diff --git a/common/cuda_hip/factorization/par_ic_kernels.hpp.inc b/common/cuda_hip/factorization/par_ic_kernels.hpp.inc
index 569c2d23157..57b8189123a 100644
--- a/common/cuda_hip/factorization/par_ic_kernels.hpp.inc
+++ b/common/cuda_hip/factorization/par_ic_kernels.hpp.inc
@@ -44,6 +44,9 @@ __global__ __launch_bounds__(default_block_size) void ic_init(
     }
     auto l_nz = l_row_ptrs[row + 1] - 1;
     auto diag = sqrt(l_vals[l_nz]);
+    printf("%d %lf %lf \n", static_cast<int>(row),
+           static_cast<double>(real(l_vals[l_nz])),
+           static_cast<double>(real(diag)));
     if (is_finite(diag)) {
         l_vals[l_nz] = diag;
     } else {
diff --git a/core/base/extended_float.hpp b/core/base/extended_float.hpp
index bf79c75b519..6898c4ffaa2 100644
--- a/core/base/extended_float.hpp
+++ b/core/base/extended_float.hpp
@@ -54,6 +54,8 @@ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #include <hip/hip_fp16.h>
 
 
+#else
+class __half;
 #endif  // __CUDA_ARCH__
 
 
@@ -101,7 +103,7 @@ struct basic_float_traits<float16> {
     static constexpr bool rounds_to_nearest = true;
 };
 
-#if defined(__CUDA_ARCH__) || defined(__HIP_DEVICE_COMPILE__)
+// #if defined(__CUDA_ARCH__) || defined(__HIP_DEVICE_COMPILE__)
 template <>
 struct basic_float_traits<__half> {
     using type = __half;
@@ -110,7 +112,7 @@ struct basic_float_traits<__half> {
     static constexpr int exponent_bits = 5;
     static constexpr bool rounds_to_nearest = true;
 };
-#endif
+// #endif
 
 template <>
 struct basic_float_traits<float32> {
@@ -598,17 +600,17 @@ class complex<gko::half> {
     value_type imag() const noexcept { return imag_; }
 
 
-    operator std::complex<gko::float32>() const noexcept
+    operator std::complex<float>() const noexcept
     {
-        return std::complex<gko::float32>(static_cast<gko::float32>(real_),
-                                          static_cast<gko::float32>(imag_));
+        return std::complex<float>(static_cast<float>(real_),
+                                   static_cast<float>(imag_));
     }
 
-    operator std::complex<double>() const noexcept
-    {
-        return std::complex<double>(static_cast<double>(real_),
-                                    static_cast<double>(imag_));
-    }
+    // operator std::complex<double>() const noexcept
+    // {
+    //     return std::complex<double>(static_cast<double>(real_),
+    //                                 static_cast<double>(imag_));
+    // }
 
     template <typename V>
     complex& operator=(const V& val)
diff --git a/core/preconditioner/jacobi.cpp b/core/preconditioner/jacobi.cpp
index 5f8d194fcd6..75f5e941303 100644
--- a/core/preconditioner/jacobi.cpp
+++ b/core/preconditioner/jacobi.cpp
@@ -319,7 +319,9 @@ void Jacobi<ValueType, IndexType>::generate(const LinOp* system_matrix,
                               ->extract_diagonal_linop());
         auto diag_vt =
             ::gko::detail::temporary_conversion<matrix::Diagonal<ValueType>>::
-                template create<matrix::Diagonal<previous_precision<ValueType>>>(
+                template create<matrix::Diagonal<previous_precision<ValueType>>,
+                                matrix::Diagonal<previous_precision<
+                                    previous_precision<ValueType>>>>(
                     diag.get());
         if (!diag_vt) {
             GKO_NOT_SUPPORTED(system_matrix);
diff --git a/hip/base/types.hip.hpp b/hip/base/types.hip.hpp
index df867993551..4f40b81ad47 100644
--- a/hip/base/types.hip.hpp
+++ b/hip/base/types.hip.hpp
@@ -122,9 +122,9 @@ __device__ __forceinline__ __half abs(const __half& val) { return __habs(val); }
 
 #endif
 
-#if defined(__HIPCC__)
+// #if defined(__HIPCC__)
 __device__ __forceinline__ float sqrt(float val) { return sqrtf(val); }
-__device__ __forceinline__ double sqrt(double val) { return sqrt(val); }
+// __device__ __forceinline__ double sqrt(double val) { return sqrt(val); }
 __device__ __forceinline__ thrust::complex<float> sqrt(
     thrust::complex<float> val)
 {
@@ -144,7 +144,7 @@ __device__ __forceinline__ __half sqrt(__half val)
 #else
 __device__ __forceinline__ __half sqrt(__half val) { return hsqrt(val); }
 #endif
-#endif
+// #endif
 
 
 namespace kernels {
diff --git a/hip/preconditioner/jacobi_generate_instantiate.inc.hip.cpp b/hip/preconditioner/jacobi_generate_instantiate.inc.hip.cpp
index cb37fd3d0f2..6589033fb56 100644
--- a/hip/preconditioner/jacobi_generate_instantiate.inc.hip.cpp
+++ b/hip/preconditioner/jacobi_generate_instantiate.inc.hip.cpp
@@ -53,18 +53,18 @@ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 
 namespace gko {
-namespace detail {
-#if !defined(__HIP_DEVICE_COMPILE__)
-template <>
-struct basic_float_traits<__half> {
-    using type = __half;
-    static constexpr int sign_bits = 1;
-    static constexpr int significand_bits = 10;
-    static constexpr int exponent_bits = 5;
-    static constexpr bool rounds_to_nearest = true;
-};
-#endif
-}  // namespace detail
+// namespace detail {
+// #if !defined(__HIP_DEVICE_COMPILE__)
+// template <>
+// struct basic_float_traits<__half> {
+//     using type = __half;
+//     static constexpr int sign_bits = 1;
+//     static constexpr int significand_bits = 10;
+//     static constexpr int exponent_bits = 5;
+//     static constexpr bool rounds_to_nearest = true;
+// };
+// #endif
+// }  // namespace detail
 namespace kernels {
 namespace hip {
 /**
diff --git a/include/ginkgo/core/base/half.hpp b/include/ginkgo/core/base/half.hpp
new file mode 100644
index 00000000000..7ef38f68247
--- /dev/null
+++ b/include/ginkgo/core/base/half.hpp
@@ -0,0 +1,24 @@
+#ifndef GKO_BASE_HALF_HPP_
+#define GKO_BASE_HALF_HPP_
+#include <complex>
+#include <type_traits>
+
+
+#ifdef __CUDA_ARCH__
+
+
+#include <cuda_fp16.h>
+
+
+#elif defined(__HIP_DEVICE_COMPILE__)
+
+
+#include <hip/hip_fp16.h>
+
+
+#endif  // __CUDA_ARCH__
+
+
+namespace gko {}
+
+#endif  // GKO_BASE_HALF_HPP_
diff --git a/reference/matrix/ell_kernels.cpp b/reference/matrix/ell_kernels.cpp
index 1b2a1b4b243..802ab25699c 100644
--- a/reference/matrix/ell_kernels.cpp
+++ b/reference/matrix/ell_kernels.cpp
@@ -137,7 +137,8 @@ void advanced_spmv(std::shared_ptr<const ReferenceExecutor> exec,
 
     for (size_type j = 0; j < c->get_size()[1]; j++) {
         for (size_type row = 0; row < a->get_size()[0]; row++) {
-            arithmetic_type result = c->at(row, j);
+            arithmetic_type result =
+                static_cast<arithmetic_type>(c->at(row, j));
             result *= beta_val;
             for (size_type i = 0; i < num_stored_elements_per_row; i++) {
                 auto val = a_vals(row + i * stride);