WIP: update new profiling data

Author: Mes0903

statistic/README

@@ -3,3 +3,4 @@
 3: Intel(R) Core(TM) i7-1065G7 CPU @ 1.30GHz
 4: arm (ThunderX2-99xx)
 5: mag
+6: M2 Pro Max

statistic/boot time difference.png

@@ -1,46 +1,42 @@
 #!/usr/bin/env python3
 """
-plot_results.py
+draw.py
 
-Reads 5 "results_summary_{N}.txt" files, each containing columns:
+We have 6 data files:
+  results_summary_1.txt
+  results_summary_2.txt
+  results_summary_3.txt
+  results_summary_4.txt
+  results_summary_5.txt
+  results_summary_6.txt
+
+Each has 11 columns (skip the first row as a header):
  1) SMP
  2) real_boot_time
- 3) times_call_semu_timer_clocksource
+ 3) times_called
  4) ns_per_call
  5) predict_sec
  6) scale_factor
- 7) percentage
- 8) real_ns_per_call
-
-We skip the header row in each file. For each column 2..8, we create
-one figure, plotting SMP on the x-axis vs. that column's values
-on the y-axis. We have 5 lines in each plot, one line per environment.
-
-Additionally, we create one more figure for the difference:
-  (predict_sec [col 5]) - (real_boot_time [col 2])
-
-Environment CPU Descriptions:
-1: 13th Gen Intel(R) Core(TM) i7-13700
-2: Intel(R) Xeon(R) CPU E5-2640 v4 @ 2.40GHz
-3: Intel(R) Core(TM) i7-1065G7 CPU @ 1.30GHz
-4: arm (ThunderX2-99xx)
-5: mag
+ 7) test_total_clocksource_ns
+ 8) real_total_clocksource_ns
+ 9) percentage
+ 10) real_ns_per_call
+ 11) diff_ns_per_call
+
+We want to plot columns 2..11 (10 columns) vs. column 1 (SMP).
+Then create 1 more difference plot for (predict_sec - real_boot_time).
+
+Thus, we produce 11 plots total:
+ - 10 from col=2..11
+ - 1 difference plot (predict_sec - real_boot_time).
 """
 
 import matplotlib
-# Force a non-interactive backend (no hover tooltips):
-matplotlib.use("Agg")
+matplotlib.use("Agg")  # for non-interactive environments
 
 import numpy as np
 import matplotlib.pyplot as plt
 
-# Filenames expected
-#   results_summary_1.txt
-#   results_summary_2.txt
-#   results_summary_3.txt
-#   results_summary_4.txt
-#   results_summary_5.txt
-
 ENV_FILES = [
     "results_summary_1.txt",
     "results_summary_2.txt",
@@ -56,91 +52,101 @@
     "3: Env3",
     "4: Env4",
     "5: Env5",
-    "6: Env6"
+    "6: Env6",
 ]
 
-# Map from column index to a label (column 1 is SMP)
-# We want columns 2..8
+# A dictionary for column labels (2..11):
 COL_LABELS = {
-    2: "real boot time (sec)",
-    3: "times call 'semu_timer_clocksource'",
-    4: "ns_per_call",
-    5: "predict boot time (sec)",
+    2: "real_boot_time",
+    3: "times_called",
+    4: "predict_ns_per_call",
+    5: "predict_boot_time",
     6: "scale_factor",
+    7: "test_total_clocksource_ns",
+    8: "real_total_clocksource_ns",
+    9: "percentage",
+    10: "real_ns_per_call",
+    11: "diff_ns_per_call",
 }
 
 def main():
-    # 1) Plot each data column from 2..8
-    for col in range(2, 7):
+    # 1) Create a figure for each column from 2..11
+    for col in range(2, 12):  # 2..11 inclusive => 10 columns
         label = COL_LABELS[col]
 
         plt.figure(figsize=(8, 6))
         plt.title(f"{label} vs. SMP")
         plt.xlabel("SMP")
         plt.ylabel(label)
 
+        # For each environment data
         for i, fname in enumerate(ENV_FILES):
             try:
-                # Load numeric data, skipping first row (header).
+                # skiprows=1 => skip header line
+                # Load all columns => we have 11 total
                 data = np.loadtxt(fname, skiprows=1)
             except OSError:
-                print(f"Warning: could not open {fname}, skipping.")
+                print(f"Warning: could not open {fname}. Skipping.")
+                continue
+            except ValueError as e:
+                print(f"ValueError loading {fname}: {e}")
                 continue
 
-            if data.shape[1] < 6:
-                print(f"Warning: {fname} does not have at least 8 columns, skipping.")
+            # We expect at least 11 columns
+            if data.shape[1] < 11:
+                print(f"Warning: {fname} does not have 11 columns. Found shape: {data.shape}")
                 continue
 
-            x = data[:, 0]              # SMP is col 1 => index=0
-            y = data[:, col - 1]       # col -> zero-based
+            x = data[:, 0]           # SMP is col=1 => index=0
+            y = data[:, col - 1]     # col => zero-based index
 
             plt.plot(x, y, marker='o', label=ENV_NAMES[i])
 
         plt.legend(loc="best")
         plt.grid(True)
-
-        outname = f"{label}.png"
-        plt.savefig(outname, dpi=150)
+        outfile = f"{label}.png"
+        plt.savefig(outfile, dpi=150)
         plt.close()
-        print(f"Created {outname}")
-
-    print("All standard column plots generated!")
-
-    # 2) Create an additional figure for (predict_sec - real_boot_time).
-    #    predict_sec is column 5 => index 4
-    #    real_boot_time is column 2 => index 1
-    label_diff = "boot time difference"  # name of the new figure
+        print(f"Created {outfile}")
 
+    # 2) Create an additional difference plot for (predict_sec - real_boot_time).
+    #    predict_sec => col=5 => index=4
+    #    real_boot_time => col=2 => index=1
+    label_diff = "diff_boot_time"
     plt.figure(figsize=(8, 6))
     plt.title(f"{label_diff} vs. SMP")
     plt.xlabel("SMP")
-    plt.ylabel("boot time difference (seconds)")
+    plt.ylabel("predict_sec - real_boot_time")
+
     for i, fname in enumerate(ENV_FILES):
         try:
             data = np.loadtxt(fname, skiprows=1)
         except OSError:
-            print(f"Warning: could not open {fname}, skipping.")
+            print(f"Warning: could not open {fname}. Skipping.")
+            continue
+        except ValueError as e:
+            print(f"ValueError loading {fname}: {e}")
             continue
 
-        if data.shape[1] < 6:
-            print(f"Warning: {fname} does not have at least 8 columns, skipping.")
+        if data.shape[1] < 11:
+            print(f"Warning: {fname} does not have 11 columns. Found shape: {data.shape}")
             continue
 
-        x = data[:, 0]  # SMP
-        predict_sec = data[:, 4]      # col5 => index=4
-        real_bt = data[:, 1]         # col2 => index=1
-        diff = predict_sec - real_bt # difference
+        x = data[:, 0]
+        predict_sec = data[:, 4]     # col=5 => index=4
+        real_bt = data[:, 1]        # col=2 => index=1
+        diff = predict_sec - real_bt
 
         plt.plot(x, diff, marker='o', label=ENV_NAMES[i])
 
     plt.legend(loc="best")
     plt.grid(True)
-    outdiff = f"{label_diff}.png"
-    plt.savefig(outdiff, dpi=150)
+    outfile_diff = f"{label_diff}.png"
+    plt.savefig(outfile_diff, dpi=150)
     plt.close()
-    print(f"Created {outdiff}")
+    print(f"Created {outfile_diff}")
 
-    print("Done! All plots generated including difference (predict - real_boot_time).")
+    print("Done! Generated 10 standard plots (cols=2..11) + 1 difference plot (predict - real).")
 
 if __name__ == "__main__":
     main()