CUTLASS 2.4 (Implicit GEMM convolution) (#147)

CUTLASS 2.4 (Implicit GEMM Convolution)

Co-authored-by: Manish Gupta <manigupta@nvidia.com>, Haicheng Wu <haichengw@nvidia.com>, Dustyn Blasig <dblasig@nvidia.com>, Andrew Kerr <akerr@nvidia.com>

examples/03_visualize_layout/CMakeLists.txt

@@ -20,9 +20,15 @@
 # STRICT LIABILITY, OR TOR (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
+set(TEST_COMMAND_00 RowMajor --extent=16,16)
+set(TEST_COMMAND_01 "ColumnMajorInterleaved<4>" --extent=32,8 --output-shape=16 --vectorize=4)
+
 cutlass_example_add_executable(
   03_visualize_layout
   visualize_layout.cpp
   register_layout.cu
+  TEST_COMMAND_OPTIONS
+  TEST_COMMAND_00
+  TEST_COMMAND_01
   )
 

examples/03_visualize_layout/visualize_layout.cpp

@@ -32,6 +32,8 @@
 #include <iomanip>
 #include <memory>
 
+#include <cutlass/cutlass.h>
+
 #include "options.h"
 #include "register_layout.h"
 
@@ -133,6 +135,8 @@ int main(int argc, char const *arg[]) {
 
   layout_it->second->print_csv(std::cout);
 
+  cudaFree(0); // Ensure CUDA is available.
+
   return 0;
 }