summaryrefslogtreecommitdiff
path: root/src/routines/level2/xher2.cpp
diff options
context:
space:
mode:
authorCedric Nugteren <web@cedricnugteren.nl>2016-07-10 20:32:01 +0200
committerCedric Nugteren <web@cedricnugteren.nl>2016-07-10 20:32:01 +0200
commitc87e877bf23d2fe38a7da2898e1734a3cdeaf48c (patch)
treed091dfdd826dd11e5c9e533eb46b22aeb7f6f823 /src/routines/level2/xher2.cpp
parent57f09178d89a1cf4f38a0bb338c864ed850d5470 (diff)
Now passing alpha/beta to the kernel as arguments as before fp16 support; in case of fp16 arguments are cast on host and in kernel
Diffstat (limited to 'src/routines/level2/xher2.cpp')
-rw-r--r--src/routines/level2/xher2.cpp6
1 files changed, 1 insertions, 5 deletions
diff --git a/src/routines/level2/xher2.cpp b/src/routines/level2/xher2.cpp
index 50572cea..3d57a9b9 100644
--- a/src/routines/level2/xher2.cpp
+++ b/src/routines/level2/xher2.cpp
@@ -58,10 +58,6 @@ StatusCode Xher2<T>::DoHer2(const Layout layout, const Triangle triangle,
status = TestVectorY(n, y_buffer, y_offset, y_inc);
if (ErrorIn(status)) { return status; }
- // Upload the scalar argument as a constant buffer to the device (needed for half-precision)
- auto alpha_buffer = Buffer<T>(context_, 1);
- alpha_buffer.Write(queue_, 1, &alpha);
-
// Retrieves the kernel from the compiled binary
try {
const auto program = GetProgramFromCache(context_, PrecisionValue<T>(), routine_name_);
@@ -69,7 +65,7 @@ StatusCode Xher2<T>::DoHer2(const Layout layout, const Triangle triangle,
// Sets the kernel arguments
kernel.SetArgument(0, static_cast<int>(n));
- kernel.SetArgument(1, alpha_buffer());
+ kernel.SetArgument(1, GetRealArg(alpha));
kernel.SetArgument(2, x_buffer());
kernel.SetArgument(3, static_cast<int>(x_offset));
kernel.SetArgument(4, static_cast<int>(x_inc));