update automationto automatically detect devices and visualize results

Author: sheralskumar

automation/run_benchmarks.py

@@ -3,6 +3,9 @@
 import os
 import sys
 from pathlib import Path
+import pyopencl as cl
+import matplotlib.pyplot as plt
+import pandas as pd
 
 # Define path to the benchmark executable
 SCRIPT_DIR = Path(__file__).parent.parent
@@ -11,6 +14,29 @@
 SAVE_DIR = SCRIPT_DIR / "automation" / "benchmark_results.csv"
 
 
+# Automatically detect all GPU devices
+def get_gpu_devices():
+    platforms = cl.get_platforms()
+    devices = [dev for p in platforms for dev in p.get_devices(device_type=cl.device_type.GPU)]
+
+    unique_devices = []
+    seen_names = set()
+
+    for dev in devices:
+        if dev.name not in seen_names:
+            unique_devices.append(dev)
+            seen_names.add(dev.name)
+
+    for i, dev in enumerate(unique_devices):
+        print(f"{i}: {dev.name}")
+        print(f"  Vendor: {dev.vendor}")
+        print(f"  Max Compute Units: {dev.max_compute_units}")
+        print(f"  Max Clock Frequency: {dev.max_clock_frequency} MHz")
+        print(f"  Global Memory Size: {dev.global_mem_size // (1024**2)} MB")
+        print(f"  Local Memory Size: {dev.local_mem_size // 1024} KB")
+
+    return unique_devices
+
 
 def run_benchmark(device, N):
     if not BENCHMARK_EXECUTABLE.exists():
@@ -33,26 +59,45 @@ def run_benchmark(device, N):
         print(f"Exception occurred while running benchmark for device {device} with N={N}: {e}")
         return None
 
+def plot_results():
+    df = pd.read_csv(SAVE_DIR)
+
+    for device_index in df["Device"].unique():
+        subset = df[df["Device"] == device_index]
+        plt.plot(subset["N"], subset["Time (ms)"], label=f"Device {device_index}")
+
+    plt.xlabel("N (size)")
+    plt.ylabel("Time (s)")
+    plt.xscale("log", base=2)
+    plt.yscale("log")
+    plt.title("GPU Benchmark Results")
+    plt.legend()
+    plt.grid(True)
+    plt.savefig(SCRIPT_DIR / "automation" / "benchmark_results.png")
+    plt.show()
+
 def main():
-    devices = [0, 1]  # List of device IDs to benchmark
-    N_values = [1 << 22, 1 << 24, 1 << 26]  # Different sizes for the benchmark
+    devices = get_gpu_devices() # List of device IDs to benchmark
+    N_values = [1 << 22, 1 << 24, 1 << 26, 1 << 28]  # Different sizes for the benchmark
 
     results = []
     for N in N_values:
-        for device in devices:
-            print(f"Running benchmark for device {device} with N={N}...")
-            output = run_benchmark(device, N)
+        for i, dev in enumerate(devices):
+            print(f"Running benchmark for device {i} with N={N}...")
+            output = run_benchmark(i, N)
             if output:
-                results.append((device, N, output))
+                results.append((i, N, output))
                 print(f"Result: {output}")
 
     # Save results to CSV
     with open(SAVE_DIR, "w", newline='') as csvfile:
         csvwriter = csv.writer(csvfile)
-        csvwriter.writerow(["Device", "N", "Output"])
+        csvwriter.writerow(["Device", "N", "Time (ms)"])
         csvwriter.writerows(results)
 
     print("Benchmarking completed. Results saved to ", SAVE_DIR)
 
+    plot_results()
+
 if __name__ == "__main__":
     main()