diff --git a/samples/sgemm_cuda.cpp b/samples/sgemm_cuda.cpp
index ed2ad588..f1138316 100644
--- a/samples/sgemm_cuda.cpp
+++ b/samples/sgemm_cuda.cpp
@@ -19,7 +19,7 @@
 #include <vector>
 
 // Includes the CUDA driver API
-#include <cuda>
+#include <cuda.h>
 
 // Includes the CLBlast library
 #include <clblast_cuda.h>
@@ -43,14 +43,15 @@ int main() {
   const auto c_ld = n;
 
   // Initializes the OpenCL device
+  cuInit(0);
   CUdevice device;
   cuDeviceGet(&device, device_id);
 
   // Creates the OpenCL context and stream
   CUcontext context;
-  cuCtxCreate(context, 0, device);
+  cuCtxCreate(&context, 0, device);
   CUstream stream;
-  cuStreamCreate(queue, CU_STREAM_NON_BLOCKING);
+  cuStreamCreate(&stream, CU_STREAM_NON_BLOCKING);
 
   // Populate host matrices with some example data
   auto host_a = std::vector<float>(m*k);
@@ -64,12 +65,12 @@ int main() {
   CUdeviceptr device_a;
   CUdeviceptr device_b;
   CUdeviceptr device_c;
-  cuMemAlloc(device_a, host_a.size()*sizeof(float));
-  cuMemAlloc(device_b, host_b.size()*sizeof(float));
-  cuMemAlloc(device_c, host_c.size()*sizeof(float));
-  cuMemcpyHtoDAsync(device_a, host_a.data()), host_a.size()*sizeof(T), queue);
-  cuMemcpyHtoDAsync(device_b, host_c.data()), host_b.size()*sizeof(T), queue);
-  cuMemcpyHtoDAsync(device_c, host_b.data()), host_c.size()*sizeof(T), queue);
+  cuMemAlloc(&device_a, host_a.size()*sizeof(float));
+  cuMemAlloc(&device_b, host_b.size()*sizeof(float));
+  cuMemAlloc(&device_c, host_c.size()*sizeof(float));
+  cuMemcpyHtoDAsync(device_a, host_a.data(), host_a.size()*sizeof(float), stream);
+  cuMemcpyHtoDAsync(device_b, host_c.data(), host_b.size()*sizeof(float), stream);
+  cuMemcpyHtoDAsync(device_c, host_b.data(), host_c.size()*sizeof(float), stream);
 
   // Start the timer
   auto start_time = std::chrono::steady_clock::now();
@@ -79,11 +80,12 @@ int main() {
                               clblast::Transpose::kNo, clblast::Transpose::kNo,
                               m, n, k,
                               alpha,
-                              device_a(), 0, a_ld,
-                              device_b(), 0, b_ld,
+                              device_a, 0, a_ld,
+                              device_b, 0, b_ld,
                               beta,
-                              device_c(), 0, c_ld,
+                              device_c, 0, c_ld,
                               context, device);
+  cuStreamSynchronize(stream);
 
   // Record the execution time
   auto elapsed_time = std::chrono::steady_clock::now() - start_time;
diff --git a/scripts/generator/generator/cpp.py b/scripts/generator/generator/cpp.py
index 5413906a..2d18655f 100644
--- a/scripts/generator/generator/cpp.py
+++ b/scripts/generator/generator/cpp.py
@@ -56,7 +56,8 @@ def clblast_cc(routine, cuda=False):
             result += "    auto queue_cpp = Queue(context_cpp, device_cpp);" + NL
         else:
             result += "    auto queue_cpp = Queue(*queue);" + NL
-        result += "    auto routine = X" + routine.plain_name() + "<" + routine.template.template + ">(queue_cpp, event);" + NL
+        event = "nullptr" if cuda else "event"
+        result += "    auto routine = X" + routine.plain_name() + "<" + routine.template.template + ">(queue_cpp, " + event + ");" + NL
         if routine.batched:
             result += "    " + (NL + "    ").join(routine.batched_transform_to_cpp()) + NL
         result += "    routine.Do" + routine.capitalized_name() + "("
diff --git a/src/clblast_cuda.cpp b/src/clblast_cuda.cpp
index f9a24236..0e3d949d 100644
--- a/src/clblast_cuda.cpp
+++ b/src/clblast_cuda.cpp
@@ -120,7 +120,7 @@ StatusCode Swap(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xswap<T>(queue_cpp, event);
+    auto routine = Xswap<T>(queue_cpp, nullptr);
     routine.DoSwap(n,
                    Buffer<T>(x_buffer), x_offset, x_inc,
                    Buffer<T>(y_buffer), y_offset, y_inc);
@@ -158,7 +158,7 @@ StatusCode Scal(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xscal<T>(queue_cpp, event);
+    auto routine = Xscal<T>(queue_cpp, nullptr);
     routine.DoScal(n,
                    alpha,
                    Buffer<T>(x_buffer), x_offset, x_inc);
@@ -196,7 +196,7 @@ StatusCode Copy(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xcopy<T>(queue_cpp, event);
+    auto routine = Xcopy<T>(queue_cpp, nullptr);
     routine.DoCopy(n,
                    Buffer<T>(x_buffer), x_offset, x_inc,
                    Buffer<T>(y_buffer), y_offset, y_inc);
@@ -235,7 +235,7 @@ StatusCode Axpy(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xaxpy<T>(queue_cpp, event);
+    auto routine = Xaxpy<T>(queue_cpp, nullptr);
     routine.DoAxpy(n,
                    alpha,
                    Buffer<T>(x_buffer), x_offset, x_inc,
@@ -280,7 +280,7 @@ StatusCode Dot(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xdot<T>(queue_cpp, event);
+    auto routine = Xdot<T>(queue_cpp, nullptr);
     routine.DoDot(n,
                   Buffer<T>(dot_buffer), dot_offset,
                   Buffer<T>(x_buffer), x_offset, x_inc,
@@ -315,7 +315,7 @@ StatusCode Dotu(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xdotu<T>(queue_cpp, event);
+    auto routine = Xdotu<T>(queue_cpp, nullptr);
     routine.DoDotu(n,
                    Buffer<T>(dot_buffer), dot_offset,
                    Buffer<T>(x_buffer), x_offset, x_inc,
@@ -345,7 +345,7 @@ StatusCode Dotc(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xdotc<T>(queue_cpp, event);
+    auto routine = Xdotc<T>(queue_cpp, nullptr);
     routine.DoDotc(n,
                    Buffer<T>(dot_buffer), dot_offset,
                    Buffer<T>(x_buffer), x_offset, x_inc,
@@ -374,7 +374,7 @@ StatusCode Nrm2(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xnrm2<T>(queue_cpp, event);
+    auto routine = Xnrm2<T>(queue_cpp, nullptr);
     routine.DoNrm2(n,
                    Buffer<T>(nrm2_buffer), nrm2_offset,
                    Buffer<T>(x_buffer), x_offset, x_inc);
@@ -412,7 +412,7 @@ StatusCode Asum(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xasum<T>(queue_cpp, event);
+    auto routine = Xasum<T>(queue_cpp, nullptr);
     routine.DoAsum(n,
                    Buffer<T>(asum_buffer), asum_offset,
                    Buffer<T>(x_buffer), x_offset, x_inc);
@@ -450,7 +450,7 @@ StatusCode Sum(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xsum<T>(queue_cpp, event);
+    auto routine = Xsum<T>(queue_cpp, nullptr);
     routine.DoSum(n,
                   Buffer<T>(sum_buffer), sum_offset,
                   Buffer<T>(x_buffer), x_offset, x_inc);
@@ -488,7 +488,7 @@ StatusCode Amax(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xamax<T>(queue_cpp, event);
+    auto routine = Xamax<T>(queue_cpp, nullptr);
     routine.DoAmax(n,
                    Buffer<unsigned int>(imax_buffer), imax_offset,
                    Buffer<T>(x_buffer), x_offset, x_inc);
@@ -526,7 +526,7 @@ StatusCode Amin(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xamin<T>(queue_cpp, event);
+    auto routine = Xamin<T>(queue_cpp, nullptr);
     routine.DoAmin(n,
                    Buffer<unsigned int>(imin_buffer), imin_offset,
                    Buffer<T>(x_buffer), x_offset, x_inc);
@@ -564,7 +564,7 @@ StatusCode Max(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xmax<T>(queue_cpp, event);
+    auto routine = Xmax<T>(queue_cpp, nullptr);
     routine.DoMax(n,
                   Buffer<unsigned int>(imax_buffer), imax_offset,
                   Buffer<T>(x_buffer), x_offset, x_inc);
@@ -602,7 +602,7 @@ StatusCode Min(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xmin<T>(queue_cpp, event);
+    auto routine = Xmin<T>(queue_cpp, nullptr);
     routine.DoMin(n,
                   Buffer<unsigned int>(imin_buffer), imin_offset,
                   Buffer<T>(x_buffer), x_offset, x_inc);
@@ -648,7 +648,7 @@ StatusCode Gemv(const Layout layout, const Transpose a_transpose,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xgemv<T>(queue_cpp, event);
+    auto routine = Xgemv<T>(queue_cpp, nullptr);
     routine.DoGemv(layout, a_transpose,
                    m, n,
                    alpha,
@@ -714,7 +714,7 @@ StatusCode Gbmv(const Layout layout, const Transpose a_transpose,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xgbmv<T>(queue_cpp, event);
+    auto routine = Xgbmv<T>(queue_cpp, nullptr);
     routine.DoGbmv(layout, a_transpose,
                    m, n, kl, ku,
                    alpha,
@@ -780,7 +780,7 @@ StatusCode Hemv(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xhemv<T>(queue_cpp, event);
+    auto routine = Xhemv<T>(queue_cpp, nullptr);
     routine.DoHemv(layout, triangle,
                    n,
                    alpha,
@@ -822,7 +822,7 @@ StatusCode Hbmv(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xhbmv<T>(queue_cpp, event);
+    auto routine = Xhbmv<T>(queue_cpp, nullptr);
     routine.DoHbmv(layout, triangle,
                    n, k,
                    alpha,
@@ -864,7 +864,7 @@ StatusCode Hpmv(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xhpmv<T>(queue_cpp, event);
+    auto routine = Xhpmv<T>(queue_cpp, nullptr);
     routine.DoHpmv(layout, triangle,
                    n,
                    alpha,
@@ -906,7 +906,7 @@ StatusCode Symv(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xsymv<T>(queue_cpp, event);
+    auto routine = Xsymv<T>(queue_cpp, nullptr);
     routine.DoSymv(layout, triangle,
                    n,
                    alpha,
@@ -956,7 +956,7 @@ StatusCode Sbmv(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xsbmv<T>(queue_cpp, event);
+    auto routine = Xsbmv<T>(queue_cpp, nullptr);
     routine.DoSbmv(layout, triangle,
                    n, k,
                    alpha,
@@ -1006,7 +1006,7 @@ StatusCode Spmv(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xspmv<T>(queue_cpp, event);
+    auto routine = Xspmv<T>(queue_cpp, nullptr);
     routine.DoSpmv(layout, triangle,
                    n,
                    alpha,
@@ -1053,7 +1053,7 @@ StatusCode Trmv(const Layout layout, const Triangle triangle, const Transpose a_
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xtrmv<T>(queue_cpp, event);
+    auto routine = Xtrmv<T>(queue_cpp, nullptr);
     routine.DoTrmv(layout, triangle, a_transpose, diagonal,
                    n,
                    Buffer<T>(a_buffer), a_offset, a_ld,
@@ -1098,7 +1098,7 @@ StatusCode Tbmv(const Layout layout, const Triangle triangle, const Transpose a_
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xtbmv<T>(queue_cpp, event);
+    auto routine = Xtbmv<T>(queue_cpp, nullptr);
     routine.DoTbmv(layout, triangle, a_transpose, diagonal,
                    n, k,
                    Buffer<T>(a_buffer), a_offset, a_ld,
@@ -1143,7 +1143,7 @@ StatusCode Tpmv(const Layout layout, const Triangle triangle, const Transpose a_
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xtpmv<T>(queue_cpp, event);
+    auto routine = Xtpmv<T>(queue_cpp, nullptr);
     routine.DoTpmv(layout, triangle, a_transpose, diagonal,
                    n,
                    Buffer<T>(ap_buffer), ap_offset,
@@ -1188,7 +1188,7 @@ StatusCode Trsv(const Layout layout, const Triangle triangle, const Transpose a_
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xtrsv<T>(queue_cpp, event);
+    auto routine = Xtrsv<T>(queue_cpp, nullptr);
     routine.DoTrsv(layout, triangle, a_transpose, diagonal,
                    n,
                    Buffer<T>(a_buffer), a_offset, a_ld,
@@ -1290,7 +1290,7 @@ StatusCode Ger(const Layout layout,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xger<T>(queue_cpp, event);
+    auto routine = Xger<T>(queue_cpp, nullptr);
     routine.DoGer(layout,
                   m, n,
                   alpha,
@@ -1335,7 +1335,7 @@ StatusCode Geru(const Layout layout,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xgeru<T>(queue_cpp, event);
+    auto routine = Xgeru<T>(queue_cpp, nullptr);
     routine.DoGeru(layout,
                    m, n,
                    alpha,
@@ -1373,7 +1373,7 @@ StatusCode Gerc(const Layout layout,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xgerc<T>(queue_cpp, event);
+    auto routine = Xgerc<T>(queue_cpp, nullptr);
     routine.DoGerc(layout,
                    m, n,
                    alpha,
@@ -1410,7 +1410,7 @@ StatusCode Her(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xher<std::complex<T>,T>(queue_cpp, event);
+    auto routine = Xher<std::complex<T>,T>(queue_cpp, nullptr);
     routine.DoHer(layout, triangle,
                   n,
                   alpha,
@@ -1444,7 +1444,7 @@ StatusCode Hpr(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xhpr<std::complex<T>,T>(queue_cpp, event);
+    auto routine = Xhpr<std::complex<T>,T>(queue_cpp, nullptr);
     routine.DoHpr(layout, triangle,
                   n,
                   alpha,
@@ -1479,7 +1479,7 @@ StatusCode Her2(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xher2<T>(queue_cpp, event);
+    auto routine = Xher2<T>(queue_cpp, nullptr);
     routine.DoHer2(layout, triangle,
                    n,
                    alpha,
@@ -1517,7 +1517,7 @@ StatusCode Hpr2(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xhpr2<T>(queue_cpp, event);
+    auto routine = Xhpr2<T>(queue_cpp, nullptr);
     routine.DoHpr2(layout, triangle,
                    n,
                    alpha,
@@ -1554,7 +1554,7 @@ StatusCode Syr(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xsyr<T>(queue_cpp, event);
+    auto routine = Xsyr<T>(queue_cpp, nullptr);
     routine.DoSyr(layout, triangle,
                   n,
                   alpha,
@@ -1594,7 +1594,7 @@ StatusCode Spr(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xspr<T>(queue_cpp, event);
+    auto routine = Xspr<T>(queue_cpp, nullptr);
     routine.DoSpr(layout, triangle,
                   n,
                   alpha,
@@ -1635,7 +1635,7 @@ StatusCode Syr2(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xsyr2<T>(queue_cpp, event);
+    auto routine = Xsyr2<T>(queue_cpp, nullptr);
     routine.DoSyr2(layout, triangle,
                    n,
                    alpha,
@@ -1680,7 +1680,7 @@ StatusCode Spr2(const Layout layout, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xspr2<T>(queue_cpp, event);
+    auto routine = Xspr2<T>(queue_cpp, nullptr);
     routine.DoSpr2(layout, triangle,
                    n,
                    alpha,
@@ -1730,7 +1730,7 @@ StatusCode Gemm(const Layout layout, const Transpose a_transpose, const Transpos
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xgemm<T>(queue_cpp, event);
+    auto routine = Xgemm<T>(queue_cpp, nullptr);
     routine.DoGemm(layout, a_transpose, b_transpose,
                    m, n, k,
                    alpha,
@@ -1796,7 +1796,7 @@ StatusCode Symm(const Layout layout, const Side side, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xsymm<T>(queue_cpp, event);
+    auto routine = Xsymm<T>(queue_cpp, nullptr);
     routine.DoSymm(layout, side, triangle,
                    m, n,
                    alpha,
@@ -1862,7 +1862,7 @@ StatusCode Hemm(const Layout layout, const Side side, const Triangle triangle,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xhemm<T>(queue_cpp, event);
+    auto routine = Xhemm<T>(queue_cpp, nullptr);
     routine.DoHemm(layout, side, triangle,
                    m, n,
                    alpha,
@@ -1903,7 +1903,7 @@ StatusCode Syrk(const Layout layout, const Triangle triangle, const Transpose a_
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xsyrk<T>(queue_cpp, event);
+    auto routine = Xsyrk<T>(queue_cpp, nullptr);
     routine.DoSyrk(layout, triangle, a_transpose,
                    n, k,
                    alpha,
@@ -1962,7 +1962,7 @@ StatusCode Herk(const Layout layout, const Triangle triangle, const Transpose a_
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xherk<std::complex<T>,T>(queue_cpp, event);
+    auto routine = Xherk<std::complex<T>,T>(queue_cpp, nullptr);
     routine.DoHerk(layout, triangle, a_transpose,
                    n, k,
                    alpha,
@@ -2001,7 +2001,7 @@ StatusCode Syr2k(const Layout layout, const Triangle triangle, const Transpose a
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xsyr2k<T>(queue_cpp, event);
+    auto routine = Xsyr2k<T>(queue_cpp, nullptr);
     routine.DoSyr2k(layout, triangle, ab_transpose,
                     n, k,
                     alpha,
@@ -2067,7 +2067,7 @@ StatusCode Her2k(const Layout layout, const Triangle triangle, const Transpose a
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xher2k<T,U>(queue_cpp, event);
+    auto routine = Xher2k<T,U>(queue_cpp, nullptr);
     routine.DoHer2k(layout, triangle, ab_transpose,
                     n, k,
                     alpha,
@@ -2107,7 +2107,7 @@ StatusCode Trmm(const Layout layout, const Side side, const Triangle triangle, c
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xtrmm<T>(queue_cpp, event);
+    auto routine = Xtrmm<T>(queue_cpp, nullptr);
     routine.DoTrmm(layout, side, triangle, a_transpose, diagonal,
                    m, n,
                    alpha,
@@ -2159,7 +2159,7 @@ StatusCode Trsm(const Layout layout, const Side side, const Triangle triangle, c
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xtrsm<T>(queue_cpp, event);
+    auto routine = Xtrsm<T>(queue_cpp, nullptr);
     routine.DoTrsm(layout, side, triangle, a_transpose, diagonal,
                    m, n,
                    alpha,
@@ -2209,7 +2209,7 @@ StatusCode Omatcopy(const Layout layout, const Transpose a_transpose,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xomatcopy<T>(queue_cpp, event);
+    auto routine = Xomatcopy<T>(queue_cpp, nullptr);
     routine.DoOmatcopy(layout, a_transpose,
                        m, n,
                        alpha,
@@ -2259,7 +2259,7 @@ StatusCode Im2col(const size_t channels, const size_t height, const size_t width
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = Xim2col<T>(queue_cpp, event);
+    auto routine = Xim2col<T>(queue_cpp, nullptr);
     routine.DoIm2col(channels, height, width, kernel_h, kernel_w, pad_h, pad_w, stride_h, stride_w, dilation_h, dilation_w,
                      Buffer<T>(im_buffer), im_offset,
                      Buffer<T>(col_buffer), col_offset);
@@ -2299,7 +2299,7 @@ StatusCode AxpyBatched(const size_t n,
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = XaxpyBatched<T>(queue_cpp, event);
+    auto routine = XaxpyBatched<T>(queue_cpp, nullptr);
     auto alphas_cpp = std::vector<T>();
     auto x_offsets_cpp = std::vector<size_t>();
     auto y_offsets_cpp = std::vector<size_t>();
@@ -2362,7 +2362,7 @@ StatusCode GemmBatched(const Layout layout, const Transpose a_transpose, const T
     const auto context_cpp = Context(context);
     const auto device_cpp = Device(device);
     auto queue_cpp = Queue(context_cpp, device_cpp);
-    auto routine = XgemmBatched<T>(queue_cpp, event);
+    auto routine = XgemmBatched<T>(queue_cpp, nullptr);
     auto alphas_cpp = std::vector<T>();
     auto betas_cpp = std::vector<T>();
     auto a_offsets_cpp = std::vector<size_t>();
diff --git a/src/cupp11.hpp b/src/cupp11.hpp
index 988366ea..854c0be9 100644
--- a/src/cupp11.hpp
+++ b/src/cupp11.hpp
@@ -370,6 +370,8 @@ using ContextPointer = CUcontext*;
 // C++11 version of 'nvrtcProgram'. Additionally holds the program's source code.
 class Program {
 public:
+  Program() = default;
+
   // Note that there is no constructor based on the regular CUDA data-type because of extra state
 
   // Source-based constructor with memory management
@@ -404,7 +406,7 @@ public:
 
   // Confirms whether a certain status code is an actual compilation error or warning
   bool StatusIsCompilationWarningOrError(const nvrtcResult status) const {
-    return (status == NVRTC_ERROR_INVALID_INPUT);
+    return (status == NVRTC_ERROR_COMPILATION);
   }
 
   // Retrieves the warning/error message from the compiler (if any)
@@ -433,8 +435,8 @@ public:
   const nvrtcProgram& operator()() const { return *program_; }
 private:
   std::shared_ptr<nvrtcProgram> program_;
-  const std::string source_;
-  const bool from_binary_;
+  std::string source_;
+  bool from_binary_;
 };
 
 // =================================================================================================
@@ -730,7 +732,7 @@ public:
   // TODO: Implement this function
   void Launch(const Queue &queue, const std::vector<size_t> &global,
               const std::vector<size_t> &local, EventPointer event,
-              std::vector<Event>& waitForEvents) {
+              const std::vector<Event>& waitForEvents) {
     if (local.size() == 0) {
       throw LogicError("Kernel: launching with a default workgroup size is not implemented for the CUDA back-end");
     }
diff --git a/src/cxpp11_common.hpp b/src/cxpp11_common.hpp
index 6ac008be..5097eac4 100644
--- a/src/cxpp11_common.hpp
+++ b/src/cxpp11_common.hpp
@@ -15,6 +15,7 @@
 #ifndef CLBLAST_CXPP11_COMMON_H_
 #define CLBLAST_CXPP11_COMMON_H_
 
+#include <cstring>   // strchr
 #include <string>    // std::string
 #include <stdexcept> // std::runtime_error
 
diff --git a/src/kernels/opencl_to_cuda.h b/src/kernels/opencl_to_cuda.h
index 43a26a2f..2e46bc2b 100644
--- a/src/kernels/opencl_to_cuda.h
+++ b/src/kernels/opencl_to_cuda.h
@@ -11,6 +11,11 @@
 //
 // =================================================================================================
 
+// Enables loading of this file using the C++ pre-processor's #include (C++11 standard raw string
+// literal). Comment-out this line for syntax-highlighting when developing.
+R"(
+// =================================================================================================
+
 // Replaces the OpenCL keywords with CUDA equivalent
 #define __kernel __placeholder__
 #define __global
@@ -49,3 +54,9 @@ typedef struct { float s0; float s1; float s2; float s3;
                  float s12; float s13; float s14; float s15; } float16;
 
 // =================================================================================================
+
+// End of the C++11 raw string literal
+)"
+
+// =================================================================================================
+