ GpuTaskAssignments GpuTaskAssignmentsBuilder::build(const gmx::ArrayRef<const in
 
     std::exception_ptr             exceptionPtr;
     std::vector<GpuTaskAssignment> taskAssignmentOnRanksOfThisNode;
-    std::vector<int>               deviceIdsAssigned;
+    std::vector<int>               deviceIdAssignment;
     try
     {
         // Use the GPU IDs from the user if they supplied
@@ -320,7 +320,7 @@ GpuTaskAssignments GpuTaskAssignmentsBuilder::build(const gmx::ArrayRef<const in
                         host,
                         availableDevices.size())));
             }
-            deviceIdsAssigned = generatedGpuIds;
+            deviceIdAssignment = generatedGpuIds;
         }
         else
         {
@@ -342,10 +342,10 @@ GpuTaskAssignments GpuTaskAssignmentsBuilder::build(const gmx::ArrayRef<const in
             // Did the user choose compatible GPUs?
             checkUserGpuIds(hardwareInfo.deviceInfoList, availableDevices, userGpuTaskAssignment);
 
-            deviceIdsAssigned = gmx::copyOf(userGpuTaskAssignment);
+            deviceIdAssignment = gmx::copyOf(userGpuTaskAssignment);
         }
         taskAssignmentOnRanksOfThisNode =
-                buildTaskAssignment(gpuTasksOnRanksOfThisNode, deviceIdsAssigned);
+                buildTaskAssignment(gpuTasksOnRanksOfThisNode, deviceIdAssignment);
     }
     catch (...)
     {
@@ -392,7 +392,11 @@ GpuTaskAssignments GpuTaskAssignmentsBuilder::build(const gmx::ArrayRef<const in
     gpuTaskAssignments.indexOfThisRank_                 = physicalNodeComm.rank_;
     gpuTaskAssignments.numGpuTasksOnThisNode_           = numGpuTasksOnThisNode;
     gpuTaskAssignments.numRanksOnThisNode_              = numRanksOnThisNode;
-    gpuTaskAssignments.deviceIdsAssigned_               = deviceIdsAssigned;
+    gpuTaskAssignments.deviceIdsAssigned_               = deviceIdAssignment;
+    std::sort(gpuTaskAssignments.deviceIdsAssigned_.begin(), gpuTaskAssignments.deviceIdsAssigned_.end());
+    gpuTaskAssignments.deviceIdsAssigned_.erase(unique(gpuTaskAssignments.deviceIdsAssigned_.begin(),
+                                                       gpuTaskAssignments.deviceIdsAssigned_.end()),
+                                                gpuTaskAssignments.deviceIdsAssigned_.end());
     return gpuTaskAssignments;
 }
 
diff --git a/src/gromacs/taskassignment/taskassignment.h b/src/gromacs/taskassignment/taskassignment.h
index 63cf9c54ff..1267de148a 100644
--- a/src/gromacs/taskassignment/taskassignment.h
+++ b/src/gromacs/taskassignment/taskassignment.h
@@ -256,7 +256,7 @@ public:
     bool thisRankHasPmeGpuTask() const;
     //! Return whether this rank has any task running on a GPU
     bool thisRankHasAnyGpuTask() const;
-    //! Get the list of devices assigned to this node
+    //! Get the list of unique devices that have been assigned tasks on this physical node
     std::vector<int> deviceIdsAssigned() { return deviceIdsAssigned_; }
 };