Лабы по Шевченко за последний семестр

Шевченко И. В/iptables

@@ -0,0 +1,5 @@
+iptables -F
+iptables -t nat -F
+iptables -t nat -A POSTROUTING -o wlan0 -j MASQUERADE
+iptables -t nat -A PREROUTING -i wlan0 -j DNAT -p tcp --dport 2222 --to 192.168.57.102:22
+

Шевченко И. В/Параллельное программирование/Битоническая сортировка/Makefile

@@ -0,0 +1,23 @@
+CXXFLAGS=-O3 -fopenmp
+all:posled parallel pthread openmp intrin
+
+pthread:pthread.cpp
+
+openmp:openmp.cpp
+
+posled:posled.cpp
+
+intrin:intrin.cpp
+	$(CXX) -O3 -msse2 intrin.cpp -o intrin
+
+parallel:parallel.cpp
+	mpicxx parallel.cpp -o parallel
+test:parallel
+	mpiexec -n 4 ~/.mpi/parallel
+clean:
+	if [ -f posled ];then rm posled;fi
+	if [ -f parallel ];then rm parallel;fi
+	if [ -f openmp ];then rm openmp;fi
+	if [ -f pthread ];then rm pthread;fi
+	if [ -f intrin ];then rm intrin;fi
+

Шевченко И. В/Параллельное программирование/Битоническая сортировка/intrin.cpp

@@ -0,0 +1,163 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <memory.h>
+#include <time.h>
+#include <iostream>
+#include <algorithm>
+#include <vector>
+#include <emmintrin.h>
+
+short* a;
+int n=1<<21;
+int iter=0;
+
+void o(){
+	for(int i=0;i<n;i++)printf("%i ",a[i]);
+	printf("\n");
+}
+//ïðèìåíåíèå ïîëóî÷èñòèòåëÿ k
+//k=4, k=2
+void half_cleaner(const int start, const int end, const bool asc, const int k){
+	for(int i=start;i<end;i+=k)
+		for(int j=0;(j<k/2)&&(i+j+k/2<end);++j)
+			if((a[i+j]<a[i+j+k/2])==asc){
+				short t=a[i+j];
+				a[i+j]=a[i+j+k/2];
+				a[i+j+k/2]=t;
+			}
+	iter++;
+}
+//k=8
+void half_cleaner_8(const int start, const int end, const bool asc){
+	for(int i=start;i+7<end;i+=8)
+	{
+		__m64 t = *(__m64*)&a[i];
+		__m64 d = *(__m64*)&a[i+4];
+		__m64 mn = _mm_min_pi16(t,d);
+		__m64 mx = _mm_max_pi16(t,d);
+		if (asc) {
+			*(__m64*)&a[i]=mx;
+			*(__m64*)&a[i+4]=mn;
+		}
+		else {
+			*(__m64*)&a[i]=mn;
+			*(__m64*)&a[i+4]=mx;
+		}	
+	}
+	iter++;
+}
+//k=16
+void half_cleaner_16(const int start, const int end, const bool asc){
+	for(int i=start;i+15<end;i+=16)
+	{
+		//êñòàòè a[i] çäåñü âûðàâíåíî ïîä 16, ïîýòîìó ìîæíî áåç u, íî âðåìåíè ýòî íå ïðèáàâëÿåò
+		__m128i t = _mm_loadu_si128((__m128i*)&a[i]);
+		__m128i d = _mm_loadu_si128((__m128i*)&a[i+8]);
+		__m128i mn = _mm_min_epi16(t,d);
+		__m128i mx = _mm_max_epi16(t,d);
+		if (asc) {
+			_mm_storeu_si128((__m128i*)&a[i],mx);
+			_mm_storeu_si128((__m128i*)&a[i+8],mn);
+		}
+		else {
+			_mm_storeu_si128((__m128i*)&a[i],mn);
+			_mm_storeu_si128((__m128i*)&a[i+8],mx);
+		}	
+	}
+	iter++;
+}
+//k>16
+void half_cleaner_gt16(const int start, const int end, const bool asc, const int k){
+	for(int i=start;i<end;i+=k)
+	{
+		for(int j=0;(j<k/2)&&(i+j+k/2<end);j+=8)
+		{
+			__m128i t = _mm_loadu_si128((__m128i*)&a[i+j]);
+			__m128i d = _mm_loadu_si128((__m128i*)&a[i+j+k/2]);
+			__m128i mn = _mm_min_epi16(t,d);
+			__m128i mx = _mm_max_epi16(t,d);
+			if (asc) {
+				_mm_storeu_si128((__m128i*)&a[i+j],mx);
+				_mm_storeu_si128((__m128i*)&a[i+j+k/2],mn);
+			}
+			else {
+				_mm_storeu_si128((__m128i*)&a[i+j],mn);
+				_mm_storeu_si128((__m128i*)&a[i+j+k/2],mx);
+			}
+			/*for(int j=0;(j<k/2)&&(i+j+k/2<end);++j)
+			if((a[i+j]<a[i+j+k/2])==asc){
+				int t=a[i+j];
+				a[i+j]=a[i+j+k/2];
+				a[i+j+k/2]=t;
+			}*/
+		}
+
+	}
+	iter++;
+}
+//ïðèìåíåíèå ïîëóî÷èñòèòåëÿ k ñ ÷åðåäîâàíèåì íàïðàâëåíèÿ óïîðÿäî÷èâàíèÿ
+// k=2
+void half_cleaner_alter(int k){
+	bool asc=false;
+	for(int i=0;i<n;i+=k)
+	{	for(int j=0;(j<k/2)&&(i+j+k/2<n);++j)
+			if((a[i+j]<a[i+j+k/2])==asc){
+				short t=a[i+j];
+				a[i+j]=a[i+j+k/2];
+				a[i+j+k/2]=t;
+			}
+		asc=!asc;
+	}
+}
+//ñîðòèðîâêà áèòîíè÷åñêèì ñëÿíèåì k
+void bitonic_sort(int s,int e,bool a,int k)
+{
+	while(k!=1) {
+		if(k>16) half_cleaner_gt16(s,e,a,k);
+		else if(k<8) half_cleaner(s,e,a,k);
+		else if(k==8) half_cleaner_8(s,e,a);
+		else if(k==16) half_cleaner_16(s,e,a);
+		else half_cleaner(s,e,a,k);
+		k=(k/2);
+	}
+}
+
+void all_sort() {
+	//ìîæíî óñêîðèòü, åñëè ñäåëàòü øàã "ïèëû" íå 2, à 16: ïðèìåíÿòü bitonic_sort_16elems îòñþäà:
+	//github.com/hayamiz/sse-practice-sort/blob/master/sse-practice.c
+	half_cleaner_alter(2);
+	for(int k=4;k<=n;k<<=1){
+		bool asc=false;
+		for(int i=0;i+k<=n;i+=k)
+		{
+			bitonic_sort(i,i+k,asc,k);
+			asc=!asc;
+		}
+	}
+}
+#ifdef __GNUC__
+#define MALLOC(pplist,num)\
+	posix_memalign(pplist, 16, sizeof(short) * num)
+#else
+#include <malloc.h>
+#define MALLOC(pplist,num)\
+	*(pplist)=_aligned_malloc(sizeof(short) * num, 16)
+#endif
+
+int main(int argc, char* argv[])
+{
+	MALLOC((void**)&a,n);
+	for(int i=0;i<n;++i)a[i]=rand()%100;
+
+	time_t start=clock();
+	all_sort();
+	time_t end = clock();
+
+	bool ok=true;
+	for(int i=1;i<n;++i)if(a[i-1]>a[i])ok=false;
+	double workTime = end-start;
+	workTime /= CLOCKS_PER_SEC;
+	if(ok)printf("Iter: %i, Calculation takes: %.5f sec\n", iter, workTime);
+	return 0;
+}
+

Шевченко И. В/Параллельное программирование/Битоническая сортировка/openmp.cpp

@@ -0,0 +1,93 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <memory.h>
+#include <time.h>
+#include <iostream>
+#include <algorithm>
+#include <vector>
+#include <omp.h>
+
+int a[1<<21];
+int n=1<<21;
+int iter=0;
+
+void o(){
+	for(int i=0;i<n;i++)printf("%i ",a[i]);
+	printf("\n");
+}
+//ïðèìåíåíèå ïîëóî÷èñòèòåëÿ k
+void half_cleaner(int start, int end, bool asc, int k){
+	for(int i=start;i<end;i+=k)
+		for(int j=0;(j<k/2)&&(i+j+k/2<end);++j)
+			if((a[i+j]<a[i+j+k/2])==asc){
+				int t=a[i+j];
+				a[i+j]=a[i+j+k/2];
+				a[i+j+k/2]=t;
+			}
+	iter++;
+}
+//ïðèìåíåíèå ïîëóî÷èñòèòåëÿ k ñ ÷åðåäîâàíèåì íàïðàâëåíèÿ óïîðÿäî÷èâàíèÿ
+void half_cleaner_alter(int k){
+	bool asc=false;
+	for(int i=0;i<n;i+=k)
+	{	for(int j=0;(j<k/2)&&(i+j+k/2<n);++j)
+			if((a[i+j]<a[i+j+k/2])==asc){
+				int t=a[i+j];
+				a[i+j]=a[i+j+k/2];
+				a[i+j+k/2]=t;
+			}
+		asc=!asc;
+	}
+}
+//ñîðòèðîâêà áèòîíè÷åñêèì ñëÿíèåì k
+void bitonic_sort(int s,int e,bool a,int k)
+{
+	while(k!=1) {
+		half_cleaner(s,e,a,k);
+		k=(k/2);
+	}
+}
+void all_sort() {
+	half_cleaner_alter(2);
+	int k=4;
+	for(;k<=n/4;k<<=1)
+#pragma omp parallel firstprivate(k) shared(a,n)
+		{
+			bool asc=omp_get_thread_num()%2;
+#pragma omp for 
+			for(int i=0;i<=n-k;i+=k)
+			{
+				bitonic_sort(i,i+k,asc,k);
+				asc=!asc;
+			}
+		}
+
+	for(;k<=n;k<<=1){bool asc=false;
+	for(int i=0;i+k<=n;i+=k)
+	{
+		bitonic_sort(i,i+k,asc,k);
+		asc=!asc;
+	}}
+}
+
+//bool desc (int i,int j) { return (i>j); }
+int main(int argc, char* argv[])
+{
+	for(int i=0;i<n;++i)a[i]=rand()%100;
+	//std::vector<int> myvector (a, a+n);
+	//sort(myvector.begin(), myvector.begin()+n/2);
+	//sort(myvector.begin()+n/2, myvector.begin()+n,desc);
+	//memcpy( a, &myvector[0], sizeof( int ) * myvector.size() );
+
+	time_t start=clock();
+	all_sort();
+	time_t end = clock();
+
+	bool ok=true;
+	for(int i=1;i<n;++i)if(a[i-1]>a[i])ok=false;
+	double workTime = end-start;
+	workTime /= CLOCKS_PER_SEC;
+	if(ok)printf("Iter: %i, Calculation takes: %.5f sec\n", 4*iter, workTime);
+	return 0;
+}
+