summaryrefslogtreecommitdiff
path: root/src/routines/level3/xgemm.cc
diff options
context:
space:
mode:
Diffstat (limited to 'src/routines/level3/xgemm.cc')
-rw-r--r--src/routines/level3/xgemm.cc5
1 files changed, 3 insertions, 2 deletions
diff --git a/src/routines/level3/xgemm.cc b/src/routines/level3/xgemm.cc
index aa081e81..3699b548 100644
--- a/src/routines/level3/xgemm.cc
+++ b/src/routines/level3/xgemm.cc
@@ -184,12 +184,13 @@ StatusCode Xgemm<T>::DoGemm(const Layout layout,
// Launches the kernel
auto eventKernel = Event();
- status = RunKernel(kernel, global, local, eventKernel.pointer(), eventWaitList);
+ auto eventPointer = (!c_no_temp) ? eventKernel.pointer() : event_;
+ status = RunKernel(kernel, global, local, eventPointer, eventWaitList);
if (ErrorIn(status)) { return status; }
- eventWaitList.push_back(eventKernel);
// Runs the post-processing kernel if needed
if (!c_no_temp) {
+ eventWaitList.push_back(eventKernel);
status = PadCopyTransposeMatrix(event_, eventWaitList,
m_ceiled, n_ceiled, m_ceiled, 0, c_temp,
c_one, c_two, c_ld, c_offset, c_buffer,