Refactor CMakeLists.txt for eviction sources and add plot_hand_position script for visualizing hand position and retention ratio

Author: 1a1a11a

libCacheSim/cache/CMakeLists.txt

@@ -83,24 +83,24 @@ set(eviction_sources_cpp
 
 # LRB
 if (ENABLE_LRB)
-    set (eviction_sources_cpp ${eviction_sources_cpp} 
+    set (eviction_sources_cpp ${eviction_sources_cpp}
         eviction/LRB/lrb.cpp
         eviction/LRB/LRB_Interface.cpp
         )
 endif()
 
 # 3L Cache
 if (ENABLE_3L_CACHE)
-    set (eviction_sources_cpp ${eviction_sources_cpp} 
+    set (eviction_sources_cpp ${eviction_sources_cpp}
         eviction/3LCache/ThreeLCache_Interface.cpp
         eviction/3LCache/ThreeLCache.cpp
         )
 endif()
 
 # GLCache
 if (ENABLE_GLCACHE)
-    set(eviction_sources_c 
-        ${eviction_sources_c} 
+    set(eviction_sources_c
+        ${eviction_sources_c}
         eviction/GLCache/bucket.c
         eviction/GLCache/dataPrep.c
         eviction/GLCache/eviction.c

scripts/plot_hand_position.py

@@ -0,0 +1,198 @@
+"""
+Plot hand position (scan depth) and retention ratio for CLOCK and SIEVE.
+
+Usage:
+    python3 scripts/plot_hand_position.py tracking_*.csv
+"""
+
+import os
+import sys
+import numpy as np
+import matplotlib.pyplot as plt
+
+sys.path.append(os.path.dirname(os.path.abspath(__file__)))
+from utils.plot_utils import *
+
+
+def plot_tracking(csv_files, output_name="hand_position", trace_name=None):
+    has_hand_pos = False
+
+    # load data
+    data = {}
+    for f in csv_files:
+        # extract algo name from filename: tracking_[trace_]algo_size.csv
+        base = os.path.basename(f)
+        # strip extensions like .csv, .csv.plot
+        for ext in [".csv.plot", ".plot", ".csv"]:
+            if base.endswith(ext):
+                base = base[:-len(ext)]
+                break
+        if base.startswith("tracking_"):
+            base = base[len("tracking_"):]
+        # split off the trailing cache size (last _number segment)
+        parts = base.rsplit("_", 1)
+        label = parts[0] if len(parts) == 2 and parts[1].isdigit() else base
+        with open(f) as fh:
+            header = fh.readline().strip().split(",")
+        cols = np.genfromtxt(f, delimiter=",", skip_header=1)
+        if cols.ndim == 1:
+            continue
+        data[label] = (header, cols)
+        if "hand_pos" in header:
+            has_hand_pos = True
+
+    if not data:
+        print("No valid data found")
+        return
+
+    n_plots = 3 if has_hand_pos else 2
+    fig, axes = plt.subplots(n_plots, 1, figsize=(10, 4 * n_plots), sharex=True)
+
+    colors = ["navy", "darkorange", "tab:green", "cornflowerblue", "tab:red", "tab:purple"]
+    linestyles = ["-", "--", "-.", ":"]
+
+    for i, (label, (header, cols)) in enumerate(data.items()):
+        color = colors[i % len(colors)]
+        ls = linestyles[i % len(linestyles)]
+        vtime = cols[:, 0]
+
+        # scan depth
+        scan_idx = header.index("avg_scan_depth")
+        axes[0].plot(vtime, cols[:, scan_idx], label=label, color=color,
+                     linestyle=ls, linewidth=1.5)
+
+        # retention ratio
+        ret_idx = header.index("retention_ratio")
+        axes[1].plot(vtime, cols[:, ret_idx], label=label, color=color,
+                     linestyle=ls, linewidth=1.5)
+
+        # hand position (SIEVE only)
+        if has_hand_pos and "hand_pos" in header:
+            hp_idx = header.index("hand_pos")
+            axes[2].plot(vtime, cols[:, hp_idx], label=label, color=color,
+                         linestyle=ls, linewidth=1.5)
+
+    axes[0].set_ylabel("Avg Scan Depth\n(objects per eviction)")
+    axes[0].legend(frameon=False)
+    axes[0].grid(axis="y", linestyle="--", alpha=0.5)
+
+    axes[1].set_ylabel("Retention Ratio")
+    axes[1].legend(frameon=False)
+    axes[1].grid(axis="y", linestyle="--", alpha=0.5)
+    axes[1].set_ylim(-0.05, 1.05)
+
+    if has_hand_pos:
+        axes[2].set_ylabel("Hand Position\n(0=tail, 1=head)")
+        axes[2].legend(frameon=False)
+        axes[2].grid(axis="y", linestyle="--", alpha=0.5)
+        axes[2].set_ylim(-0.05, 1.05)
+
+    axes[-1].set_xlabel("Virtual Time (requests)")
+
+    if trace_name:
+        fig.suptitle(trace_name, fontsize=14, y=1.02)
+
+    plt.tight_layout()
+    os.makedirs("figure", exist_ok=True)
+    plt.savefig("figure/{}.pdf".format(output_name), bbox_inches="tight")
+    plt.savefig("figure/{}.png".format(output_name), bbox_inches="tight", dpi=150)
+    plt.close()
+    print("Plots saved to figure/{}.pdf and figure/{}.png".format(output_name, output_name))
+
+
+def plot_binned_retention(csv_files, output_name="retention_by_position", trace_name=None):
+    """Box plot of retention ratio at different hand positions in SIEVE."""
+    fig, ax = plt.subplots(1, 1, figsize=(10, 5))
+    plt.rcParams.update({'font.size': 16})
+
+    colors = ["navy", "darkorange", "tab:green", "cornflowerblue"]
+
+    # load all valid files first
+    all_data = []
+    for f in csv_files:
+        label = os.path.basename(f)
+        for ext in [".csv.plot", ".plot", ".csv"]:
+            if label.endswith(ext):
+                label = label[:-len(ext)]
+                break
+        # strip any single-char temp prefixes like "a_" or "b_"
+        if len(label) > 2 and label[1] == '_' and label[0].isalpha():
+            label = label[2:]
+        if label.startswith("tracking_"):
+            label = label[len("tracking_"):]
+        label = label.replace("_binned", "")
+        # remove duplicate trace prefix (e.g. "w100_tracking_w100_Sieve" -> "Sieve")
+        if "_tracking_" in label:
+            label = label.split("_tracking_")[-1]
+        parts = label.rsplit("_", 1)
+        label = parts[0] if len(parts) == 2 and parts[1].isdigit() else label
+
+        with open(f) as fh:
+            header = fh.readline().strip().split(",")
+        cols = np.genfromtxt(f, delimiter=",", skip_header=1)
+        if cols.ndim == 1 or cols.shape[0] < 2:
+            continue
+        all_data.append((label, header, cols))
+
+    if not all_data:
+        print("No valid data found")
+        return
+
+    n_files = len(all_data)
+    n_bins = len(all_data[0][1]) - 1
+    width = 0.7 / n_files
+    bin_labels = []
+
+    for fi, (label, header, cols) in enumerate(all_data):
+        bin_data = []
+        bin_labels = []
+        for b in range(n_bins):
+            vals = cols[:, b + 1]
+            vals = vals[vals >= 0]
+            bin_data.append(vals)
+            bin_labels.append(header[b + 1].replace("ret_", "") + "%")
+
+        offset = (fi - (n_files - 1) / 2) * width
+        positions = np.arange(n_bins) + offset
+        bp = ax.boxplot(bin_data, positions=positions, widths=width * 0.9,
+                        patch_artist=True, showfliers=False,
+                        medianprops=dict(color="black", linewidth=1.5))
+        color = colors[fi % len(colors)]
+        for patch in bp['boxes']:
+            patch.set_facecolor(color)
+            patch.set_alpha(0.7)
+        # legend entry
+        ax.plot([], [], color=color, linewidth=8, alpha=0.7, label=label)
+
+    ax.set_xticks(range(n_bins))
+    ax.set_xticklabels(bin_labels, rotation=45, ha="right", fontsize=12)
+    ax.legend(frameon=False)
+    ax.set_xlabel("Hand Position (% from tail toward head)")
+    ax.set_ylabel("Retention Ratio")
+    ax.set_ylim(-0.05, 1.05)
+    ax.grid(axis="y", linestyle="--", alpha=0.5)
+
+    if trace_name:
+        ax.set_title(trace_name)
+
+    plt.tight_layout()
+    os.makedirs("figure", exist_ok=True)
+    plt.savefig("figure/{}.pdf".format(output_name), bbox_inches="tight")
+    plt.savefig("figure/{}.png".format(output_name), bbox_inches="tight", dpi=150)
+    plt.close()
+    print("Plots saved to figure/{}.pdf and figure/{}.png".format(output_name, output_name))
+
+
+if __name__ == "__main__":
+    import argparse
+    p = argparse.ArgumentParser(description="Plot hand position and retention ratio")
+    p.add_argument("files", nargs="+", help="tracking CSV files")
+    p.add_argument("--name", type=str, default="hand_position", help="output file name")
+    p.add_argument("--trace", type=str, default=None, help="trace name for plot title")
+    p.add_argument("--binned", action="store_true", help="plot binned retention box plot")
+    args = p.parse_args()
+
+    if args.binned:
+        plot_binned_retention(args.files, args.name, args.trace)
+    else:
+        plot_tracking(args.files, args.name, args.trace)