diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorExecutor.h b/unsupported/Eigen/CXX11/src/Tensor/TensorExecutor.h
index 6bbf235cc..f71625ae5 100644
--- a/unsupported/Eigen/CXX11/src/Tensor/TensorExecutor.h
+++ b/unsupported/Eigen/CXX11/src/Tensor/TensorExecutor.h
@@ -125,7 +125,7 @@ class TensorExecutor<Expression, ThreadPoolDevice, Vectorizable>
 
       int blocksz = std::ceil<int>(static_cast<float>(size)/device.numThreads()) + PacketSize - 1;
       const Index blocksize = numext::maxi<Index>(PacketSize, (blocksz - (blocksz % PacketSize)));
-      const Index numblocks = size / blocksize;
+      const unsigned int numblocks = static_cast<unsigned int>(size / blocksize);
 
       Barrier barrier(numblocks);
       for (int i = 0; i < numblocks; ++i) {
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h b/unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h
index 489451215..0ce2517d6 100644
--- a/unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h
+++ b/unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h
@@ -253,7 +253,7 @@ struct FullReducer<Self, Op, ThreadPoolDevice, false> {
       return;
     } else {
       const Index blocksize = std::floor<Index>(static_cast<float>(num_coeffs) / num_threads);
-      const Index numblocks = blocksize > 0 ? num_coeffs / blocksize : 0;
+      const unsigned int numblocks = blocksize > 0 ? static_cast<unsigned int>(num_coeffs / blocksize) : 0;
       eigen_assert(num_coeffs >= numblocks * blocksize);
 
       Barrier barrier(numblocks);
@@ -300,7 +300,7 @@ struct FullReducer<Self, Op, ThreadPoolDevice, true> {
       return;
     }
     const Index blocksize = std::floor<Index>(static_cast<float>(num_coeffs) / num_threads);
-    const Index numblocks = blocksize > 0 ? num_coeffs / blocksize : 0;
+    const unsigned int numblocks = blocksize > 0 ? static_cast<unsigned int>(num_coeffs / blocksize) : 0;
     eigen_assert(num_coeffs >= numblocks * blocksize);
 
     Barrier barrier(numblocks);