jacobian

commit a6fd46e0f458d0deb3160c54a240f03e37911021
parent 1c820db25f3367af48b601704371484eaa0821da
Author: David Freifeld <freifeld.david@gmail.com>
Date:   Sun, 20 Sep 2020 13:51:51 -0700

Headway into faster file read + parse

Diffstat:
M
src/experimental/fileread.cpp
 | 
28
+++++++++++++++++++++++++---

1 file changed, 25 insertions(+), 3 deletions(-)
diff --git a/src/experimental/fileread.cpp b/src/experimental/fileread.cpp
@@ -14,6 +14,7 @@
 #include <ctime>
 #include <chrono>
 #include <iostream>
+#include <stdint.h>
 
 int main (){
     auto start = std::chrono::high_resolution_clock::now();
@@ -26,10 +27,10 @@ int main (){
 
     /* Advise the kernel of our access pattern.  */
     // posix_fadvise(fd, 0, 0, 1);  // FDADVICE_SEQUENTIAL
-
+    
     char buf[BUFFER_SIZE + 1];
     uintmax_t lines = 0;
-
+    float tmp;
     while(size_t bytes_read = read(fd, buf, BUFFER_SIZE))
     {
         if(bytes_read == (size_t)-1) {
@@ -37,10 +38,21 @@ int main (){
             return 1;
         }
         if (!bytes_read) break;
+        int prev = 0;
         for(char *p = buf; (p = (char*) memchr(p, '\n', (buf + bytes_read) - p)); ++p) {
+            long bound = (char*) memchr(p+1, '\n', (buf + bytes_read) - p+1) - p;
+            tmp = strtod(p, NULL);
+            for (int i = 0; i < bound; i++) {
+                if (*(p+i) ==  ',') {
+                    tmp = strtod(p+i+1, NULL);
+                    //printf("Comma %f\n", tmp);
+                }
+                // printf("%c", *(p+i));
+            }
             ++lines;
         }
     }
+    printf("Final %f\n", tmp);
     auto end = std::chrono::high_resolution_clock::now();
     double time = std::chrono::duration_cast<std::chrono::nanoseconds>(end - start).count();
     
@@ -48,9 +60,19 @@ int main (){
     int flines = 0;
     FILE* fptr = fopen("../../data_banknote_authentication.txt", "r");
     char fbuf[1024];
-    while (fgets(fbuf, 1024, fptr) != NULL) flines++;
+    float ftmp;
+    while (fgets(fbuf, 1024, fptr) != NULL) {
+        char *p;
+        p = strtok(fbuf,",");
+        for (int j = 0; j < 4; j++) {
+            ftmp = strtod(p, NULL);
+            p = strtok(NULL,",");
+        }
+        flines++;
+    }
     auto f_end = std::chrono::high_resolution_clock::now();
     std::cout << flines << " " << lines << "\n";
+    std::cout << ftmp << " " << tmp << "\n";
     double ftime = std::chrono::duration_cast<std::chrono::nanoseconds>(f_end - f_start).count();
     std::cout << ftime << " " << time << " " << ftime/time << "\n";
 }