diff --git a/Projet/CODE/apm/Makefile b/Projet/CODE/apm/Makefile
index ecb8e1884e9c44118e4c59f97f516ff61047d29a..5e1a6860bf1eb672380f1774015d3c3e666fdca9 100644
--- a/Projet/CODE/apm/Makefile
+++ b/Projet/CODE/apm/Makefile
@@ -1,7 +1,8 @@
-ICLUDED=include
+INCLUDED=include
CC=gcc
CFLAGS=-O3 -I $(INCLUDED) -pg -g
+CFLAGS_OMP= -fopenmp -DUSE_OMP $(CFLAGS)
LDFLAGS=
OBJ=obj/apm.o
@@ -12,6 +13,9 @@ all: dir apm apm_omp #apm_gpu
dir:
mkdir -p obj
+obj/apm_omp.o : src/apm_omp.c
+ $(CC) $(CFLAGS_OMP) -c -o $@ $^
+
obj/%.o: src/%.c
$(CC) $(CFLAGS) -c -o $@ $^
@@ -19,7 +23,7 @@ apm: obj/apm.o
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $^
apm_omp: obj/apm_omp.o
- $(CC) $(CFLAGS) $(LDFLAGS) -o $@ $^
+ $(CC) $(CFLAGS_OMP) $(LDFLAGS) -o $@ $^
apm_gpu:
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $^
diff --git a/Projet/CODE/apm/src/.apm.c.swp b/Projet/CODE/apm/src/.apm.c.swp
deleted file mode 100644
index 28e989bebdb020ae307b90cfb0a208c9ff1045df..0000000000000000000000000000000000000000
Binary files a/Projet/CODE/apm/src/.apm.c.swp and /dev/null differ
diff --git a/Projet/CODE/apm/src/apm_omp.c b/Projet/CODE/apm/src/apm_omp.c
index f5db3e57414d6a84a2d1b472364efaca25ed21e4..d40a9cdef1c8db1582b3b5efb5a73b0e8aa58012 100644
--- a/Projet/CODE/apm/src/apm_omp.c
+++ b/Projet/CODE/apm/src/apm_omp.c
@@ -1,6 +1,7 @@
/**
* APPROXIMATE PATTERN MATCHING
*
+ * INF560 X2016
*/
#include <string.h>
#include <stdio.h>
@@ -9,12 +10,13 @@
#include <unistd.h>
#include <sys/time.h>
+#include "omp.h"
+
#define APM_DEBUG 0
-char *
-read_input_file(char *filename, int *size)
+char * read_input_file(char * filename, int * size)
{
- char *buf;
+ char * buf;
off_t fsize;
int fd = 0;
int n_bytes = 1;
@@ -30,7 +32,6 @@ read_input_file(char *filename, int *size)
/* Get the number of characters in the textfile */
fsize = lseek(fd, 0, SEEK_END);
lseek(fd, 0, SEEK_SET);
-
/* TODO check return of lseek */
#if APM_DEBUG
@@ -38,11 +39,11 @@ read_input_file(char *filename, int *size)
#endif
/* Allocate data to copy the target text */
- buf = (char *) malloc(fsize * sizeof (char));
+ buf = (char *)malloc(fsize * sizeof (char));
if (buf == NULL)
{
fprintf(stderr, "Unable to allocate %lld byte(s) for main array\n",
- fsize);
+ fsize);
return NULL;
}
@@ -50,8 +51,8 @@ read_input_file(char *filename, int *size)
if (n_bytes != fsize)
{
fprintf(stderr,
- "Unable to copy %lld byte(s) from text file (%d byte(s) copied)\n",
- fsize, n_bytes);
+ "Unable to copy %lld byte(s) from text file (%d byte(s) copied)\n",
+ fsize, n_bytes);
return NULL;
}
@@ -60,18 +61,13 @@ read_input_file(char *filename, int *size)
#endif
*size = n_bytes;
-
close(fd);
-
return buf;
}
+#define MIN3(a, b, c) ((a)<(b) ? ((a)<(c) ? (a) : (c)) : ((b)<(c) ? (b) : (c)))
-#define MIN3(a, b, c)\
- ((a) < (b) ? ((a) < (c) ? (a) : (c)) : ((b) < (c) ? (b) : (c)))
-
-int
-levenshtein(char *s1, char *s2, int len, int * column)
+int levenshtein(char *s1, char *s2, int len, int * column)
{
unsigned int x, y, lastdiag, olddiag;
@@ -82,15 +78,13 @@ levenshtein(char *s1, char *s2, int len, int * column)
for (x = 1; x <= len; x++)
{
column[0] = x;
- lastdiag = x-1 ;
+ lastdiag = x-1;
for (y = 1; y <= len; y++)
{
olddiag = column[y];
- column[y] = MIN3(
- column[y] + 1,
+ column[y] = MIN3(column[y] + 1,
column[y-1] + 1,
- lastdiag + (s1[y-1] == s2[x-1] ? 0 : 1)
- );
+ lastdiag + (s1[y-1] == s2[x-1] ? 0 : 1));
lastdiag = olddiag;
}
}
@@ -98,87 +92,79 @@ levenshtein(char *s1, char *s2, int len, int * column)
}
-int
-main(int argc, char **argv)
+int main(int argc, char ** argv)
{
- char **pattern;
- char *filename;
+ char ** pattern;
+ char * filename;
int approx_factor = 0;
int nb_patterns = 0;
int i, j;
- char *buf;
+ char * buf;
struct timeval t1, t2;
double duration;
int n_bytes;
- int *n_matches;
+ int * n_matches;
/* Check number of arguments */
if (argc < 4)
{
printf("Usage: %s approximation_factor "
- "dna_database pattern1 pattern2 ...\n",
- argv[0]);
+ "dna_database pattern1 pattern2 ...\n",
+ argv[0]);
return 1;
}
- /* Get the distance factor */
- approx_factor = atoi(argv[1]);
-
- /* Grab the filename containing the target text */
- filename = argv[2];
-
- /* Get the number of patterns that the user wants to search for */
- nb_patterns = argc - 3;
-
- /* Fill the pattern array */
- pattern = (char **)malloc(nb_patterns * sizeof(char *));
- if (pattern == NULL)
+ approx_factor = atoi(argv[1]);/* Get the distance factor */
+ filename = argv[2];/* Grab the filename containing the target text */
+ nb_patterns = argc - 3;/* Get the number of patterns to search for */
+
+ pattern = (char **)malloc(nb_patterns * sizeof(char*));
+ if (pattern == NULL)/*Fill the pattern*/
{
fprintf(stderr,
- "Unable to allocate array of pattern of size %d\n",
- nb_patterns );
+ "Unable to allocate array of pattern of size %d\n",
+ nb_patterns);
return 1;
}
- /* Grab the patterns */
- for (i = 0; i < nb_patterns; i++)
+ for (i=0; i < nb_patterns; i++) /* Grab the patterns */
{
int l;
-
l = strlen(argv[i+3]);
- if (l <= 0)
+
+ if (l <= 0)
{
- fprintf(stderr, "Error while parsing argument %d\n", i + 3);
+ fprintf(stderr, "Error while parsing argument %d\n", i+3);
return 1;
}
-
- pattern[i] = (char *) malloc((l+1) * sizeof(char));
+
+ pattern[i] = (char *)malloc((l+1) * sizeof(char));
if (pattern[i] == NULL)
{
fprintf(stderr, "Unable to allocate string of size %d\n", l);
return 1;
}
- strncpy(pattern[i], argv[i + 3], (l + 1));
+ strncpy(pattern[i], argv[i+3], (l+1));
}
printf("Approximate Pattern Mathing: "
- "looking for %d pattern(s) in file %s w/ distance of %d\n",
- nb_patterns, filename, approx_factor);
+ "looking for %d pattern(s) in file %s w/ distance of %d\n",
+ nb_patterns, filename, approx_factor);
buf = read_input_file(filename, &n_bytes);
if (buf == NULL)
{
+ fprintf(stderr, "Error: NULL pointer from reading input file.");
return 1;
}
- /* Allocate the array of matches */
- n_matches = (int *) malloc(nb_patterns * sizeof(int));
+ n_matches = (int *)malloc(nb_patterns * sizeof(int));/*Alloc the matches*/
if (n_matches == NULL)
{
fprintf(stderr, "Error: unable to allocate memory for %ldB\n",
- nb_patterns * sizeof(int));
+ nb_patterns * sizeof(int));
return 1;
}
@@ -191,26 +177,25 @@ main(int argc, char **argv)
for (i = 0; i < nb_patterns; i++)
{
-
int size_pattern = strlen(pattern[i]);
-
int * column;
+ int distance = 0;
n_matches[i] = 0;
- column = (int *) malloc((size_pattern + 1) * sizeof(int));
+ column = (int *)malloc((size_pattern+1) * n_bytes * sizeof(int));
if (column == NULL)
{
- fprintf(stderr, "Error: unable to allocate memory for column "
- "(%ldB)\n",
- (size_pattern + 1) * sizeof(int));
+ fprintf(stderr,
+ "Error: unable to allocate memory for column (%ldB)\n",
+ (size_pattern+1) * sizeof(int));
return 1;
}
+ int matches = 0;
+ #pragma omp parallel for reduction(+:matches) private(distance)
for (j = 0; j < n_bytes; j++)
{
- int distance = 0;
- int size;
#if APM_DEBUG
if (j % 100 == 0)
@@ -219,27 +204,31 @@ main(int argc, char **argv)
}
#endif
- size = size_pattern;
+ size_pattern;
if (n_bytes - j < size_pattern)
{
- size = n_bytes - j;
+ //size = n_bytes - j;
+ //NO ! we do not want to match substring of our input, wth
+ continue;
}
- distance = levenshtein(pattern[i], &buf[j], size, column);
+ distance = levenshtein(pattern[i], &buf[j], size_pattern,
+ column+(j*(size_pattern+1)));
if (distance <= approx_factor)
{
- n_matches[i]++;
+ matches++;
}
}
+ n_matches[i] = matches;
free(column);
}
/* Timer stop */
gettimeofday(&t2, NULL);
- duration = (t2.tv_sec - t1.tv_sec) + ((t2.tv_usec - t1.tv_usec) / 1e6);
+ duration = (t2.tv_sec -t1.tv_sec)+((t2.tv_usec-t1.tv_usec)/1e6);
printf("APM done in %lf s\n", duration);
@@ -250,7 +239,7 @@ main(int argc, char **argv)
for (i = 0; i < nb_patterns; i++)
{
printf("Number of matches for pattern <%s>: %d\n",
- pattern[i], n_matches[i]);
+ pattern[i], n_matches[i]);
}
return 0;