v4.2 release. (#2587)
* Fix default cluster callback values to 1 to avoid profiler failure when these values are not set in command line. * v4.2 release.
This commit is contained in:
@@ -40,14 +40,12 @@
|
||||
Note that in general the fragment passed to the OutputOp could
|
||||
span multiple rows but it does not happen with the configurations we have
|
||||
*/
|
||||
|
||||
#pragma once
|
||||
|
||||
#include <cuda/std/cassert>
|
||||
|
||||
#include "cutlass/aligned_buffer.h"
|
||||
#include "cutlass/array.h"
|
||||
#include "cutlass/cutlass.h"
|
||||
#include CUDA_STD_HEADER(cassert)
|
||||
#include "cutlass/functional.h"
|
||||
#include "cutlass/layout/tensor.h"
|
||||
#include "cutlass/layout/vector.h"
|
||||
|
||||
@@ -42,12 +42,10 @@
|
||||
*/
|
||||
|
||||
#pragma once
|
||||
|
||||
#include <cuda/std/cassert>
|
||||
|
||||
#include "cutlass/cutlass.h"
|
||||
#include CUDA_STD_HEADER(cassert)
|
||||
#include "cutlass/aligned_buffer.h"
|
||||
#include "cutlass/array.h"
|
||||
#include "cutlass/cutlass.h"
|
||||
#include "cutlass/functional.h"
|
||||
#include "cutlass/layout/tensor.h"
|
||||
#include "cutlass/layout/vector.h"
|
||||
|
||||
Reference in New Issue
Block a user