CIS565_Project2

README.md

@@ -3,12 +3,90 @@ CUDA Stream Compaction
 
 **University of Pennsylvania, CIS 565: GPU Programming and Architecture, Project 2**
 
-* (TODO) YOUR NAME HERE
-  * (TODO) [LinkedIn](), [personal website](), [twitter](), etc.
-* Tested on: (TODO) Windows 22, i7-2222 @ 2.22GHz 22GB, GTX 222 222MB (Moore 2222 Lab)
+* Xinyu Lin
+[Linkedin](https://www.linkedin.com/in/xinyu-lin-138352125/)
+* Tested on: Windows 10, Intel(R) Core(TM) i7-6700HQ [email protected], 16GB, GTX960M(Private Computer)
 
-### (TODO: Your README)
+### Features
 
-Include analysis, etc. (Remember, this is public, so don't put
-anything here that you don't want to share with the world.)
+* CPU Scan & Stream Compaction
+* Naive GPU Scan Algorithm
+* Work-Efficient GPU Scan
+* Work-Efficient GPU Stream Compaction
+* Thrust Implementation
 
+### Performance Analysis
+* BlockSize over Scan methods(ArraySize: 256)
+![](img/bsscan.png)
+
+* BlockSize over Compact methods(ArraySize: 256)
+![](img/bscompact.png)
+
+* ArraySize over Scan methods(BlockSize: 512)
+![](img/as256scan.png)
+![](img/as512scan'.png)
+![](img/as1024scan.png)
+
+* ArraySize over Compact methods(BlockSize: 512)
+![](img/as256compact.png)
+![](img/as512compact.png)
+![](img/as1024compact.png)
+
+### Result
+```
+****************
+** SCAN TESTS **
+****************
+    [  33  32  23  11   4  10  29  49   9  36  47  43  19 ...  20   0 ]
+==== cpu scan, power-of-two ====
+   elapsed time: 0.001185ms    (std::chrono Measured)
+    [   0  33  65  88  99 103 113 142 191 200 236 283 326 ... 6332 6352 ]
+==== cpu scan, non-power-of-two ====
+   elapsed time: 0.001185ms    (std::chrono Measured)
+    [   0  33  65  88  99 103 113 142 191 200 236 283 326 ... 6208 6251 ]
+    passed
+==== naive scan, power-of-two ====
+   elapsed time: 0.240992ms    (CUDA Measured)
+    passed
+==== naive scan, non-power-of-two ====
+   elapsed time: 0.189888ms    (CUDA Measured)
+    passed
+==== work-efficient scan, power-of-two ====
+   elapsed time: 0.262976ms    (CUDA Measured)
+    passed
+==== work-efficient scan, non-power-of-two ====
+   elapsed time: 0.228768ms    (CUDA Measured)
+    passed
+==== thrust scan, power-of-two ====
+   elapsed time: 0.001088ms    (CUDA Measured)
+    passed
+==== thrust scan, non-power-of-two ====
+   elapsed time: 0.001088ms    (CUDA Measured)
+    passed
+
+*****************************
+** STREAM COMPACTION TESTS **
+*****************************
+    [   1   0   2   2   0   2   3   2   0   0   3   1   0 ...   0   0 ]
+==== cpu compact without scan, power-of-two ====
+   elapsed time: 0.001975ms    (std::chrono Measured)
+    [   1   2   2   2   3   2   3   1   3   2   1   2   1 ...   2   3 ]
+    passed
+==== cpu compact without scan, non-power-of-two ====
+   elapsed time: 0.001976ms    (std::chrono Measured)
+    [   1   2   2   2   3   2   3   1   3   2   1   2   1 ...   2   3 ]
+    passed
+==== cpu compact with scan ====
+   elapsed time: 0.008692ms    (std::chrono Measured)
+    [   1   2   2   2   3   2   3   1   3   2   1   2   1 ...   2   3 ]
+    passed
+==== work-efficient compact, power-of-two ====
+   elapsed time: 0.499232ms    (CUDA Measured)
+    passed
+==== work-efficient compact, non-power-of-two ====
+   elapsed time: 0.43552ms    (CUDA Measured)
+    passed
+```
+
+
+

src/main.cpp

@@ -79,6 +79,17 @@ int main(int argc, char* argv[]) {
     StreamCompaction::Efficient::scan(NPOT, c, a);
     printElapsedTime(StreamCompaction::Efficient::timer().getGpuElapsedTimeForPreviousOperation(), "(CUDA Measured)");
     //printArray(NPOT, c, true);
+	//std::cout << "my result: "<< std::endl;
+	//for (int i = 0; i < SIZE; ++i)
+	//{
+	//	std::cout << c[i] << " ";
+	//}
+	//std::cout << std::endl;
+	//std::cout << "correct result: " << std::endl;
+	//for (int i = 0; i < SIZE; ++i)
+	//{
+	//	std::cout << b[i] << " ";
+	//}
     printCmpResult(NPOT, b, c);
 
     zeroArray(SIZE, c);

stream_compaction/CMakeLists.txt

@@ -13,5 +13,5 @@ set(SOURCE_FILES
 
 cuda_add_library(stream_compaction
     ${SOURCE_FILES}
-    OPTIONS -arch=sm_20
+    OPTIONS -arch=sm_50
     )

stream_compaction/common.h

@@ -12,7 +12,7 @@
 
 #define FILENAME (strrchr(__FILE__, '/') ? strrchr(__FILE__, '/') + 1 : __FILE__)
 #define checkCUDAError(msg) checkCUDAErrorFn(msg, FILENAME, __LINE__)
-
+#define blockSize 1024
 /**
  * Check for CUDA errors; print and exit if there was a problem.
  */

stream_compaction/cpu.cu

@@ -1,15 +1,15 @@
 #include <cstdio>
 #include "cpu.h"
 
-#include "common.h"
+#include "common.h"
 
 namespace StreamCompaction {
     namespace CPU {
-        using StreamCompaction::Common::PerformanceTimer;
-        PerformanceTimer& timer()
-        {
-	        static PerformanceTimer timer;
-	        return timer;
+        using StreamCompaction::Common::PerformanceTimer;
+        PerformanceTimer& timer()
+        {
+	        static PerformanceTimer timer;
+	        return timer;
         }
 
         /**
@@ -20,6 +20,13 @@ namespace StreamCompaction {
         void scan(int n, int *odata, const int *idata) {
 	        timer().startCpuTimer();
             // TODO
+			// Serial Scan: Exclusive
+			int acc = 0;
+			for (int i = 0; i < n; i++)
+			{
+				odata[i] = acc;
+				acc = acc + idata[i];			
+			}
 	        timer().endCpuTimer();
         }
 
@@ -31,8 +38,19 @@ namespace StreamCompaction {
         int compactWithoutScan(int n, int *odata, const int *idata) {
 	        timer().startCpuTimer();
             // TODO
+			int flag = 0;
+			for (int i = 0; i < n; ++i)
+			{
+				if (idata[i] != 0)
+				{
+					odata[flag] = idata[i];
+					flag++;
+				}
+
+			}
+
 	        timer().endCpuTimer();
-            return -1;
+            return flag;
         }
 
         /**
@@ -43,8 +61,46 @@ namespace StreamCompaction {
         int compactWithScan(int n, int *odata, const int *idata) {
 	        timer().startCpuTimer();
 	        // TODO
+			// predicate
+			int acc = 0;
+			int* tempdata = new int[n];
+			for (int i = 0; i < n; ++i)
+			{
+				if (idata[i] == 0)
+				{
+					tempdata[i] = 0;
+
+				}
+				else
+				{
+					tempdata[i] = 1;
+					acc++;
+				}
+			}
+			// ======= tempdata[] = 1110011111001111
+			// scan sum
+			int accSum = 0;
+			int* tempData2 = new int[n];
+			for (int i = 0; i < n; i++)
+			{
+				tempData2[i] = accSum;
+				accSum += tempdata[i];
+
+			}
+
+			// idata[] = 3120043215002222
+			// ======= tempdata[] = 1110011111001111
+			// tempData2[] = 0123334567899910....
+			// scatter
+			for (int i = 0; i < n; ++i)
+			{
+				odata[tempData2[i]] = idata[i];
+			}
+
+			delete[] tempdata;
+			delete[] tempData2;
 	        timer().endCpuTimer();
-            return -1;
+            return acc;
         }
     }
 }