fixes

.gitignore

@@ -1,3 +1,5 @@
+*.GCC_
+*.exe
 main
 *.o
 .vscode/

ex3/code/main.cpp

@@ -8,10 +8,9 @@
 #include "task_7.h"
 #include "timing.h"
 #include <cblas.h>               // cBLAS Library
-#include <lapacke.h>
-
 #include <iomanip>
 #include <iostream>
+#include <lapacke.h>
 
 void task_1() {
     printf("\n\n-------------- Task 1 --------------\n\n");
@@ -235,31 +234,33 @@ void task_7() {
     // Benchmark
 
     cout << fixed << setprecision(4);
-    size_t NLOOPS = 1000;
+    size_t NLOOPS = 200;
+    size_t Nrhs = 2000;
+    cout << "Solution time per right hand side in milliseconds: sec*1000/Nrhs" << endl;
     cout << "N    = " << "     | 1      | 2      | 4      | 8      | 16     | 32 " << endl;
-    cout <<       "---------|--------|--------|--------|--------|--------|-------" << endl;
+    cout <<       "------------|--------|--------|--------|--------|--------|-------" << endl;
-    for (int exp = 1; exp < 10; ++exp) {
+    for (int k = 1; k < 10; ++k) {
-        cout << "Nrhs = " << static_cast<size_t>(pow(2,exp));
+        cout << "Nrhs = " << Nrhs*k;
         for (size_t N : {1, 2, 4, 8, 16, 32}) {
             tic();
             for (size_t i = 0; i < NLOOPS; ++i) {
-                benchmark_lapacke(N, static_cast<size_t>(pow(2,exp)));
+                benchmark_lapacke(N, Nrhs*k);
             }
-            double sec = toc();
+            double sec = toc()*1000 / (static_cast<double>(Nrhs)*k);
             cout << " | " << sec;
         }
         cout << endl;
     }
-    printf("\nFor fixed n, the solution time per rhs does not slow down consistently and scales very well.\nIts faster than expected.");
+    printf("\nFor fixed n, the solution time per rhs stays roughly constant.");
 }
 
 int main() {
-    task_1();
+    // task_1();
-    task_2();
+    // task_2();
-    task_3();
+    // task_3();
-    task_4();
+    // task_4();
-    task_5();
+    // task_5();
-    task_6();
+    // task_6();
     task_7();
     printf("\n\n");
 

ex3/code/task_3.cpp

@@ -1,7 +1,7 @@
 #include "task_3.h"
-#include <vector>
 #include <cassert>
 #include <cmath>
+#include <vector>
 using namespace std;
 
 

ex3/code/task_4+6.cpp

@@ -2,8 +2,8 @@
 #include "task_4+6.h"
 #include "timing.h"
 #include <cblas.h>               // cBLAS Library
-#include <vector>
 #include <iostream>
+#include <vector>
 using namespace std;
 
 void print_performance(double sec, size_t memory, size_t flops, unsigned int size) {

ex3/code/task_5.cpp

@@ -1,9 +1,9 @@
 #include "task_4+6.h"
 #include "task_5.h"
 #include "timing.h"
-#include <vector>
-#include <iostream>
 #include <cmath>
+#include <iostream>
+#include <vector>
 using namespace std;
 
 double norm(vector<double> const &x) {

ex3/code/task_7.cpp

@@ -1,8 +1,8 @@
 #include "task_7.h"
+#include <cmath>
+#include <iostream>
 #include <lapacke.h>
 #include <vector>
-#include <iostream>
-#include <cmath>
 using namespace std;
 
 tuple<vector<double>, vector<double>> init_M(size_t N, size_t Nrhs) {

ex3/ex3_results.txt

@@ -183,20 +183,20 @@ A * x =
 0.000000 1.000000 
 
 
+Solution time per right hand side in milliseconds: sec*1000/Nrhs
 N    =      | 1      | 2      | 4      | 8      | 16     | 32
----------|--------|--------|--------|--------|--------|-------
+------------|--------|--------|--------|--------|--------|-------
-Nrhs = 2 | 0.0047 | 0.0045 | 0.0046 | 0.0130 | 0.0203 | 0.0476
+Nrhs = 2000 | 0.2122 | 0.0048 | 0.0079 | 0.0544 | 0.0850 | 0.1523
-Nrhs = 4 | 0.0027 | 0.0031 | 0.0033 | 0.0046 | 0.0085 | 0.0250
+Nrhs = 4000 | 0.0036 | 0.0053 | 0.0085 | 0.0142 | 0.0748 | 0.1596
-Nrhs = 8 | 0.0035 | 0.0035 | 0.0045 | 0.0061 | 0.0119 | 0.0300
+Nrhs = 6000 | 0.0033 | 0.0043 | 0.0069 | 0.0147 | 0.0327 | 0.1801
-Nrhs = 16 | 0.0085 | 0.0062 | 0.0221 | 0.0113 | 0.0599 | 0.0757
+Nrhs = 8000 | 0.0036 | 0.0041 | 0.0069 | 0.0197 | 0.0342 | 0.1795
-Nrhs = 32 | 0.0122 | 0.0165 | 0.0112 | 0.0123 | 0.0238 | 0.0834
+Nrhs = 10000 | 0.0028 | 0.0042 | 0.0067 | 0.0140 | 0.0328 | 0.1883
-Nrhs = 64 | 0.0072 | 0.0078 | 0.0164 | 0.0133 | 0.0421 | 0.0666
+Nrhs = 12000 | 0.0025 | 0.0044 | 0.0067 | 0.0157 | 0.0320 | 0.1981
-Nrhs = 128 | 0.0073 | 0.0189 | 0.0269 | 0.0199 | 0.0337 | 0.1041
+Nrhs = 14000 | 0.0023 | 0.0038 | 0.0065 | 0.0157 | 0.0333 | 0.1916
-Nrhs = 256 | 0.0107 | 0.0135 | 0.0279 | 0.0351 | 0.0582 | 0.1438
+Nrhs = 16000 | 0.0022 | 0.0036 | 0.0073 | 0.0140 | 0.0324 | 0.1824
-Nrhs = 512 | 0.0276 | 0.0174 | 0.0237 | 0.1027 | 0.1113 | 0.2417
+Nrhs = 18000 | 0.0028 | 0.0038 | 0.0066 | 0.0145 | 0.0337 | 0.1741
 
-For fixed n, the solution time per rhs does not slow down consistently and scales very well.
+For fixed n, the solution time per rhs stays roughly constant.
-Its faster than expected.
 
 
 -------------- Task 8 --------------

ex3/seq/skalar/compile.log

@@ -1 +0,0 @@
-

ex3/seq/skalar_stl/compile.log

@@ -1 +0,0 @@
-