add cpu cache

1-cpu_cache/branch_predict/branch_predict.cpp

+#include "stdio.h"
+#include <stdlib.h>
+#include <time.h>
+#include <algorithm>
+#include <iostream>
+#include <fstream>
+#include <unistd.h>
+
+using namespace std;
+
+#define TESTN 128*1024*1024L
+
+long timediff(clock_t t1, clock_t t2) {
+    long elapsed;
+    elapsed = ((double)t2 - t1) / CLOCKS_PER_SEC * 1000;
+    return elapsed;
+}
+
+int main(int argc, char** argv) {
+	int mode = 1;
+	int ch;
+	while((ch = getopt(argc, argv, "fsg")) != -1) {
+		switch(ch)
+		{
+		   case 's':
+			  mode = 1;
+			  break;
+		   case 'f':
+			  mode = 2;
+			  break;
+		   case 'g':
+			  mode = 3;
+			  break;
+		}
+	}
+
+	unsigned char* arr = new unsigned char[TESTN];
+
+	if (3 == mode) {
+		for (long i = 0; i < TESTN; i++) arr[i] = rand() % 256;
+		ofstream ofs;
+		ofs.open("rand.array", ios::out | ios::binary);
+		ofs.write((const char*)arr, TESTN);
+		ofs.close();
+		sort(arr,arr+TESTN);
+		ofs.open("sort.array", ios::out | ios::binary);
+		ofs.write((const char*)arr, TESTN);
+		ofs.close();
+	} else {
+		const char* fname;
+		if ( 1 == mode) {
+			fname = "rand.array";
+		}else if (2 == mode) {
+			fname = "sort.array";
+		}
+		ifstream ifs;
+		ifs.open(fname);
+		ifs.read((char *)arr, TESTN);
+
+		clock_t start,end;
+		start =clock();
+		for(long i = 0; i < TESTN; i++) {
+			if (arr[i] < 128) arr[i] = 0;
+		}
+		end =clock();
+	    cout<<" "<<timediff(start,end)<<endl;
+	}
+}

1-cpu_cache/cpu_migrate/cpu_migrate.cpp

+#include "stdio.h"
+#include <stdlib.h>
+#include <time.h>
+#include <sys/time.h>
+#include <iostream>
+#include <unistd.h>
+#include <pthread.h>
+#include <sched.h>
+
+
+using namespace std;
+
+#define TESTN 16*1024L
+
+bool setaffinity = false;
+
+void* loopcalc(void* args) {
+	if (setaffinity) {
+		cpu_set_t mask;  //CPU核的集合
+		cpu_set_t get;	 //获取在集合中的CPU
+		int *thread_num = (int *)args; 
+
+		CPU_ZERO(&mask);    //置空
+	    CPU_SET(*thread_num,&mask);   //设置亲和力值
+		if (sched_setaffinity(0, sizeof(mask), &mask) == -1)//设置线程CPU亲和力
+		{
+			cout<<"warning: could not set CPU affinity, continuing...\n";
+		}
+	}
+	timeval tStart,tEnd;
+	gettimeofday(&tStart, 0);
+
+	unsigned char* arr = new unsigned char[TESTN];
+	for (long i = 0; i < TESTN; i++) arr[i] = rand() % 256;
+	for (int j = 1; j < 16*1024; j++) {
+		for (long i = 0; i < TESTN; i++) arr[i] += 1;
+	}
+
+	gettimeofday(&tEnd, 0);
+	
+	*(long*)args = (1000000LL * (tEnd.tv_sec-tStart.tv_sec) + (tEnd.tv_usec-tStart.tv_usec))/1000;
+}
+
+int main(int argc, char** argv) {
+	int threadnum = 2;
+	int ch;
+	while((ch = getopt(argc, argv, "t:fs")) != -1) {
+		switch(ch)
+		{
+		   case 't':
+			  threadnum = atoi(optarg);
+			  break;
+		   case 'f':
+		      setaffinity = true;
+			  break;
+		   case 's':
+		   	  setaffinity = false;
+			  break;
+		}
+	}
+
+	pthread_t* id = new pthread_t[threadnum];
+	long* timecost = new long[threadnum];
+	for(int i = 0; i < threadnum; i++) {
+		timecost[i] = i;
+		int ret=pthread_create(&id[i],NULL,loopcalc,&timecost[i]); 
+		
+		if(ret!=0){ 
+			cout<<"Create pthread error!\n"; 
+			exit (1); 
+		} 
+	}
+
+	long costsum = 0;
+	for(int i = 0; i < threadnum; i++) {
+		pthread_join(id[i],NULL);
+		costsum += timecost[i];
+	}
+	cout<<"costsum: "<<costsum<<", avg: "<<costsum/threadnum<<endl;
+}
+

1-cpu_cache/traverse_1d_array/traverse_1d_array.cpp

+#include <time.h>
+#include <iostream>
+#include <unistd.h>
+#include <stdlib.h>
+
+using namespace std;
+
+long timediff(clock_t t1, clock_t t2) {
+    long elapsed;
+    elapsed = ((double)t2 - t1) / CLOCKS_PER_SEC * 1000;
+    return elapsed;
+}
+
+int main(int argc, char** argv) {
+	int step = 1,ch;
+	bool slowMode = true;
+	long TESTN = 1024*1024*1024*8L;
+	while((ch = getopt(argc, argv, "s:")) != -1) {
+		switch(ch)
+		{
+		   case 's':
+			  step = atoi(optarg);
+			  break;
+		}
+	}
+
+	char* arr = new char[TESTN];
+	clock_t start, end;
+	long total = TESTN/1024,cnt = 0;
+	long i = 0;
+	start =clock();
+	while (++cnt < total) {
+	    arr[i] = 0;
+		i += step;
+	}
+	end =clock();
+    cout<<timediff(start,end)<<",access count:"<<cnt<<endl;
+}
+

1-cpu_cache/traverse_2d_array/README.md

+* C++程序traverse_2d_array.cpp
+** 编译程序
+*** 安装编译依赖的软件
+如Linux中需要安装gcc-c++，CentOS中可用yum install gcc-c++安装，Ubuntu中可用apt-get install gcc-c++
+*** 编译程序
+g++ traverse_2d_array.cpp -o traverse_2d_array
+*** 运行验证
+**** 使用array[i][j]遍历数组
+./traverse_2d_array -f
+**** 使用array[j][i]遍历数组
+./traverse_2d_array -s
+*** 使用perf验证缓存命中率
+**** 使用array[i][j]遍历数组
+perf stat -e cache-references,cache-misses,instructions,cycles,L1-dcache-load-misses,L1-dcache-loads ./traverse_2d_array -f
+**** 使用array[j][i]遍历数组
+perf stat -e cache-references,cache-misses,instructions,cycles,L1-dcache-load-misses,L1-dcache-loads ./traverse_2d_array -s
+* python程序traverse_2d_array.py
\ No newline at end of file

1-cpu_cache/traverse_2d_array/traverse_2d_array.cpp

+#include <time.h>
+#include <iostream>
+#include <unistd.h>
+#include <stdlib.h>
+
+using namespace std;
+
+long timediff(clock_t t1, clock_t t2) {
+    long elapsed;
+    elapsed = ((double)t2 - t1) / CLOCKS_PER_SEC * 1000;
+    return elapsed;
+}
+
+#define TESTN 2048
+
+
+int main(int argc, char** argv) {
+	int ch;
+	bool slowMode = true;
+	while((ch = getopt(argc, argv, "fsn:")) != -1) {
+		switch(ch)
+		{
+		   case 's':
+			  slowMode = true;
+			  break;
+		   case 'f':
+			  slowMode = false;
+			  break;
+		}
+	}
+	cout<<(slowMode?"arr[j][i]":"arr[i][j]")<<endl;
+
+	char arr[TESTN][TESTN];
+	clock_t start, end;
+
+	if (!slowMode) {
+		start =clock();
+		for(int i = 0; i < TESTN; i++) {
+		    for(int j = 0; j < TESTN; j++) {
+		        arr[i][j] = 0;
+		    }
+		}
+		end =clock();
+	    cout<<timediff(start,end)<<endl;
+	}else {
+		start =clock();
+		for(int i = 0; i < TESTN; i++) {
+		    for(int j = 0; j < TESTN; j++) {
+		        arr[j][i] = 0;
+		    }
+		}
+		end =clock();
+		cout<<timediff(start,end)<<endl;
+	}
+}

1-cpu_cache/traverse_2d_array/traverse_2d_array.py

+import time
+
+TESTN = 10240
+arr = [[0 for col in range(TESTN)] for row in range(TESTN)]
+t1 = time.time()
+for i in range(TESTN):
+    for j in range(TESTN):
+        arr[i][j] = 1
+t2 = time.time()
+print t2-t1
+
+t1 = time.time()
+for i in range(TESTN):
+    for j in range(TESTN):
+        arr[j][i] = 1
+t2 = time.time()
+print t2-t1

README.md

-perf stat -e cache-references,cache-misses,instructions,cycles,L1-dcache-load-misses,L1-dcache-loads,L1-icache-load-misses,branch-load-misses,branch-loads ./branch -f
-g++ branch.cpp -o branch
-g++ cpu_migrate.cpp -o cpu_migrate -lpthread
+《极客时间：分布式高性能程序优化30讲》示例代码
+1-cpu_cache目录：第1讲 CPU缓存：怎样写代码能够提升缓存命中率？
\ No newline at end of file