ConvertData/graph.py at master · laBobberto/ConvertData · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
import csv
import os
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

file_map = {
    'clang_linux_benchmark_analysis.csv': 'Clang (Linux)',
    'g++_linux_benchmark_analysis.csv': 'G++ (Linux)',
    'mingw_win_benchmark_analysis.csv': 'MinGW (Win, per-call)',
    'msvc_win_benchmark_analysis.csv': 'MSVC (Win, per-call)',
    'time_mingw_win_benchmark_analysis.csv': 'MinGW (Win, batch)',
    'time_msvc_win_benchmark_analysis.csv': 'MSVC (Win, batch)'
}

metrics = ['Average_ns', 'Median_ns', 'P95_ns', 'P99_ns']
all_results = []

for filename, source_name in file_map.items():
    if not os.path.exists(filename):
        continue

    try:
        with open(filename, mode='r', encoding='utf-8') as f:
            reader = csv.DictReader(f)
            source_data = list(reader)

            baseline = None
            for row in source_data:
                version = row['Version'].strip('_')
                if version == 'Original':
                    baseline = float(row['Average_ns'])
                    break

            if baseline is None:
                continue

            for row in source_data:
                version = row['Version'].strip('_')
                avg_ns = float(row['Average_ns'])
                median_ns = float(row['Median_ns'])
                p95_ns = float(row['P95_ns'])
                p99_ns = float(row['P99_ns'])
                speedup = baseline / avg_ns if avg_ns > 0 else 0

                all_results.append({
                    'Source': source_name,
                    'Version': version,
                    'Average_ns': avg_ns,
                    'Median_ns': median_ns,
                    'P95_ns': p95_ns,
                    'P99_ns': p99_ns,
                    'Speedup': speedup
                })
    except Exception as e:
        print(f"Error processing {filename}: {e}")

if not all_results:
    print("No results to process.")
    exit()

df = pd.DataFrame(all_results)
df.to_csv('performance_comparison_detailed.csv', index=False)

# Main Summary Plot (Average & Speedup)
sns.set_theme(style="whitegrid", context="talk")
fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(14, 12))

sns.barplot(data=df, x='Source', y='Average_ns', hue='Version', ax=ax1)
ax1.set_title('Average Execution Time (ns) - Lower is Better')
ax1.set_ylabel('Time (ns)')
ax1.set_xlabel('')
ax1.tick_params(axis='x', rotation=15)
ax1.legend(title='Version', bbox_to_anchor=(1.05, 1), loc='upper left')

sns.barplot(data=df, x='Source', y='Speedup', hue='Version', ax=ax2)
ax2.set_title('Speedup vs Original (Average) - Higher is Better')
ax2.set_ylabel('Speedup (x)')
ax2.set_xlabel('Compiler / Platform')
ax2.tick_params(axis='x', rotation=15)
ax2.axhline(1, ls='--', color='red', alpha=0.5)
ax2.legend(title='Version', bbox_to_anchor=(1.05, 1), loc='upper left')

plt.tight_layout()
plt.savefig('benchmark_summary_charts.png')

# Detailed Plot (Tail Latencies)
fig_det, (ax3, ax4) = plt.subplots(2, 1, figsize=(14, 12))

df_melted_tail = df.melt(id_vars=['Source', 'Version'], value_vars=['P95_ns', 'P99_ns'], var_name='Metric', value_name='Time_ns')
sns.barplot(data=df_melted_tail, x='Source', y='Time_ns', hue='Version', ax=ax3)
ax3.set_title('Tail Latency (P95 & P99) - Lower is Better')
ax3.set_ylabel('Time (ns)')
ax3.set_xlabel('')
ax3.tick_params(axis='x', rotation=15)
ax3.legend(title='Version', bbox_to_anchor=(1.05, 1), loc='upper left')

# Speedup Heatmap (Version vs Source)
pivot_speedup = df.pivot(index='Version', columns='Source', values='Speedup')
sns.heatmap(pivot_speedup, annot=True, fmt=".2f", cmap="YlGnBu", ax=ax4)
ax4.set_title('Speedup Relative to Original (Heatmap)')

plt.tight_layout()
plt.savefig('benchmark_summary_charts_detailed.png')

print("\n=== SUMMARY ANALYSIS ===")
for source in df['Source'].unique():
    source_df = df[df['Source'] == source]
    best_row = source_df.loc[source_df['Speedup'].idxmax()]
    print(f"{source:<25}: Best is {best_row['Version']} ({best_row['Speedup']:.2f}x speedup)")

print("\nCharts updated:")
print("- benchmark_summary_charts.png: Average Time and Speedup")
print("- benchmark_summary_charts_detailed.png: Tail Latencies and Speedup Heatmap")
print("- performance_comparison_detailed.csv: Raw data table")