Commented out some minor optimizations in AMOEBA that were making the forces...

Commented out some minor optimizations in AMOEBA that were making the forces incorrect.  AMOEBA energies and forces now agree with TINKER to within 1e-9 (fractional energy) 1e-5 (fractional RMS force error) for NPT simulations on rectangular water boxes with PME and direct or mutual polarizability.

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

@@ -529,7 +529,8 @@ extern "C" __global__ void computeFixedField(
 #endif
 #ifdef USE_CUTOFF
                 unsigned int flags = (numTiles <= maxTiles ? interactionFlags[pos] : 0xFFFFFFFF);
-                if (!hasExclusions && flags != 0xFFFFFFFF) {
+                if (!hasExclusions && flags == 0) { // TODO: Why doesn't the flags != 0 block work?
+//                if (!hasExclusions && flags != 0xFFFFFFFF) {
                     if (flags == 0) {
                         // No interactions in this tile.
                     }

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

@@ -298,7 +298,8 @@ extern "C" __global__ void computeElectrostatics(
                 localData[threadIdx.x].torque = make_real3(0);
 #ifdef USE_CUTOFF
                 unsigned int flags = (numTiles <= maxTiles ? interactionFlags[pos] : 0xFFFFFFFF);
-                if (!hasExclusions && flags != 0xFFFFFFFF) {
+                if (!hasExclusions && flags == 0) { // TODO: Why doesn't the flags != 0 block work?
+//                if (!hasExclusions && flags != 0xFFFFFFFF) {
                     if (flags == 0) {
                         // No interactions in this tile.
                     }