Material Práctica 6

S6/material/BAJA_PRIO_FIB

@@ -0,0 +1,8 @@
+#!/bin/bash
+
+i=0
+while [ $i -lt $1 ]
+do
+	/usr/bin/time nice ./fib 45&
+	i=`expr $i + 1` #incrementa en 1 la variable i
+done

S6/material/FIB

@@ -0,0 +1,17 @@
+#!/bin/bash
+
+if [ "$#" -ne 1 ]
+then
+	echo "usage $0 num" 
+	echo "num: numero de procesos fibonacci a ejecutar en paralelo"
+	exit
+fi
+
+echo $1
+
+i=0
+while [ $i -lt $1 ]
+do
+	/usr/bin/time ./fib 45&
+	i=`expr $i + 1` 
+done

S6/material/Makefile

@@ -0,0 +1,10 @@
+all:fibonacci mem1
+
+fibonacci:fibonacci.c
+	gcc -o fib fibonacci.c
+
+mem1:mem1.c
+	gcc -o mem1 mem1.c
+
+clean:
+	rm -f fib mem1

S6/material/fibonacci.c

@@ -0,0 +1,29 @@
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+
+long fibonacci(unsigned i) {
+	if (i == 0) return 0;
+	if (i == 1) return 1;
+	return fibonacci(i-1) + fibonacci(i-2);
+}
+
+void usage() {
+	char buffer[80];
+
+	strcpy(buffer,"fib n\n");
+	write(1,buffer,strlen(buffer));
+	exit(1);
+}
+void main(int argc,char *argv[]) {
+	unsigned n;
+	long f;
+	char buffer[80];
+
+	if (argc != 2) usage();
+	n = atoi(argv[1]);
+	f = fibonacci(n);
+	sprintf(buffer,"El fibonacci de %u es %ld\n",n,f);
+	write(1, buffer, strlen(buffer));
+}
+

S6/material/mem1.c

@@ -0,0 +1,79 @@
+#include <stdlib.h>
+#include <stdio.h>
+#include <unistd.h>
+#include <string.h>
+
+#define REGION_SIZE 256*1024*1024
+
+/*
+ * rdtsc is an assembler instruction that get the cycles executed by the processor since boot time.
+ */
+#define getCycles(low,high) __asm__ __volatile__("rdtsc": "=a"(low), "=d"(high))
+
+static inline unsigned long long getTime(void)
+{
+	unsigned long low, high;
+	getCycles(low, high);
+	return ((unsigned long long)high << 32) + low;
+
+}
+
+int main(int argc, char **argv)
+{
+	int i, j;
+	int *p;
+	unsigned long long initTime;
+	unsigned long long endTime;
+	long long region_size;
+	int nelts;
+	int nits;
+	int nprocs;
+	int ret;
+	char buff[256];					
+
+	if (argc != 4) {
+		sprintf( buff, "usage: mem2_4 region_size processes iterations \n");
+		write(1, buff, strlen(buff));	
+		exit(1);
+	}
+
+	region_size = atoll(argv[1]);
+	nprocs = atoi(argv[2]);
+	nits = atoi(argv[3]);
+
+	// compute the number of integers that fits in the region
+	nelts = region_size / sizeof(int);
+
+	//loop to create the number of processes required
+	i = 1;
+	ret = 1;
+	while ((i < nprocs) && (ret > 0)) {
+		ret = fork();
+		i++;
+	}
+
+	//code that executes each process
+	p = malloc(region_size);
+	if (p == NULL) {
+		sprintf( buff, "Unable to allocate the memory region\n");
+		write(1, buff, strlen(buff));	
+		exit(1);
+	}
+	//loop to control how many times each process accesses all the elements of the vector
+	i = 0;
+	while (i < nits) {
+		initTime = getTime();
+		//loop to access all the elements in the vector
+		for (j = 0; j < nelts; j++)
+			p[j] = j;
+		endTime = getTime();
+		sprintf( buff, "process: %d, iteration %d, access loop time: %llu\n",
+		      getpid(), i, (endTime - initTime));
+		write(1, buff, strlen(buff));	
+		i++;
+
+	}
+	sprintf( buff, "Entering endless loop\n");
+	write(1, buff, strlen(buff));	
+	while (1) ;
+}