README updated for BFS, HSTI, HSTO, RSCD, RSCT, and SSSP

CUDA-D-Sim/BFS/README

@@ -14,6 +14,28 @@ For more options:
     ./bfs -h
 
 
+Note:
+The input folder contains two graphs from the 9th DIMACS Implementation Challenge 
+(http://www.dis.uniroma1.it/challenge9/download.shtml). This benchmark can use 
+any input graph, provided that the format is as follows:
+(Beginning of the file)
+#Nodes #Edges Source_node
+
+A0 B0
+A1 B1
+...
+
+C0 D0
+C1 D1
+...
+
+Each tuple (Ai, Bi) represents one node. Each tuple (Cj, Dj) represents one edge.
+Thus, the file contains the list of nodes, followed by the list of edges.
+Ai indicates the position where the edges of node i start in the list of edges.
+Bi means the number of edges of node i.
+Cj is the node where edge j terminates (i.e., the head of the edge). Dj is the
+cost of edge j.
+
 Read more:
 L. Luo, M. Wong, and W.-m. Hwu, “An effective GPU implementation of 
 breadth-first search,” in Proceedings of the 47th Design Automation Conference, 

CUDA-D-Sim/HSTI/README

@@ -12,3 +12,14 @@ Execution instructions
 For more options:
 
     ./hsti -h
+
+
+Note:
+The input folder contains one image from Van Hateren's natural image database 
+(http://www.kyb.tuebingen.mpg.de/?id=227). Image pixels are 12-bit depth. Thus, 
+for calculation of the B-bin histogram of an image, the corresponding histogram 
+bin is computed as ((pixel * B) >> 12).
+Monochrome images from other databases or synthetic images can also be used. The 
+read input function (in main.cpp) might need to be changed accordingly. If image 
+pixels are b-bit depth and the histogram contains B bins, the histogram bin will 
+be computed as ((pixel * B) >> b).

CUDA-D-Sim/HSTO/README

@@ -12,3 +12,14 @@ Execution instructions
 For more options:
 
     ./hsto -h
+
+
+Note:
+The input folder contains one image from Van Hateren's natural image database 
+(http://www.kyb.tuebingen.mpg.de/?id=227). Image pixels are 12-bit depth. Thus, 
+for calculation of the B-bin histogram of an image, the corresponding histogram 
+bin is computed as ((pixel * B) >> 12).
+Monochrome images from other databases or synthetic images can also be used. The 
+read input function (in main.cpp) might need to be changed accordingly. If image 
+pixels are b-bit depth and the histogram contains B bins, the histogram bin will 
+be computed as ((pixel * B) >> b).

CUDA-D-Sim/RSCD/README

@@ -14,6 +14,13 @@ For more options:
     ./rscd -h
 
 
+Note:
+The input folder contains a file with a list of flow vectors from a video frame 
+with the format (x, y, vx, vy). (x, y) is the tail of the vector, and (vx, vy) is 
+the head.
+The input file could be synthetically generated with every x, y, vx, and vy within 
+frame dimensions.
+
 Read more:
 M. A. Fischler and R. C. Bolles, “Random sample consensus: a paradigm for model 
 fitting with applications to image analysis and automated cartography,” 

CUDA-D-Sim/RSCT/README

@@ -14,6 +14,13 @@ For more options:
     ./rsct -h
 
 
+Note:
+The input folder contains a file with a list of flow vectors from a video frame 
+with the format (x, y, vx, vy). (x, y) is the tail of the vector, and (vx, vy) is 
+the head.
+The input file could be synthetically generated with every x, y, vx, and vy within 
+frame dimensions.
+
 Read more:
 M. A. Fischler and R. C. Bolles, “Random sample consensus: a paradigm for model 
 fitting with applications to image analysis and automated cartography,” 

CUDA-D-Sim/SSSP/README

@@ -14,6 +14,28 @@ For more options:
     ./sssp -h
 
 
+Note:
+The input folder contains two graphs from the 9th DIMACS Implementation Challenge 
+(http://www.dis.uniroma1.it/challenge9/download.shtml). This benchmark can use 
+any input graph, provided that the format is as follows:
+(Beginning of the file)
+#Nodes #Edges Source_node
+
+A0 B0
+A1 B1
+...
+
+C0 D0
+C1 D1
+...
+
+Each tuple (Ai, Bi) represents one node. Each tuple (Cj, Dj) represents one edge.
+Thus, the file contains the list of nodes, followed by the list of edges.
+Ai indicates the position where the edges of node i start in the list of edges.
+Bi means the number of edges of node i.
+Cj is the node where edge j terminates (i.e., the head of the edge). Dj is the
+cost of edge j.
+
 Read more:
 L. Luo, M. Wong, and W.-m. Hwu, “An effective GPU implementation of 
 breadth-first search,” in Proceedings of the 47th Design Automation Conference, 

CUDA-D/BFS/README

@@ -14,6 +14,28 @@ For more options:
     ./bfs -h
 
 
+Note:
+The input folder contains two graphs from the 9th DIMACS Implementation Challenge 
+(http://www.dis.uniroma1.it/challenge9/download.shtml). This benchmark can use 
+any input graph, provided that the format is as follows:
+(Beginning of the file)
+#Nodes #Edges Source_node
+
+A0 B0
+A1 B1
+...
+
+C0 D0
+C1 D1
+...
+
+Each tuple (Ai, Bi) represents one node. Each tuple (Cj, Dj) represents one edge.
+Thus, the file contains the list of nodes, followed by the list of edges.
+Ai indicates the position where the edges of node i start in the list of edges.
+Bi means the number of edges of node i.
+Cj is the node where edge j terminates (i.e., the head of the edge). Dj is the
+cost of edge j.
+
 Read more:
 L. Luo, M. Wong, and W.-m. Hwu, “An effective GPU implementation of 
 breadth-first search,” in Proceedings of the 47th Design Automation Conference, 

CUDA-D/HSTI/README

@@ -12,3 +12,14 @@ Execution instructions
 For more options:
 
     ./hsti -h
+
+
+Note:
+The input folder contains one image from Van Hateren's natural image database 
+(http://www.kyb.tuebingen.mpg.de/?id=227). Image pixels are 12-bit depth. Thus, 
+for calculation of the B-bin histogram of an image, the corresponding histogram 
+bin is computed as ((pixel * B) >> 12).
+Monochrome images from other databases or synthetic images can also be used. The 
+read input function (in main.cpp) might need to be changed accordingly. If image 
+pixels are b-bit depth and the histogram contains B bins, the histogram bin will 
+be computed as ((pixel * B) >> b).

CUDA-D/HSTO/README

@@ -12,3 +12,14 @@ Execution instructions
 For more options:
 
     ./hsto -h
+
+
+Note:
+The input folder contains one image from Van Hateren's natural image database 
+(http://www.kyb.tuebingen.mpg.de/?id=227). Image pixels are 12-bit depth. Thus, 
+for calculation of the B-bin histogram of an image, the corresponding histogram 
+bin is computed as ((pixel * B) >> 12).
+Monochrome images from other databases or synthetic images can also be used. The 
+read input function (in main.cpp) might need to be changed accordingly. If image 
+pixels are b-bit depth and the histogram contains B bins, the histogram bin will 
+be computed as ((pixel * B) >> b).

CUDA-D/RSCD/README

@@ -14,6 +14,13 @@ For more options:
     ./rscd -h
 
 
+Note:
+The input folder contains a file with a list of flow vectors from a video frame 
+with the format (x, y, vx, vy). (x, y) is the tail of the vector, and (vx, vy) is 
+the head.
+The input file could be synthetically generated with every x, y, vx, and vy within 
+frame dimensions.
+
 Read more:
 M. A. Fischler and R. C. Bolles, “Random sample consensus: a paradigm for model 
 fitting with applications to image analysis and automated cartography,” 

CUDA-D/RSCT/README

@@ -14,6 +14,13 @@ For more options:
     ./rsct -h
 
 
+Note:
+The input folder contains a file with a list of flow vectors from a video frame 
+with the format (x, y, vx, vy). (x, y) is the tail of the vector, and (vx, vy) is 
+the head.
+The input file could be synthetically generated with every x, y, vx, and vy within 
+frame dimensions.
+
 Read more:
 M. A. Fischler and R. C. Bolles, “Random sample consensus: a paradigm for model 
 fitting with applications to image analysis and automated cartography,” 

CUDA-D/SSSP/README

@@ -14,6 +14,28 @@ For more options:
     ./sssp -h
 
 
+Note:
+The input folder contains two graphs from the 9th DIMACS Implementation Challenge 
+(http://www.dis.uniroma1.it/challenge9/download.shtml). This benchmark can use 
+any input graph, provided that the format is as follows:
+(Beginning of the file)
+#Nodes #Edges Source_node
+
+A0 B0
+A1 B1
+...
+
+C0 D0
+C1 D1
+...
+
+Each tuple (Ai, Bi) represents one node. Each tuple (Cj, Dj) represents one edge.
+Thus, the file contains the list of nodes, followed by the list of edges.
+Ai indicates the position where the edges of node i start in the list of edges.
+Bi means the number of edges of node i.
+Cj is the node where edge j terminates (i.e., the head of the edge). Dj is the
+cost of edge j.
+
 Read more:
 L. Luo, M. Wong, and W.-m. Hwu, “An effective GPU implementation of 
 breadth-first search,” in Proceedings of the 47th Design Automation Conference, 

CUDA-U-Sim/BFS/README

@@ -14,6 +14,28 @@ For more options:
     ./bfs -h
 
 
+Note:
+The input folder contains two graphs from the 9th DIMACS Implementation Challenge 
+(http://www.dis.uniroma1.it/challenge9/download.shtml). This benchmark can use 
+any input graph, provided that the format is as follows:
+(Beginning of the file)
+#Nodes #Edges Source_node
+
+A0 B0
+A1 B1
+...
+
+C0 D0
+C1 D1
+...
+
+Each tuple (Ai, Bi) represents one node. Each tuple (Cj, Dj) represents one edge.
+Thus, the file contains the list of nodes, followed by the list of edges.
+Ai indicates the position where the edges of node i start in the list of edges.
+Bi means the number of edges of node i.
+Cj is the node where edge j terminates (i.e., the head of the edge). Dj is the
+cost of edge j.
+
 Read more:
 L. Luo, M. Wong, and W.-m. Hwu, “An effective GPU implementation of 
 breadth-first search,” in Proceedings of the 47th Design Automation Conference, 

CUDA-U-Sim/HSTI/README

@@ -12,3 +12,14 @@ Execution instructions
 For more options:
 
     ./hsti -h
+
+
+Note:
+The input folder contains one image from Van Hateren's natural image database 
+(http://www.kyb.tuebingen.mpg.de/?id=227). Image pixels are 12-bit depth. Thus, 
+for calculation of the B-bin histogram of an image, the corresponding histogram 
+bin is computed as ((pixel * B) >> 12).
+Monochrome images from other databases or synthetic images can also be used. The 
+read input function (in main.cpp) might need to be changed accordingly. If image 
+pixels are b-bit depth and the histogram contains B bins, the histogram bin will 
+be computed as ((pixel * B) >> b).

CUDA-U-Sim/HSTO/README

@@ -12,3 +12,14 @@ Execution instructions
 For more options:
 
     ./hsto -h
+
+
+Note:
+The input folder contains one image from Van Hateren's natural image database 
+(http://www.kyb.tuebingen.mpg.de/?id=227). Image pixels are 12-bit depth. Thus, 
+for calculation of the B-bin histogram of an image, the corresponding histogram 
+bin is computed as ((pixel * B) >> 12).
+Monochrome images from other databases or synthetic images can also be used. The 
+read input function (in main.cpp) might need to be changed accordingly. If image 
+pixels are b-bit depth and the histogram contains B bins, the histogram bin will 
+be computed as ((pixel * B) >> b).

CUDA-U-Sim/RSCD/README

@@ -14,6 +14,13 @@ For more options:
     ./rscd -h
 
 
+Note:
+The input folder contains a file with a list of flow vectors from a video frame 
+with the format (x, y, vx, vy). (x, y) is the tail of the vector, and (vx, vy) is 
+the head.
+The input file could be synthetically generated with every x, y, vx, and vy within 
+frame dimensions.
+
 Read more:
 M. A. Fischler and R. C. Bolles, “Random sample consensus: a paradigm for model 
 fitting with applications to image analysis and automated cartography,” 

CUDA-U-Sim/RSCT/README

@@ -14,6 +14,13 @@ For more options:
     ./rsct -h
 
 
+Note:
+The input folder contains a file with a list of flow vectors from a video frame 
+with the format (x, y, vx, vy). (x, y) is the tail of the vector, and (vx, vy) is 
+the head.
+The input file could be synthetically generated with every x, y, vx, and vy within 
+frame dimensions.
+
 Read more:
 M. A. Fischler and R. C. Bolles, “Random sample consensus: a paradigm for model 
 fitting with applications to image analysis and automated cartography,” 

CUDA-U-Sim/SSSP/README

@@ -14,6 +14,28 @@ For more options:
     ./sssp -h
 
 
+Note:
+The input folder contains two graphs from the 9th DIMACS Implementation Challenge 
+(http://www.dis.uniroma1.it/challenge9/download.shtml). This benchmark can use 
+any input graph, provided that the format is as follows:
+(Beginning of the file)
+#Nodes #Edges Source_node
+
+A0 B0
+A1 B1
+...
+
+C0 D0
+C1 D1
+...
+
+Each tuple (Ai, Bi) represents one node. Each tuple (Cj, Dj) represents one edge.
+Thus, the file contains the list of nodes, followed by the list of edges.
+Ai indicates the position where the edges of node i start in the list of edges.
+Bi means the number of edges of node i.
+Cj is the node where edge j terminates (i.e., the head of the edge). Dj is the
+cost of edge j.
+
 Read more:
 L. Luo, M. Wong, and W.-m. Hwu, “An effective GPU implementation of 
 breadth-first search,” in Proceedings of the 47th Design Automation Conference, 

CUDA-U/BFS/README

@@ -14,6 +14,28 @@ For more options:
     ./bfs -h
 
 
+Note:
+The input folder contains two graphs from the 9th DIMACS Implementation Challenge 
+(http://www.dis.uniroma1.it/challenge9/download.shtml). This benchmark can use 
+any input graph, provided that the format is as follows:
+(Beginning of the file)
+#Nodes #Edges Source_node
+
+A0 B0
+A1 B1
+...
+
+C0 D0
+C1 D1
+...
+
+Each tuple (Ai, Bi) represents one node. Each tuple (Cj, Dj) represents one edge.
+Thus, the file contains the list of nodes, followed by the list of edges.
+Ai indicates the position where the edges of node i start in the list of edges.
+Bi means the number of edges of node i.
+Cj is the node where edge j terminates (i.e., the head of the edge). Dj is the
+cost of edge j.
+
 Read more:
 L. Luo, M. Wong, and W.-m. Hwu, “An effective GPU implementation of 
 breadth-first search,” in Proceedings of the 47th Design Automation Conference, 

CUDA-U/HSTI/README

@@ -12,3 +12,14 @@ Execution instructions
 For more options:
 
     ./hsti -h
+
+
+Note:
+The input folder contains one image from Van Hateren's natural image database 
+(http://www.kyb.tuebingen.mpg.de/?id=227). Image pixels are 12-bit depth. Thus, 
+for calculation of the B-bin histogram of an image, the corresponding histogram 
+bin is computed as ((pixel * B) >> 12).
+Monochrome images from other databases or synthetic images can also be used. The 
+read input function (in main.cpp) might need to be changed accordingly. If image 
+pixels are b-bit depth and the histogram contains B bins, the histogram bin will 
+be computed as ((pixel * B) >> b).