Added loop over particles for torque mapping

plugins/amoeba/platforms/cuda/src/kernels/kCalculateAmoebaCudaPmeDirectElectrostatic.cu

@@ -1089,8 +1089,8 @@ static void kReduceTorque(amoebaGpuContext amoebaGpu )
 {
     gpuContext gpu = amoebaGpu->gpuContext;
     kReduceFields_kernel<<<gpu->sim.nonbond_blocks, gpu->sim.bsf_reduce_threads_per_block>>>(
-                             gpu->sim.paddedNumberOfAtoms*3, gpu->sim.outputBuffers,
-                             amoebaGpu->psWorkArray_3_1->_pDevData, amoebaGpu->psTorque->_pDevData );
+                           gpu->sim.paddedNumberOfAtoms*3, gpu->sim.outputBuffers,
+                           amoebaGpu->psWorkArray_3_1->_pDevData, amoebaGpu->psTorque->_pDevData );
     LAUNCHERROR("kReducePmeDirectElectrostaticTorque");
 }
 

plugins/amoeba/platforms/cuda/src/kernels/kCalculateAmoebaCudaRotateFrame.cu

@@ -147,7 +147,6 @@ void kCudaComputeLabFrameMoments_kernel( void )
     float vectorZ[3];
  
     int particleIndex            = __mul24(blockIdx.x,blockDim.x) + threadIdx.x;
-    int numberOfParticles        = cSim.atoms;
 
     float4* particleCoord        = cSim.pPosq;
     int4* multiPoleParticles     = cAmoebaSim.pMultipoleParticlesIdsAndAxisType;
@@ -162,7 +161,7 @@ void kCudaComputeLabFrameMoments_kernel( void )
  
     // code common to ZThenX and Bisector
     
-    while( particleIndex < numberOfParticles )
+    while( particleIndex < cSim.atoms )
     {
 
         if( multiPoleParticles[particleIndex].x >= 0 && multiPoleParticles[particleIndex].z >= 0 )