Replace memset with fill to work for non-trivial scalars.

For custom scalars, zero is not necessarily represented by
a zeroed-out memory block (e.g. gnu MPFR). We therefore
cannot rely on `memset` if we want to fill a matrix or tensor
with zeroes. Instead, we should rely on `fill`, which for trivial
types does end up getting converted to a `memset` under-the-hood
(at least with gcc/clang).

Requires adding a `fill(begin, end, v)` to `TensorDevice`.

Replaced all potentially bad instances of memset with fill.

Fixes #2245.

unsupported/Eigen/CXX11/src/Tensor/TensorDeviceThreadPool.h

@@ -122,6 +122,11 @@ struct ThreadPoolDevice {
     ::memset(buffer, c, n);
   }
 
+  template<typename T>
+  EIGEN_STRONG_INLINE void fill(T* begin, T* end, const T& value) const {
+    std::fill(begin, end, value);
+  }
+
   EIGEN_STRONG_INLINE int numThreads() const {
     return num_threads_;
   }