Remove redundant <fstream> includes (#563)

* Remove redundant <fstream> includes

* Fix fstream in examples/

* Fix <fstream> in test/

* Use consistent order for <fstream> (always after <iostream>)

* Remove an unneeded include in a file where std::ofstream usage is commented out

Co-authored-by: Ivan Komarov <dfyz@yandex-team.ru>

examples/04_tile_iterator/tile_iterator.cu

@@ -50,7 +50,6 @@
 #include <iostream>
 #include <sstream>
 #include <vector>
-#include <fstream>
 
 // CUTLASS includes
 #include "cutlass/transform/threadblock/predicated_tile_iterator.h"

examples/09_turing_tensorop_conv2dfprop/turing_tensorop_conv2dfprop.cu

@@ -124,6 +124,7 @@ compare if the output from CUTLASS kernel is same as the reference implicit GEMM
 */
 
 #include <iostream>
+#include <fstream>
 #include <sstream>
 
 #include "cutlass/cutlass.h"

examples/10_planar_complex/planar_complex.cu

@@ -74,7 +74,6 @@
 */
 
 #include <iostream>
-#include <fstream>
 #include <sstream>
 
 #include "cutlass/cutlass.h"

examples/11_planar_complex_array/planar_complex_array.cu

@@ -72,7 +72,6 @@
 */
 
 #include <iostream>
-#include <fstream>
 #include <sstream>
 
 #include "cutlass/cutlass.h"

examples/16_ampere_tensorop_conv2dfprop/ampere_tensorop_conv2dfprop.cu

@@ -111,6 +111,7 @@ compare if the output from CUTLASS kernel is same as the reference implicit GEMM
 */
 
 #include <iostream>
+#include <fstream>
 #include <sstream>
 
 #include "cutlass/cutlass.h"

examples/22_quaternion_conv/quaternion_conv.cu

@@ -30,6 +30,7 @@
  **************************************************************************************************/
 
 #include <iostream>
+#include <fstream>
 #include <sstream>
 
 #include "cutlass/cutlass.h"

examples/23_ampere_gemm_operand_reduction_fusion/ampere_gemm_operand_reduction_fusion.cu

@@ -41,6 +41,7 @@ epilogue/threadblock/epilogue_gemm_k_reduction.h
 */
 
 #include <iostream>
+#include <fstream>
 #include <sstream>
 
 #include "cutlass/cutlass.h"

examples/25_ampere_fprop_mainloop_fusion/ampere_fprop_mainloop_fusion.cu

@@ -52,6 +52,7 @@ line is the same.
 */
 
 #include <iostream>
+#include <fstream>
 #include <sstream>
 
 #include "cutlass/cutlass.h"

examples/26_ampere_wgrad_mainloop_fusion/ampere_wgrad_mainloop_fusion.cu

@@ -49,6 +49,7 @@ technical details.
 */
 
 #include <iostream>
+#include <fstream>
 #include <sstream>
 
 #include "cutlass/cutlass.h"

examples/28_ampere_3xtf32_fast_accurate_tensorop_fprop/ampere_3xtf32_fast_accurate_tensorop_fprop.cu

@@ -36,6 +36,7 @@ compared with CUDA Cores.  See example 27 for the trick of 3xTF32.
 */
 
 #include <iostream>
+#include <fstream>
 #include <sstream>
 
 #include "cutlass/cutlass.h"

examples/30_wgrad_split_k/30_wgrad_split_k.cu

@@ -40,6 +40,7 @@ to correctly instantiate the GEMM template.
 */
 
 #include <iostream>
+#include <fstream>
 #include <sstream>
 
 #include "cutlass/cutlass.h"

examples/36_gather_scatter_fusion/gather_scatter_fusion.cu

@@ -69,6 +69,7 @@
 
 #include <algorithm>
 #include <iostream>
+#include <fstream>
 #include <random>
 #include <numeric>