diff options
Diffstat (limited to 'src/routines/level3/xherk.cpp')
-rw-r--r-- | src/routines/level3/xherk.cpp | 10 |
1 files changed, 3 insertions, 7 deletions
diff --git a/src/routines/level3/xherk.cpp b/src/routines/level3/xherk.cpp index 6ef7f21f..cc87e3e9 100644 --- a/src/routines/level3/xherk.cpp +++ b/src/routines/level3/xherk.cpp @@ -98,13 +98,9 @@ StatusCode Xherk<T,U>::DoHerk(const Layout layout, const Triangle triangle, cons auto b_temp = (b_no_temp) ? a_buffer : Buffer<T>(context_, k_ceiled*n_ceiled); auto c_temp = Buffer<T>(context_, n_ceiled*n_ceiled); - // Upload the scalar arguments as constant buffers to the device (needed for half-precision) + // Convert the arguments to complex versions auto complex_alpha = T{alpha, static_cast<U>(0.0)}; auto complex_beta = T{beta, static_cast<U>(0.0)}; - auto alpha_buffer = Buffer<T>(context_, 1); - auto beta_buffer = Buffer<T>(context_, 1); - alpha_buffer.Write(queue_, 1, &complex_alpha); - beta_buffer.Write(queue_, 1, &complex_beta); // Events of all kernels (including pre/post processing kernels) auto eventWaitList = std::vector<Event>(); @@ -152,8 +148,8 @@ StatusCode Xherk<T,U>::DoHerk(const Layout layout, const Triangle triangle, cons // Sets the kernel arguments kernel.SetArgument(0, static_cast<int>(n_ceiled)); kernel.SetArgument(1, static_cast<int>(k_ceiled)); - kernel.SetArgument(2, alpha_buffer()); - kernel.SetArgument(3, beta_buffer()); + kernel.SetArgument(2, GetRealArg(complex_alpha)); + kernel.SetArgument(3, GetRealArg(complex_beta)); kernel.SetArgument(4, a_temp()); kernel.SetArgument(5, b_temp()); kernel.SetArgument(6, c_temp()); |