diff --git a/script/analyze_build/README.md b/script/analyze_build/README.md
index 582a01f91d..4a1bcca50c 100644
--- a/script/analyze_build/README.md
+++ b/script/analyze_build/README.md
@@ -6,10 +6,17 @@ Simple, fast tools for analyzing Clang `-ftime-trace` build performance data.
 
 This directory provides straightforward Python tools for analyzing the JSON trace files generated during compilation with `-ftime-trace`. The focus is on simplicity and speed - no caching, no complexity, just fast parallel I/O and pandas DataFrames.
 
-**Key principle: Fresh analysis every time is faster and simpler than managing caches.**
-
 ## Quick Start
 
+Configure a build directory `build-trace` and edit CMakeCache to add `-ftime-trace` to the `CMAKE_CXX_FLAGS`. With `-ftime-trace` enabled, the clang compiler will generate `.json` trace files alongside each compiled object file. These trace files contain detailed timing information about:
+
+- Template instantiations
+- Function parsing
+- Code generation phases
+- Optimization passes
+
+These JSON files are what the analysis tools in `script/analyze_build/` are designed to process.
+
 ```bash
 # Analyze all trace files in a directory
 cd script/analyze_build/examples
@@ -57,6 +64,7 @@ python examples/analyze_build.py ../../build-trace
 ```
 
 This will:
+
 - Find all `.json` files recursively
 - Process them in parallel using all CPU cores
 - Display comprehensive build statistics
@@ -93,6 +101,7 @@ print(f"Template time: {templates_df['dur'].sum() / 1e6:.2f}s")
 For interactive analysis, see the comprehensive example notebook:
 
 **[notebooks/comprehensive_example.ipynb](notebooks/comprehensive_example.ipynb)** - Complete guide covering:
+
 - Single file analysis with detailed explanations
 - Multi-file parallel processing
 - Build-wide statistics and template analysis
@@ -130,11 +139,13 @@ print(event_totals.head(10))
 ## Performance
 
 **Typical performance on 4,484 trace files (~46 GB):**
+
 - Parsing: ~26 seconds (174 files/sec)
 - Memory: ~1-2 GB
 - Throughput: I/O limited (uses all CPU cores)
 
 **Why no caching?**
+
 - Fresh analysis is faster than cache management overhead
 - Simpler code (60% less code than cached version)
 - No cache invalidation issues
@@ -164,6 +175,7 @@ The trace files use the [Chrome Trace Event Format](https://docs.google.com/docu
 ```
 
 **Key fields:**
+
 - `name`: Event type (e.g., "InstantiateClass", "ParseFunctionDefinition")
 - `dur`: Duration in microseconds
 - `ts`: Timestamp in microseconds
@@ -248,7 +260,7 @@ template_time = templates_df['dur'].sum()
 print(f"Template time: {(template_time / total_time) * 100:.1f}%")
 ```
 
-## Tips
+## Build time analysis philosophy
 
 - **Use all CPU cores**: The tools automatically use all available cores for parallel processing
 - **Memory is cheap**: 1-2GB for 4,484 files is acceptable on modern systems
diff --git a/script/analyze_build/notebooks/comprehensive_example.ipynb b/script/analyze_build/notebooks/comprehensive_example.ipynb
index a48ac0fbbb..491e373f07 100644
--- a/script/analyze_build/notebooks/comprehensive_example.ipynb
+++ b/script/analyze_build/notebooks/comprehensive_example.ipynb
@@ -40,9 +40,19 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 1,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Note: Install plotly for visualizations: pip install plotly\n",
+      "Using 384 CPU cores for parallel processing\n",
+      "Pandas version: 2.3.3\n"
+     ]
+    }
+   ],
    "source": [
     "from importlib.util import find_spec\n",
     "from multiprocessing import cpu_count\n",
@@ -88,9 +98,20 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 2,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Found 1,253 trace files\n",
+      "\n",
+      "Using sample file: device_conv1d_bwd_data_xdl_nwc_kxc_nwk_int8_instance.cpp.json\n",
+      "File size: 11652.5 KB\n"
+     ]
+    }
+   ],
    "source": [
     "# Configure the path to your trace files\n",
     "TRACE_DIR = Path(\"../../../build-trace\")\n",
@@ -127,9 +148,24 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 3,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Parsed 15,110 events in 0.043s\n",
+      "Transformed to Pandas tables in 1.115s\n",
+      "\n",
+      "Pandas DataFrames:\n",
+      "  templates           :  8,703 rows,   8.10 MB | template_id, template_name, full_signature, depth, arg_count\n",
+      "  instantiations      :  9,838 rows,   0.31 MB | instantiation_id, template_id, file_id, dur_us, ts_us, event_type\n",
+      "  template_args       : 51,474 rows,   8.73 MB | parent_template_id, arg_position, arg_template_id, arg_type, arg_text\n",
+      "  events              : 15,110 rows,   0.53 MB | name, dur, ts, pid, tid, ph, ts_absolute_us\n"
+     ]
+    }
+   ],
    "source": [
     "if sample_files:\n",
     "    # Parse the trace file\n",
@@ -172,9 +208,22 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 4,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Compilation Unit Summary:\n",
+      "  Trace file: device_conv1d_bwd_data_xdl_nwc_kxc_nwk_int8_instance.cpp.json\n",
+      "  Trace file size: 11652.5 KB\n",
+      "  Start time: 2026-01-03 22:51:51.980489\n",
+      "  Total compilation time: 178.01s\n",
+      "  Total events: 15,110\n"
+     ]
+    }
+   ],
    "source": [
     "if sample_files:\n",
     "    print(\"Compilation Unit Summary:\")\n",
@@ -196,9 +245,187 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 5,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Templates DataFrame Schema:\n",
+      "Column                    Type             Memory (MB)   % of Total\n",
+      "-------------------------------------------------------------------\n",
+      "template_id               int32                   0.03         0.4%\n",
+      "template_name             category                0.03         0.4%\n",
+      "full_signature            object                  8.02        99.0%\n",
+      "depth                     int8                    0.01         0.1%\n",
+      "arg_count                 int8                    0.01         0.1%\n",
+      "Index                     RangeIndex              0.00         0.0%\n",
+      "-------------------------------------------------------------------\n",
+      "TOTAL                                             8.10       100.0%\n",
+      "\n",
+      "Total templates: 8,703\n",
+      "CK templates: 6,331 (72.7%)\n",
+      "Other templates: 2,372\n",
+      "\n",
+      "Sample CK templates:\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>template_id</th>\n",
+       "      <th>template_name</th>\n",
+       "      <th>full_signature</th>\n",
+       "      <th>depth</th>\n",
+       "      <th>arg_count</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>45</th>\n",
+       "      <td>45</td>\n",
+       "      <td>ck::Tuple</td>\n",
+       "      <td>ck::Tuple&lt;ck::integral_constant&lt;int, 2&gt;, _BitInt(4)&gt;</td>\n",
+       "      <td>2</td>\n",
+       "      <td>2</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>46</th>\n",
+       "      <td>46</td>\n",
+       "      <td>ck::Tuple</td>\n",
+       "      <td>ck::Tuple&lt;ck::integral_constant&lt;int, 16&gt;, _BitInt(6)&gt;</td>\n",
+       "      <td>2</td>\n",
+       "      <td>2</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>47</th>\n",
+       "      <td>47</td>\n",
+       "      <td>ck::Tuple</td>\n",
+       "      <td>ck::Tuple&lt;float, float&gt;</td>\n",
+       "      <td>1</td>\n",
+       "      <td>2</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>48</th>\n",
+       "      <td>48</td>\n",
+       "      <td>ck::vector_type</td>\n",
+       "      <td>ck::vector_type&lt;float, 2&gt;::(unnamed union at /home/AMD/jshumway/composable_k...</td>\n",
+       "      <td>1</td>\n",
+       "      <td>2</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>49</th>\n",
+       "      <td>49</td>\n",
+       "      <td>ck::vector_type</td>\n",
+       "      <td>ck::vector_type&lt;float, 2&gt;</td>\n",
+       "      <td>1</td>\n",
+       "      <td>2</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>50</th>\n",
+       "      <td>50</td>\n",
+       "      <td>ck::Tuple</td>\n",
+       "      <td>ck::Tuple&lt;float, float, float, float&gt;</td>\n",
+       "      <td>1</td>\n",
+       "      <td>4</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>51</th>\n",
+       "      <td>51</td>\n",
+       "      <td>ck::vector_type</td>\n",
+       "      <td>ck::vector_type&lt;float, 4&gt;::(unnamed union at /home/AMD/jshumway/composable_k...</td>\n",
+       "      <td>1</td>\n",
+       "      <td>2</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>52</th>\n",
+       "      <td>52</td>\n",
+       "      <td>ck::vector_type</td>\n",
+       "      <td>ck::vector_type&lt;float, 4&gt;</td>\n",
+       "      <td>1</td>\n",
+       "      <td>2</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>53</th>\n",
+       "      <td>53</td>\n",
+       "      <td>ck::Tuple</td>\n",
+       "      <td>ck::Tuple&lt;float, float, float, float, float, float, float, float&gt;</td>\n",
+       "      <td>1</td>\n",
+       "      <td>8</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>54</th>\n",
+       "      <td>54</td>\n",
+       "      <td>ck::detail::TupleElementKeyData</td>\n",
+       "      <td>ck::detail::TupleElementKeyData&lt;ck::detail::TupleElementKey&lt;3&gt;, float __attr...</td>\n",
+       "      <td>2</td>\n",
+       "      <td>2</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "    template_id                    template_name  \\\n",
+       "45           45                        ck::Tuple   \n",
+       "46           46                        ck::Tuple   \n",
+       "47           47                        ck::Tuple   \n",
+       "48           48                  ck::vector_type   \n",
+       "49           49                  ck::vector_type   \n",
+       "50           50                        ck::Tuple   \n",
+       "51           51                  ck::vector_type   \n",
+       "52           52                  ck::vector_type   \n",
+       "53           53                        ck::Tuple   \n",
+       "54           54  ck::detail::TupleElementKeyData   \n",
+       "\n",
+       "                                                                     full_signature  \\\n",
+       "45                             ck::Tuple<ck::integral_constant<int, 2>, _BitInt(4)>   \n",
+       "46                            ck::Tuple<ck::integral_constant<int, 16>, _BitInt(6)>   \n",
+       "47                                                          ck::Tuple<float, float>   \n",
+       "48  ck::vector_type<float, 2>::(unnamed union at /home/AMD/jshumway/composable_k...   \n",
+       "49                                                        ck::vector_type<float, 2>   \n",
+       "50                                            ck::Tuple<float, float, float, float>   \n",
+       "51  ck::vector_type<float, 4>::(unnamed union at /home/AMD/jshumway/composable_k...   \n",
+       "52                                                        ck::vector_type<float, 4>   \n",
+       "53                ck::Tuple<float, float, float, float, float, float, float, float>   \n",
+       "54  ck::detail::TupleElementKeyData<ck::detail::TupleElementKey<3>, float __attr...   \n",
+       "\n",
+       "    depth  arg_count  \n",
+       "45      2          2  \n",
+       "46      2          2  \n",
+       "47      1          2  \n",
+       "48      1          2  \n",
+       "49      1          2  \n",
+       "50      1          4  \n",
+       "51      1          2  \n",
+       "52      1          2  \n",
+       "53      1          8  \n",
+       "54      2          2  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "if sample_files:\n",
     "    templates_df = tables[\"templates\"]\n",
@@ -250,9 +477,166 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 6,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Template Instantiation Summary:\n",
+      "  Unique templates: 8,703\n",
+      "  Total instantiations: 9,838\n",
+      "  Template time: 39.65s\n",
+      "  Percentage of build: 22.3%\n",
+      "  Avg per instantiation: 4.03 ms\n",
+      "\n",
+      "Top 10 Templates by Total Time:\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>signature</th>\n",
+       "      <th>count</th>\n",
+       "      <th>total_ms</th>\n",
+       "      <th>mean_ms</th>\n",
+       "      <th>depth</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>670</th>\n",
+       "      <td>ck::generate_tuple&lt;(lambda at /home/AMD/jshumway/composable_kernel/include/c...</td>\n",
+       "      <td>12</td>\n",
+       "      <td>444.005</td>\n",
+       "      <td>37.000417</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>669</th>\n",
+       "      <td>ck::generate_tuple_for&lt;(lambda at /home/AMD/jshumway/composable_kernel/inclu...</td>\n",
+       "      <td>12</td>\n",
+       "      <td>442.665</td>\n",
+       "      <td>36.888750</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1411</th>\n",
+       "      <td>ck::detail::applier&lt;int, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10&gt;::operator()&lt;(lamb...</td>\n",
+       "      <td>34</td>\n",
+       "      <td>365.628</td>\n",
+       "      <td>10.753765</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6424</th>\n",
+       "      <td>ck::detail::applier&lt;int, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12&gt;::operator...</td>\n",
+       "      <td>2</td>\n",
+       "      <td>346.742</td>\n",
+       "      <td>173.371000</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4079</th>\n",
+       "      <td>ck::generate_tuple&lt;(lambda at /home/AMD/jshumway/composable_kernel/include/c...</td>\n",
+       "      <td>4</td>\n",
+       "      <td>330.928</td>\n",
+       "      <td>82.732000</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4078</th>\n",
+       "      <td>ck::generate_tuple_for&lt;(lambda at /home/AMD/jshumway/composable_kernel/inclu...</td>\n",
+       "      <td>4</td>\n",
+       "      <td>330.380</td>\n",
+       "      <td>82.595000</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4859</th>\n",
+       "      <td>ck::transform_tensor_descriptor&lt;ck::TensorDescriptor&lt;ck::Tuple&lt;ck::UnMerge&lt;c...</td>\n",
+       "      <td>1</td>\n",
+       "      <td>327.756</td>\n",
+       "      <td>327.756000</td>\n",
+       "      <td>6</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1470</th>\n",
+       "      <td>ck::generate_tuple&lt;(lambda at /home/AMD/jshumway/composable_kernel/include/c...</td>\n",
+       "      <td>6</td>\n",
+       "      <td>311.490</td>\n",
+       "      <td>51.915000</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1469</th>\n",
+       "      <td>ck::generate_tuple_for&lt;(lambda at /home/AMD/jshumway/composable_kernel/inclu...</td>\n",
+       "      <td>6</td>\n",
+       "      <td>310.729</td>\n",
+       "      <td>51.788167</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2362</th>\n",
+       "      <td>ck::generate_tuple&lt;(lambda at /home/AMD/jshumway/composable_kernel/include/c...</td>\n",
+       "      <td>5</td>\n",
+       "      <td>295.890</td>\n",
+       "      <td>59.178000</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                                            signature  \\\n",
+       "670   ck::generate_tuple<(lambda at /home/AMD/jshumway/composable_kernel/include/c...   \n",
+       "669   ck::generate_tuple_for<(lambda at /home/AMD/jshumway/composable_kernel/inclu...   \n",
+       "1411  ck::detail::applier<int, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10>::operator()<(lamb...   \n",
+       "6424  ck::detail::applier<int, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12>::operator...   \n",
+       "4079  ck::generate_tuple<(lambda at /home/AMD/jshumway/composable_kernel/include/c...   \n",
+       "4078  ck::generate_tuple_for<(lambda at /home/AMD/jshumway/composable_kernel/inclu...   \n",
+       "4859  ck::transform_tensor_descriptor<ck::TensorDescriptor<ck::Tuple<ck::UnMerge<c...   \n",
+       "1470  ck::generate_tuple<(lambda at /home/AMD/jshumway/composable_kernel/include/c...   \n",
+       "1469  ck::generate_tuple_for<(lambda at /home/AMD/jshumway/composable_kernel/inclu...   \n",
+       "2362  ck::generate_tuple<(lambda at /home/AMD/jshumway/composable_kernel/include/c...   \n",
+       "\n",
+       "      count  total_ms     mean_ms  depth  \n",
+       "670      12   444.005   37.000417      1  \n",
+       "669      12   442.665   36.888750      1  \n",
+       "1411     34   365.628   10.753765      1  \n",
+       "6424      2   346.742  173.371000      1  \n",
+       "4079      4   330.928   82.732000      1  \n",
+       "4078      4   330.380   82.595000      1  \n",
+       "4859      1   327.756  327.756000      6  \n",
+       "1470      6   311.490   51.915000      1  \n",
+       "1469      6   310.729   51.788167      1  \n",
+       "2362      5   295.890   59.178000      1  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "if sample_files and len(tables[\"instantiations\"]) > 0:\n",
     "    # Join instantiations with templates\n",
@@ -316,9 +700,112 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 7,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Template Instantiation by Nesting Depth:\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>depth</th>\n",
+       "      <th>count</th>\n",
+       "      <th>total_ms</th>\n",
+       "      <th>mean_ms</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>1628</td>\n",
+       "      <td>8510.435</td>\n",
+       "      <td>5.227540</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>2</td>\n",
+       "      <td>1962</td>\n",
+       "      <td>5433.074</td>\n",
+       "      <td>2.769151</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>3</td>\n",
+       "      <td>1311</td>\n",
+       "      <td>5042.420</td>\n",
+       "      <td>3.846240</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>4</td>\n",
+       "      <td>2548</td>\n",
+       "      <td>7730.989</td>\n",
+       "      <td>3.034140</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>5</td>\n",
+       "      <td>2112</td>\n",
+       "      <td>10514.270</td>\n",
+       "      <td>4.978348</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>6</td>\n",
+       "      <td>257</td>\n",
+       "      <td>2398.921</td>\n",
+       "      <td>9.334323</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>7</td>\n",
+       "      <td>20</td>\n",
+       "      <td>21.944</td>\n",
+       "      <td>1.097200</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "   depth  count   total_ms   mean_ms\n",
+       "0      1   1628   8510.435  5.227540\n",
+       "1      2   1962   5433.074  2.769151\n",
+       "2      3   1311   5042.420  3.846240\n",
+       "3      4   2548   7730.989  3.034140\n",
+       "4      5   2112  10514.270  4.978348\n",
+       "5      6    257   2398.921  9.334323\n",
+       "6      7     20     21.944  1.097200"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "if sample_files and len(tables[\"templates\"]) > 0:\n",
     "    depth_stats = (\n",
@@ -347,9 +834,17 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 8,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Parallel processing function defined\n"
+     ]
+    }
+   ],
    "source": [
     "from concurrent.futures import ProcessPoolExecutor, as_completed\n",
     "\n",
@@ -380,9 +875,18 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 9,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Found 1,253 trace files\n",
+      "Total size: 18.79 GB\n"
+     ]
+    }
+   ],
    "source": [
     "# Find all trace files\n",
     "json_files = find_trace_files(TRACE_DIR)\n",
@@ -408,9 +912,47 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 10,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Processing 1,253 files with 384 workers...\n",
+      "\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/home/AMD/jshumway/composable_kernel/.venv/lib/python3.12/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
+      "  from .autonotebook import tqdm as notebook_tqdm\n",
+      "Processing: 100%|██████████| 1253/1253 [00:46<00:00, 27.13files/s] \n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "Parsing complete in 48.54s (25.8 files/sec)\n",
+      "\n",
+      "Combining results...\n",
+      "Combined in 4.73s\n",
+      "\n",
+      "Total analysis time: 53.27s\n",
+      "\n",
+      "Combined Tables:\n",
+      "  Templates: 11,751,350 rows\n",
+      "  Instantiations: 17,437,000 rows\n",
+      "  Template Args: 155,281,743 rows\n",
+      "  Events: 24,229,252 rows\n",
+      "  Total memory: 34.06 GB\n"
+     ]
+    }
+   ],
    "source": [
     "if json_files:\n",
     "    print(f\"Processing {len(json_files):,} files with {cpu_count()} workers...\\n\")\n",
@@ -515,9 +1057,26 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 11,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "================================================================================\n",
+      "BUILD-WIDE STATISTICS\n",
+      "================================================================================\n",
+      "Files processed: 1,253\n",
+      "Total events: 24,229,252\n",
+      "Total build time: 5915.00 minutes\n",
+      "Unique templates: 11,751,350\n",
+      "Template instantiations: 17,437,000\n",
+      "Template time: 2495.47 minutes (42.2%)\n",
+      "================================================================================\n"
+     ]
+    }
+   ],
    "source": [
     "if json_files and len(events_df) > 0:\n",
     "    total_build_time_us = events_df[\"dur\"].sum()\n",
@@ -550,9 +1109,326 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 12,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Aggregating template statistics...\n",
+      "Completed in 3.42s\n",
+      "\n",
+      "Top 20 Templates by Total Time:\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>full_signature</th>\n",
+       "      <th>count</th>\n",
+       "      <th>total_s</th>\n",
+       "      <th>mean_ms</th>\n",
+       "      <th>median_ms</th>\n",
+       "      <th>depth</th>\n",
+       "      <th>pct_template_time</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>11751348</th>\n",
+       "      <td>ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...</td>\n",
+       "      <td>2</td>\n",
+       "      <td>135.700625</td>\n",
+       "      <td>67850.312500</td>\n",
+       "      <td>67850.3125</td>\n",
+       "      <td>2</td>\n",
+       "      <td>0.090631</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135552</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;8, 8&gt;, ck::tensor_...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>6</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135553</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;4, 16&gt;, ck::tensor...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>7</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135554</th>\n",
+       "      <td>std::_TupleConstraints&lt;true, ck::tensor_operation::device::DeviceConvNdBwdDa...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>4</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135555</th>\n",
+       "      <td>ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>3</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135556</th>\n",
+       "      <td>ck::detail::static_ford_impl&lt;ck::Sequence&lt;16&gt;, ck::Sequence&lt;0, 1&gt;&gt;::operator...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>2</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135557</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;1, 16&gt;, ck::tensor...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>6</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135558</th>\n",
+       "      <td>std::get&lt;1UL, ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>4</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135559</th>\n",
+       "      <td>std::_Tuple_impl&lt;0, ck::tensor_operation::device::DeviceConv2dBwdDataXdl_Inp...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>4</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135560</th>\n",
+       "      <td>std::__uniq_ptr_impl&lt;ck::tensor_operation::device::DeviceConv2dBwdDataXdl_In...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>4</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135561</th>\n",
+       "      <td>ck::StaticTensorTupleOfVectorBuffer&lt;ck::AddressSpaceEnum::Vgpr, unsigned sho...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>6</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135562</th>\n",
+       "      <td>ck::StaticTensorTupleOfVectorBuffer&lt;ck::AddressSpaceEnum::Vgpr, unsigned sho...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>6</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135563</th>\n",
+       "      <td>std::tuple&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>4</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135564</th>\n",
+       "      <td>ck::TensorDescriptor&lt;ck::Tuple&lt;ck::Embed&lt;ck::Tuple&lt;int, int, int, int, int&gt;,...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>5</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135565</th>\n",
+       "      <td>std::tuple&lt;ck::tensor_operation::device::DeviceConv2dBwdDataXdl_Input_N_Hi_W...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>4</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135566</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;16, 4&gt;, ck::tensor...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>7</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135567</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;8, 4&gt;, ck::tensor_...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>7</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135568</th>\n",
+       "      <td>ck::StaticBufferTupleOfVector&lt;ck::AddressSpaceEnum::Vgpr, _Float16, 64, 1, t...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>1</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135569</th>\n",
+       "      <td>ck::TensorDescriptor&lt;ck::Tuple&lt;ck::Embed&lt;ck::Tuple&lt;ck::integral_constant&lt;int...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>5</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135570</th>\n",
+       "      <td>std::unique_ptr&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>3</td>\n",
+       "      <td>0.088275</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                                           full_signature  \\\n",
+       "11751348  ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...   \n",
+       "9135552   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<8, 8>, ck::tensor_...   \n",
+       "9135553   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<4, 16>, ck::tensor...   \n",
+       "9135554   std::_TupleConstraints<true, ck::tensor_operation::device::DeviceConvNdBwdDa...   \n",
+       "9135555   ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...   \n",
+       "9135556   ck::detail::static_ford_impl<ck::Sequence<16>, ck::Sequence<0, 1>>::operator...   \n",
+       "9135557   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<1, 16>, ck::tensor...   \n",
+       "9135558   std::get<1UL, ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_...   \n",
+       "9135559   std::_Tuple_impl<0, ck::tensor_operation::device::DeviceConv2dBwdDataXdl_Inp...   \n",
+       "9135560   std::__uniq_ptr_impl<ck::tensor_operation::device::DeviceConv2dBwdDataXdl_In...   \n",
+       "9135561   ck::StaticTensorTupleOfVectorBuffer<ck::AddressSpaceEnum::Vgpr, unsigned sho...   \n",
+       "9135562   ck::StaticTensorTupleOfVectorBuffer<ck::AddressSpaceEnum::Vgpr, unsigned sho...   \n",
+       "9135563   std::tuple<ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl...   \n",
+       "9135564   ck::TensorDescriptor<ck::Tuple<ck::Embed<ck::Tuple<int, int, int, int, int>,...   \n",
+       "9135565   std::tuple<ck::tensor_operation::device::DeviceConv2dBwdDataXdl_Input_N_Hi_W...   \n",
+       "9135566   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<16, 4>, ck::tensor...   \n",
+       "9135567   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<8, 4>, ck::tensor_...   \n",
+       "9135568   ck::StaticBufferTupleOfVector<ck::AddressSpaceEnum::Vgpr, _Float16, 64, 1, t...   \n",
+       "9135569   ck::TensorDescriptor<ck::Tuple<ck::Embed<ck::Tuple<ck::integral_constant<int...   \n",
+       "9135570   std::unique_ptr<ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...   \n",
+       "\n",
+       "          count     total_s       mean_ms   median_ms  depth  \\\n",
+       "11751348      2  135.700625  67850.312500  67850.3125      2   \n",
+       "9135552     846  132.173182    156.233076      1.0550      6   \n",
+       "9135553     846  132.173182    156.233076      1.0550      7   \n",
+       "9135554     846  132.173182    156.233076      1.0550      4   \n",
+       "9135555     846  132.173182    156.233076      1.0550      3   \n",
+       "9135556     846  132.173182    156.233076      1.0550      2   \n",
+       "9135557     846  132.173182    156.233076      1.0550      6   \n",
+       "9135558     846  132.173182    156.233076      1.0550      4   \n",
+       "9135559     846  132.173182    156.233076      1.0550      4   \n",
+       "9135560     846  132.173182    156.233076      1.0550      4   \n",
+       "9135561     846  132.173182    156.233076      1.0550      6   \n",
+       "9135562     846  132.173182    156.233076      1.0550      6   \n",
+       "9135563     846  132.173182    156.233076      1.0550      4   \n",
+       "9135564     846  132.173182    156.233076      1.0550      5   \n",
+       "9135565     846  132.173182    156.233076      1.0550      4   \n",
+       "9135566     846  132.173182    156.233076      1.0550      7   \n",
+       "9135567     846  132.173182    156.233076      1.0550      7   \n",
+       "9135568     846  132.173182    156.233076      1.0550      1   \n",
+       "9135569     846  132.173182    156.233076      1.0550      5   \n",
+       "9135570     846  132.173182    156.233076      1.0550      3   \n",
+       "\n",
+       "          pct_template_time  \n",
+       "11751348           0.090631  \n",
+       "9135552            0.088275  \n",
+       "9135553            0.088275  \n",
+       "9135554            0.088275  \n",
+       "9135555            0.088275  \n",
+       "9135556            0.088275  \n",
+       "9135557            0.088275  \n",
+       "9135558            0.088275  \n",
+       "9135559            0.088275  \n",
+       "9135560            0.088275  \n",
+       "9135561            0.088275  \n",
+       "9135562            0.088275  \n",
+       "9135563            0.088275  \n",
+       "9135564            0.088275  \n",
+       "9135565            0.088275  \n",
+       "9135566            0.088275  \n",
+       "9135567            0.088275  \n",
+       "9135568            0.088275  \n",
+       "9135569            0.088275  \n",
+       "9135570            0.088275  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "if json_files and len(instantiations_df) > 0:\n",
     "    # OPTIMIZATION: Aggregate FIRST, then join (much faster!)\n",
@@ -621,9 +1497,305 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 13,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Filtered to 9,433,574 CK templates (from 11,751,350 total)\n",
+      "CK template time: 66204864.83s\n",
+      "Percentage of total template time: 44216.7%\n",
+      "\n",
+      "Top 20 CK Templates by Total Time:\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>full_signature</th>\n",
+       "      <th>count</th>\n",
+       "      <th>total_s</th>\n",
+       "      <th>mean_ms</th>\n",
+       "      <th>median_ms</th>\n",
+       "      <th>depth</th>\n",
+       "      <th>pct_ck_time</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>11751348</th>\n",
+       "      <td>ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...</td>\n",
+       "      <td>2</td>\n",
+       "      <td>135.700625</td>\n",
+       "      <td>67850.312500</td>\n",
+       "      <td>67850.3125</td>\n",
+       "      <td>2</td>\n",
+       "      <td>0.000205</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135552</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;8, 8&gt;, ck::tensor_...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>6</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135553</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;4, 16&gt;, ck::tensor...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>7</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135555</th>\n",
+       "      <td>ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>3</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135556</th>\n",
+       "      <td>ck::detail::static_ford_impl&lt;ck::Sequence&lt;16&gt;, ck::Sequence&lt;0, 1&gt;&gt;::operator...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>2</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135557</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;1, 16&gt;, ck::tensor...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>6</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135561</th>\n",
+       "      <td>ck::StaticTensorTupleOfVectorBuffer&lt;ck::AddressSpaceEnum::Vgpr, unsigned sho...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>6</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135562</th>\n",
+       "      <td>ck::StaticTensorTupleOfVectorBuffer&lt;ck::AddressSpaceEnum::Vgpr, unsigned sho...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>6</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135564</th>\n",
+       "      <td>ck::TensorDescriptor&lt;ck::Tuple&lt;ck::Embed&lt;ck::Tuple&lt;int, int, int, int, int&gt;,...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>5</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135566</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;16, 4&gt;, ck::tensor...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>7</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135567</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;8, 4&gt;, ck::tensor_...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>7</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135568</th>\n",
+       "      <td>ck::StaticBufferTupleOfVector&lt;ck::AddressSpaceEnum::Vgpr, _Float16, 64, 1, t...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>1</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135569</th>\n",
+       "      <td>ck::TensorDescriptor&lt;ck::Tuple&lt;ck::Embed&lt;ck::Tuple&lt;ck::integral_constant&lt;int...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>5</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135571</th>\n",
+       "      <td>ck::tensor_operation::device::DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>2</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135577</th>\n",
+       "      <td>ck::make_tensor_coordinate(const ck::TensorDescriptor&lt;ck::Tuple&lt;ck::UnMerge&lt;...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>5</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135579</th>\n",
+       "      <td>ck::Tuple&lt;ck::vector_type&lt;float, 1&gt;, ck::vector_type&lt;float, 1&gt;, ck::vector_t...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>2</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135582</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;16, 16&gt;, ck::tenso...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>6</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135583</th>\n",
+       "      <td>ck::Sequence&lt;5, 6, 7, 8, 9, 10&gt;::ReorderGivenNew2Old&lt;0, 1, 2, 3, 4, 5&gt;</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>1</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135584</th>\n",
+       "      <td>ck::to_multi_index&lt;ck::Sequence&lt;4, 12&gt;&gt;</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>2</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9135585</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;16, 16&gt;, ck::tenso...</td>\n",
+       "      <td>846</td>\n",
+       "      <td>132.173182</td>\n",
+       "      <td>156.233076</td>\n",
+       "      <td>1.0550</td>\n",
+       "      <td>6</td>\n",
+       "      <td>0.000200</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                                           full_signature  \\\n",
+       "11751348  ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...   \n",
+       "9135552   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<8, 8>, ck::tensor_...   \n",
+       "9135553   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<4, 16>, ck::tensor...   \n",
+       "9135555   ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...   \n",
+       "9135556   ck::detail::static_ford_impl<ck::Sequence<16>, ck::Sequence<0, 1>>::operator...   \n",
+       "9135557   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<1, 16>, ck::tensor...   \n",
+       "9135561   ck::StaticTensorTupleOfVectorBuffer<ck::AddressSpaceEnum::Vgpr, unsigned sho...   \n",
+       "9135562   ck::StaticTensorTupleOfVectorBuffer<ck::AddressSpaceEnum::Vgpr, unsigned sho...   \n",
+       "9135564   ck::TensorDescriptor<ck::Tuple<ck::Embed<ck::Tuple<int, int, int, int, int>,...   \n",
+       "9135566   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<16, 4>, ck::tensor...   \n",
+       "9135567   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<8, 4>, ck::tensor_...   \n",
+       "9135568   ck::StaticBufferTupleOfVector<ck::AddressSpaceEnum::Vgpr, _Float16, 64, 1, t...   \n",
+       "9135569   ck::TensorDescriptor<ck::Tuple<ck::Embed<ck::Tuple<ck::integral_constant<int...   \n",
+       "9135571   ck::tensor_operation::device::DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle...   \n",
+       "9135577   ck::make_tensor_coordinate(const ck::TensorDescriptor<ck::Tuple<ck::UnMerge<...   \n",
+       "9135579   ck::Tuple<ck::vector_type<float, 1>, ck::vector_type<float, 1>, ck::vector_t...   \n",
+       "9135582   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<16, 16>, ck::tenso...   \n",
+       "9135583            ck::Sequence<5, 6, 7, 8, 9, 10>::ReorderGivenNew2Old<0, 1, 2, 3, 4, 5>   \n",
+       "9135584                                           ck::to_multi_index<ck::Sequence<4, 12>>   \n",
+       "9135585   ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<16, 16>, ck::tenso...   \n",
+       "\n",
+       "          count     total_s       mean_ms   median_ms  depth  pct_ck_time  \n",
+       "11751348      2  135.700625  67850.312500  67850.3125      2     0.000205  \n",
+       "9135552     846  132.173182    156.233076      1.0550      6     0.000200  \n",
+       "9135553     846  132.173182    156.233076      1.0550      7     0.000200  \n",
+       "9135555     846  132.173182    156.233076      1.0550      3     0.000200  \n",
+       "9135556     846  132.173182    156.233076      1.0550      2     0.000200  \n",
+       "9135557     846  132.173182    156.233076      1.0550      6     0.000200  \n",
+       "9135561     846  132.173182    156.233076      1.0550      6     0.000200  \n",
+       "9135562     846  132.173182    156.233076      1.0550      6     0.000200  \n",
+       "9135564     846  132.173182    156.233076      1.0550      5     0.000200  \n",
+       "9135566     846  132.173182    156.233076      1.0550      7     0.000200  \n",
+       "9135567     846  132.173182    156.233076      1.0550      7     0.000200  \n",
+       "9135568     846  132.173182    156.233076      1.0550      1     0.000200  \n",
+       "9135569     846  132.173182    156.233076      1.0550      5     0.000200  \n",
+       "9135571     846  132.173182    156.233076      1.0550      2     0.000200  \n",
+       "9135577     846  132.173182    156.233076      1.0550      5     0.000200  \n",
+       "9135579     846  132.173182    156.233076      1.0550      2     0.000200  \n",
+       "9135582     846  132.173182    156.233076      1.0550      6     0.000200  \n",
+       "9135583     846  132.173182    156.233076      1.0550      1     0.000200  \n",
+       "9135584     846  132.173182    156.233076      1.0550      2     0.000200  \n",
+       "9135585     846  132.173182    156.233076      1.0550      6     0.000200  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "if json_files and len(template_stats) > 0:\n",
     "    # Filter to only CK namespaces\n",
@@ -671,9 +1843,259 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 14,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Top 20 Most Frequently Instantiated Templates:\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>full_signature</th>\n",
+       "      <th>count</th>\n",
+       "      <th>total_s</th>\n",
+       "      <th>mean_ms</th>\n",
+       "      <th>depth</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>7243825</th>\n",
+       "      <td>ck::Tuple&lt;ck::PassThrough&lt;int&gt;, ck::UnMerge&lt;ck::Tuple&lt;int, int&gt;, false&gt;, ck:...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>3</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243826</th>\n",
+       "      <td>std::__is_implicitly_default_constructible&lt;ck::tensor_operation::device::Dev...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>3</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243827</th>\n",
+       "      <td>ck::StaticBufferTupleOfVector&lt;ck::AddressSpaceEnum::Vgpr, _Float16, 4, 1, tr...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243828</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;2, 2&gt;, ck::tensor_...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>6</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243829</th>\n",
+       "      <td>ck::utility::launch_and_time_kernel_with_preprocess&lt;false, ck::GridwiseGemmM...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>5</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243830</th>\n",
+       "      <td>std::unique_ptr&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>4</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243831</th>\n",
+       "      <td>std::tuple&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>4</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243832</th>\n",
+       "      <td>ck::TensorDescriptor&lt;ck::Tuple&lt;ck::Embed&lt;ck::Tuple&lt;ck::integral_constant&lt;int...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>5</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243833</th>\n",
+       "      <td>ck::generate_tuple&lt;(lambda at /home/AMD/jshumway/composable_kernel/include/c...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243834</th>\n",
+       "      <td>std::__uniq_ptr_data&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMulti...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>4</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243835</th>\n",
+       "      <td>std::allocator_traits&lt;std::allocator&lt;std::_Rb_tree_node&lt;std::pair&lt;const ck::...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>5</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243836</th>\n",
+       "      <td>std::unique_ptr&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>3</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243837</th>\n",
+       "      <td>std::__and_&lt;std::__is_implicitly_default_constructible&lt;ck::tensor_operation:...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>5</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243838</th>\n",
+       "      <td>std::tuple&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>4</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243839</th>\n",
+       "      <td>ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>5</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243840</th>\n",
+       "      <td>std::unique_ptr&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>4</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243841</th>\n",
+       "      <td>std::pair&lt;const ck::BlockGemmPipelineScheduler, std::basic_string&lt;char&gt;&gt;::pa...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>2</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243842</th>\n",
+       "      <td>ck::make_tuple&lt;ck::integral_constant&lt;int, 1&gt;, ck::integral_constant&lt;int, 4&gt;,...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>2</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243843</th>\n",
+       "      <td>ck::StaticBufferTupleOfVector&lt;ck::AddressSpaceEnum::Vgpr, unsigned short, 64...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243844</th>\n",
+       "      <td>std::is_nothrow_constructible&lt;std::unique_ptr&lt;ck::tensor_operation::device::...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>4</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                                          full_signature  \\\n",
+       "7243825  ck::Tuple<ck::PassThrough<int>, ck::UnMerge<ck::Tuple<int, int>, false>, ck:...   \n",
+       "7243826  std::__is_implicitly_default_constructible<ck::tensor_operation::device::Dev...   \n",
+       "7243827  ck::StaticBufferTupleOfVector<ck::AddressSpaceEnum::Vgpr, _Float16, 4, 1, tr...   \n",
+       "7243828  ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<2, 2>, ck::tensor_...   \n",
+       "7243829  ck::utility::launch_and_time_kernel_with_preprocess<false, ck::GridwiseGemmM...   \n",
+       "7243830  std::unique_ptr<ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...   \n",
+       "7243831  std::tuple<ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl...   \n",
+       "7243832  ck::TensorDescriptor<ck::Tuple<ck::Embed<ck::Tuple<ck::integral_constant<int...   \n",
+       "7243833  ck::generate_tuple<(lambda at /home/AMD/jshumway/composable_kernel/include/c...   \n",
+       "7243834  std::__uniq_ptr_data<ck::tensor_operation::device::DeviceGroupedConvFwdMulti...   \n",
+       "7243835  std::allocator_traits<std::allocator<std::_Rb_tree_node<std::pair<const ck::...   \n",
+       "7243836  std::unique_ptr<ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...   \n",
+       "7243837  std::__and_<std::__is_implicitly_default_constructible<ck::tensor_operation:...   \n",
+       "7243838  std::tuple<ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl...   \n",
+       "7243839  ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...   \n",
+       "7243840  std::unique_ptr<ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...   \n",
+       "7243841  std::pair<const ck::BlockGemmPipelineScheduler, std::basic_string<char>>::pa...   \n",
+       "7243842  ck::make_tuple<ck::integral_constant<int, 1>, ck::integral_constant<int, 4>,...   \n",
+       "7243843  ck::StaticBufferTupleOfVector<ck::AddressSpaceEnum::Vgpr, unsigned short, 64...   \n",
+       "7243844  std::is_nothrow_constructible<std::unique_ptr<ck::tensor_operation::device::...   \n",
+       "\n",
+       "         count    total_s   mean_ms  depth  \n",
+       "7243825   9406  27.985313  2.975262      3  \n",
+       "7243826   9406  27.985313  2.975262      3  \n",
+       "7243827   9406  27.985313  2.975262      1  \n",
+       "7243828   9406  27.985313  2.975262      6  \n",
+       "7243829   9406  27.985313  2.975262      5  \n",
+       "7243830   9406  27.985313  2.975262      4  \n",
+       "7243831   9406  27.985313  2.975262      4  \n",
+       "7243832   9406  27.985313  2.975262      5  \n",
+       "7243833   9406  27.985313  2.975262      1  \n",
+       "7243834   9406  27.985313  2.975262      4  \n",
+       "7243835   9406  27.985313  2.975262      5  \n",
+       "7243836   9406  27.985313  2.975262      3  \n",
+       "7243837   9406  27.985313  2.975262      5  \n",
+       "7243838   9406  27.985313  2.975262      4  \n",
+       "7243839   9406  27.985313  2.975262      5  \n",
+       "7243840   9406  27.985313  2.975262      4  \n",
+       "7243841   9406  27.985313  2.975262      2  \n",
+       "7243842   9406  27.985313  2.975262      2  \n",
+       "7243843   9406  27.985313  2.975262      1  \n",
+       "7243844   9406  27.985313  2.975262      4  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "if json_files and len(template_stats) > 0:\n",
     "    print(\"Top 20 Most Frequently Instantiated Templates:\")\n",
@@ -699,9 +2121,209 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 15,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Top 15 Optimization Targets (High Frequency + High Cost):\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>full_signature</th>\n",
+       "      <th>count</th>\n",
+       "      <th>total_s</th>\n",
+       "      <th>mean_ms</th>\n",
+       "      <th>priority_score</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>7243825</th>\n",
+       "      <td>ck::Tuple&lt;ck::PassThrough&lt;int&gt;, ck::UnMerge&lt;ck::Tuple&lt;int, int&gt;, false&gt;, ck:...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243826</th>\n",
+       "      <td>std::__is_implicitly_default_constructible&lt;ck::tensor_operation::device::Dev...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243827</th>\n",
+       "      <td>ck::StaticBufferTupleOfVector&lt;ck::AddressSpaceEnum::Vgpr, _Float16, 4, 1, tr...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243828</th>\n",
+       "      <td>ck::ThreadwiseTensorSliceTransfer_v3r2&lt;const ck::Sequence&lt;2, 2&gt;, ck::tensor_...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243829</th>\n",
+       "      <td>ck::utility::launch_and_time_kernel_with_preprocess&lt;false, ck::GridwiseGemmM...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243830</th>\n",
+       "      <td>std::unique_ptr&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243831</th>\n",
+       "      <td>std::tuple&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243832</th>\n",
+       "      <td>ck::TensorDescriptor&lt;ck::Tuple&lt;ck::Embed&lt;ck::Tuple&lt;ck::integral_constant&lt;int...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243833</th>\n",
+       "      <td>ck::generate_tuple&lt;(lambda at /home/AMD/jshumway/composable_kernel/include/c...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243834</th>\n",
+       "      <td>std::__uniq_ptr_data&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMulti...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243835</th>\n",
+       "      <td>std::allocator_traits&lt;std::allocator&lt;std::_Rb_tree_node&lt;std::pair&lt;const ck::...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243836</th>\n",
+       "      <td>std::unique_ptr&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243837</th>\n",
+       "      <td>std::__and_&lt;std::__is_implicitly_default_constructible&lt;ck::tensor_operation:...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243838</th>\n",
+       "      <td>std::tuple&lt;ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7243839</th>\n",
+       "      <td>ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...</td>\n",
+       "      <td>9406</td>\n",
+       "      <td>27.985313</td>\n",
+       "      <td>2.975262</td>\n",
+       "      <td>0.500016</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                                          full_signature  \\\n",
+       "7243825  ck::Tuple<ck::PassThrough<int>, ck::UnMerge<ck::Tuple<int, int>, false>, ck:...   \n",
+       "7243826  std::__is_implicitly_default_constructible<ck::tensor_operation::device::Dev...   \n",
+       "7243827  ck::StaticBufferTupleOfVector<ck::AddressSpaceEnum::Vgpr, _Float16, 4, 1, tr...   \n",
+       "7243828  ck::ThreadwiseTensorSliceTransfer_v3r2<const ck::Sequence<2, 2>, ck::tensor_...   \n",
+       "7243829  ck::utility::launch_and_time_kernel_with_preprocess<false, ck::GridwiseGemmM...   \n",
+       "7243830  std::unique_ptr<ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...   \n",
+       "7243831  std::tuple<ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl...   \n",
+       "7243832  ck::TensorDescriptor<ck::Tuple<ck::Embed<ck::Tuple<ck::integral_constant<int...   \n",
+       "7243833  ck::generate_tuple<(lambda at /home/AMD/jshumway/composable_kernel/include/c...   \n",
+       "7243834  std::__uniq_ptr_data<ck::tensor_operation::device::DeviceGroupedConvFwdMulti...   \n",
+       "7243835  std::allocator_traits<std::allocator<std::_Rb_tree_node<std::pair<const ck::...   \n",
+       "7243836  std::unique_ptr<ck::tensor_operation::device::DeviceGroupedConvFwdMultipleAB...   \n",
+       "7243837  std::__and_<std::__is_implicitly_default_constructible<ck::tensor_operation:...   \n",
+       "7243838  std::tuple<ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl...   \n",
+       "7243839  ck::tensor_operation::device::DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle_V...   \n",
+       "\n",
+       "         count    total_s   mean_ms  priority_score  \n",
+       "7243825   9406  27.985313  2.975262        0.500016  \n",
+       "7243826   9406  27.985313  2.975262        0.500016  \n",
+       "7243827   9406  27.985313  2.975262        0.500016  \n",
+       "7243828   9406  27.985313  2.975262        0.500016  \n",
+       "7243829   9406  27.985313  2.975262        0.500016  \n",
+       "7243830   9406  27.985313  2.975262        0.500016  \n",
+       "7243831   9406  27.985313  2.975262        0.500016  \n",
+       "7243832   9406  27.985313  2.975262        0.500016  \n",
+       "7243833   9406  27.985313  2.975262        0.500016  \n",
+       "7243834   9406  27.985313  2.975262        0.500016  \n",
+       "7243835   9406  27.985313  2.975262        0.500016  \n",
+       "7243836   9406  27.985313  2.975262        0.500016  \n",
+       "7243837   9406  27.985313  2.975262        0.500016  \n",
+       "7243838   9406  27.985313  2.975262        0.500016  \n",
+       "7243839   9406  27.985313  2.975262        0.500016  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "if json_files and len(template_stats) > 0:\n",
     "    # Normalize count and mean to 0-1 range\n",
@@ -736,9 +2358,158 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 16,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Template Instantiation by Nesting Depth:\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>depth</th>\n",
+       "      <th>count</th>\n",
+       "      <th>total_s</th>\n",
+       "      <th>mean_ms</th>\n",
+       "      <th>median_ms</th>\n",
+       "      <th>pct_total</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>0</td>\n",
+       "      <td>348050</td>\n",
+       "      <td>2.759851e+03</td>\n",
+       "      <td>8.315829</td>\n",
+       "      <td>1.22725</td>\n",
+       "      <td>1.843240</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2007036528</td>\n",
+       "      <td>9.720831e+06</td>\n",
+       "      <td>6.516100</td>\n",
+       "      <td>1.22500</td>\n",
+       "      <td>6492.316278</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>2</td>\n",
+       "      <td>4212093106</td>\n",
+       "      <td>2.378498e+07</td>\n",
+       "      <td>7.811773</td>\n",
+       "      <td>1.28000</td>\n",
+       "      <td>15885.434415</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>3</td>\n",
+       "      <td>2137634038</td>\n",
+       "      <td>1.261997e+07</td>\n",
+       "      <td>7.344963</td>\n",
+       "      <td>1.27700</td>\n",
+       "      <td>8428.586130</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>4</td>\n",
+       "      <td>2777069926</td>\n",
+       "      <td>1.801647e+07</td>\n",
+       "      <td>8.056019</td>\n",
+       "      <td>1.29000</td>\n",
+       "      <td>12032.783621</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>5</td>\n",
+       "      <td>1651653946</td>\n",
+       "      <td>1.169701e+07</td>\n",
+       "      <td>8.960966</td>\n",
+       "      <td>1.31100</td>\n",
+       "      <td>7812.159446</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>6</td>\n",
+       "      <td>781689120</td>\n",
+       "      <td>8.147449e+06</td>\n",
+       "      <td>16.840407</td>\n",
+       "      <td>1.76800</td>\n",
+       "      <td>5441.491328</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>7</td>\n",
+       "      <td>69349065</td>\n",
+       "      <td>7.653282e+05</td>\n",
+       "      <td>16.927265</td>\n",
+       "      <td>1.79050</td>\n",
+       "      <td>511.144843</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>8</td>\n",
+       "      <td>1516615</td>\n",
+       "      <td>1.622227e+04</td>\n",
+       "      <td>16.672371</td>\n",
+       "      <td>1.79625</td>\n",
+       "      <td>10.834476</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>9</td>\n",
+       "      <td>2150</td>\n",
+       "      <td>1.411392e+01</td>\n",
+       "      <td>7.276202</td>\n",
+       "      <td>2.63050</td>\n",
+       "      <td>0.009426</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "   depth       count       total_s    mean_ms  median_ms     pct_total\n",
+       "0      0      348050  2.759851e+03   8.315829    1.22725      1.843240\n",
+       "1      1  2007036528  9.720831e+06   6.516100    1.22500   6492.316278\n",
+       "2      2  4212093106  2.378498e+07   7.811773    1.28000  15885.434415\n",
+       "3      3  2137634038  1.261997e+07   7.344963    1.27700   8428.586130\n",
+       "4      4  2777069926  1.801647e+07   8.056019    1.29000  12032.783621\n",
+       "5      5  1651653946  1.169701e+07   8.960966    1.31100   7812.159446\n",
+       "6      6   781689120  8.147449e+06  16.840407    1.76800   5441.491328\n",
+       "7      7    69349065  7.653282e+05  16.927265    1.79050    511.144843\n",
+       "8      8     1516615  1.622227e+04  16.672371    1.79625     10.834476\n",
+       "9      9        2150  1.411392e+01   7.276202    2.63050      0.009426"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "if json_files and len(instantiations_df) > 0:\n",
     "    # OPTIMIZATION: Aggregate by depth directly from instantiations, then join\n",
@@ -790,9 +2561,133 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 17,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Template Argument Type Distribution:\n",
+      "Type                 Count   Percentage\n",
+      "----------------------------------------\n",
+      "template        125,537,546        80.8%\n",
+      "primitive       19,045,151        12.3%\n",
+      "unknown         10,699,046         6.9%\n",
+      "\n",
+      "Templates with Most Arguments:\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>full_signature</th>\n",
+       "      <th>arg_count</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>231420</th>\n",
+       "      <td>ck::Tuple&lt;ck::f8_fnuz_t, ck::f8_fnuz_t, ck::f8_fnuz_t, ck::f8_fnuz_t, ck::f8...</td>\n",
+       "      <td>23716</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>231421</th>\n",
+       "      <td>ck::detail::TupleImpl&lt;ck::Sequence&lt;0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,...</td>\n",
+       "      <td>23716</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>231422</th>\n",
+       "      <td>ck::detail::TupleImpl&lt;ck::Sequence&lt;0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,...</td>\n",
+       "      <td>23716</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>231423</th>\n",
+       "      <td>ck::Tuple&lt;_Float16 __attribute__((ext_vector_type(2))), _Float16 __attribute...</td>\n",
+       "      <td>23716</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>231424</th>\n",
+       "      <td>ck::non_native_vector_base&lt;ck::f8_fnuz_t, 64&gt;::(unnamed union at /home/AMD/j...</td>\n",
+       "      <td>23716</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>231425</th>\n",
+       "      <td>ck::Tuple&lt;_Float16, _Float16, _Float16, _Float16, _Float16, _Float16, _Float...</td>\n",
+       "      <td>23716</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>231426</th>\n",
+       "      <td>ck::non_native_vector_base&lt;ck::f8_fnuz_t, 64&gt;::(unnamed union at /home/AMD/j...</td>\n",
+       "      <td>23716</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>231427</th>\n",
+       "      <td>ck::detail::TupleImpl&lt;ck::Sequence&lt;0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,...</td>\n",
+       "      <td>23716</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>231428</th>\n",
+       "      <td>ck::vector_type&lt;_Float16, 32&gt;::(unnamed union at /home/AMD/jshumway/composab...</td>\n",
+       "      <td>23716</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>231429</th>\n",
+       "      <td>ck::vector_type&lt;signed char, 4&gt;::(unnamed union at /home/AMD/jshumway/compos...</td>\n",
+       "      <td>23716</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                                         full_signature  \\\n",
+       "231420  ck::Tuple<ck::f8_fnuz_t, ck::f8_fnuz_t, ck::f8_fnuz_t, ck::f8_fnuz_t, ck::f8...   \n",
+       "231421  ck::detail::TupleImpl<ck::Sequence<0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,...   \n",
+       "231422  ck::detail::TupleImpl<ck::Sequence<0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,...   \n",
+       "231423  ck::Tuple<_Float16 __attribute__((ext_vector_type(2))), _Float16 __attribute...   \n",
+       "231424  ck::non_native_vector_base<ck::f8_fnuz_t, 64>::(unnamed union at /home/AMD/j...   \n",
+       "231425  ck::Tuple<_Float16, _Float16, _Float16, _Float16, _Float16, _Float16, _Float...   \n",
+       "231426  ck::non_native_vector_base<ck::f8_fnuz_t, 64>::(unnamed union at /home/AMD/j...   \n",
+       "231427  ck::detail::TupleImpl<ck::Sequence<0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,...   \n",
+       "231428  ck::vector_type<_Float16, 32>::(unnamed union at /home/AMD/jshumway/composab...   \n",
+       "231429  ck::vector_type<signed char, 4>::(unnamed union at /home/AMD/jshumway/compos...   \n",
+       "\n",
+       "        arg_count  \n",
+       "231420      23716  \n",
+       "231421      23716  \n",
+       "231422      23716  \n",
+       "231423      23716  \n",
+       "231424      23716  \n",
+       "231425      23716  \n",
+       "231426      23716  \n",
+       "231427      23716  \n",
+       "231428      23716  \n",
+       "231429      23716  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "if json_files and len(template_args_df) > 0:\n",
     "    # Count argument types\n",
@@ -829,9 +2724,259 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 18,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Top 20 Event Types by Total Duration:\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>event_type</th>\n",
+       "      <th>count</th>\n",
+       "      <th>total_min</th>\n",
+       "      <th>mean_ms</th>\n",
+       "      <th>pct_total</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>47</th>\n",
+       "      <td>InstantiateFunction</td>\n",
+       "      <td>15006405</td>\n",
+       "      <td>2069.145884</td>\n",
+       "      <td>8.273051</td>\n",
+       "      <td>34.981333</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>30</th>\n",
+       "      <td>ExecuteCompiler</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>445.214199</td>\n",
+       "      <td>21353.198979</td>\n",
+       "      <td>7.526867</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>143</th>\n",
+       "      <td>Total ExecuteCompiler</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>445.214188</td>\n",
+       "      <td>21353.198487</td>\n",
+       "      <td>7.526867</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>46</th>\n",
+       "      <td>InstantiateClass</td>\n",
+       "      <td>2430595</td>\n",
+       "      <td>426.324732</td>\n",
+       "      <td>10.523960</td>\n",
+       "      <td>7.207519</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>147</th>\n",
+       "      <td>Total Frontend</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>384.832270</td>\n",
+       "      <td>18457.183201</td>\n",
+       "      <td>6.506040</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>32</th>\n",
+       "      <td>Frontend</td>\n",
+       "      <td>2419</td>\n",
+       "      <td>384.832095</td>\n",
+       "      <td>9545.235926</td>\n",
+       "      <td>6.506037</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>241</th>\n",
+       "      <td>Total Source</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>232.585137</td>\n",
+       "      <td>11155.162444</td>\n",
+       "      <td>3.932124</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>168</th>\n",
+       "      <td>Total InstantiateFunction</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>187.199754</td>\n",
+       "      <td>8978.405449</td>\n",
+       "      <td>3.164831</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>75</th>\n",
+       "      <td>PerformPendingInstantiations</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>109.696772</td>\n",
+       "      <td>5261.236084</td>\n",
+       "      <td>1.854552</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>217</th>\n",
+       "      <td>Total PerformPendingInstantiations</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>109.696762</td>\n",
+       "      <td>5261.235595</td>\n",
+       "      <td>1.854552</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>167</th>\n",
+       "      <td>Total InstantiateClass</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>101.970700</td>\n",
+       "      <td>4890.681049</td>\n",
+       "      <td>1.723934</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>211</th>\n",
+       "      <td>Total ParseClass</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>65.367502</td>\n",
+       "      <td>3135.131968</td>\n",
+       "      <td>1.105114</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>69</th>\n",
+       "      <td>ParseClass</td>\n",
+       "      <td>615160</td>\n",
+       "      <td>60.689238</td>\n",
+       "      <td>5.919361</td>\n",
+       "      <td>1.026023</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Backend</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>57.471167</td>\n",
+       "      <td>2756.410890</td>\n",
+       "      <td>0.971617</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>104</th>\n",
+       "      <td>Total Backend</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>57.471157</td>\n",
+       "      <td>2756.410412</td>\n",
+       "      <td>0.971617</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>70</th>\n",
+       "      <td>ParseDeclarationOrFunctionDefinition</td>\n",
+       "      <td>377890</td>\n",
+       "      <td>50.231873</td>\n",
+       "      <td>7.975634</td>\n",
+       "      <td>0.849229</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>212</th>\n",
+       "      <td>Total ParseDeclarationOrFunctionDefinition</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>44.632077</td>\n",
+       "      <td>2140.627196</td>\n",
+       "      <td>0.754557</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>71</th>\n",
+       "      <td>ParseFunctionDefinition</td>\n",
+       "      <td>410465</td>\n",
+       "      <td>38.754508</td>\n",
+       "      <td>5.664967</td>\n",
+       "      <td>0.655190</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>213</th>\n",
+       "      <td>Total ParseFunctionDefinition</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>38.036852</td>\n",
+       "      <td>1824.309444</td>\n",
+       "      <td>0.643057</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>68</th>\n",
+       "      <td>Optimizer</td>\n",
+       "      <td>1251</td>\n",
+       "      <td>37.729725</td>\n",
+       "      <td>1809.579141</td>\n",
+       "      <td>0.637865</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                     event_type     count    total_min  \\\n",
+       "47                          InstantiateFunction  15006405  2069.145884   \n",
+       "30                              ExecuteCompiler      1251   445.214199   \n",
+       "143                       Total ExecuteCompiler      1251   445.214188   \n",
+       "46                             InstantiateClass   2430595   426.324732   \n",
+       "147                              Total Frontend      1251   384.832270   \n",
+       "32                                     Frontend      2419   384.832095   \n",
+       "241                                Total Source      1251   232.585137   \n",
+       "168                   Total InstantiateFunction      1251   187.199754   \n",
+       "75                 PerformPendingInstantiations      1251   109.696772   \n",
+       "217          Total PerformPendingInstantiations      1251   109.696762   \n",
+       "167                      Total InstantiateClass      1251   101.970700   \n",
+       "211                            Total ParseClass      1251    65.367502   \n",
+       "69                                   ParseClass    615160    60.689238   \n",
+       "4                                       Backend      1251    57.471167   \n",
+       "104                               Total Backend      1251    57.471157   \n",
+       "70         ParseDeclarationOrFunctionDefinition    377890    50.231873   \n",
+       "212  Total ParseDeclarationOrFunctionDefinition      1251    44.632077   \n",
+       "71                      ParseFunctionDefinition    410465    38.754508   \n",
+       "213               Total ParseFunctionDefinition      1251    38.036852   \n",
+       "68                                    Optimizer      1251    37.729725   \n",
+       "\n",
+       "          mean_ms  pct_total  \n",
+       "47       8.273051  34.981333  \n",
+       "30   21353.198979   7.526867  \n",
+       "143  21353.198487   7.526867  \n",
+       "46      10.523960   7.207519  \n",
+       "147  18457.183201   6.506040  \n",
+       "32    9545.235926   6.506037  \n",
+       "241  11155.162444   3.932124  \n",
+       "168   8978.405449   3.164831  \n",
+       "75    5261.236084   1.854552  \n",
+       "217   5261.235595   1.854552  \n",
+       "167   4890.681049   1.723934  \n",
+       "211   3135.131968   1.105114  \n",
+       "69       5.919361   1.026023  \n",
+       "4     2756.410890   0.971617  \n",
+       "104   2756.410412   0.971617  \n",
+       "70       7.975634   0.849229  \n",
+       "212   2140.627196   0.754557  \n",
+       "71       5.664967   0.655190  \n",
+       "213   1824.309444   0.643057  \n",
+       "68    1809.579141   0.637865  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "if json_files and len(events_df) > 0:\n",
     "    event_stats = (\n",
@@ -873,9 +3018,118 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 19,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Found ninja log: ../../../build-trace/.ninja_log\n",
+      "Parsed 2,579 build events in 0.004s\n",
+      "\n",
+      "Builds DataFrame: 2,579 rows\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>target</th>\n",
+       "      <th>start_ms</th>\n",
+       "      <th>end_ms</th>\n",
+       "      <th>cmd_hash</th>\n",
+       "      <th>worker_id</th>\n",
+       "      <th>duration_ms</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>/home/AMD/jshumway/composable_kernel/build-trace/CMakeFiles/cmake.verify_globs</td>\n",
+       "      <td>487</td>\n",
+       "      <td>511</td>\n",
+       "      <td>20b4a9dba5446149</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>24</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>build.ninja</td>\n",
+       "      <td>512</td>\n",
+       "      <td>159966</td>\n",
+       "      <td>80f2b673d9d1b995</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>159454</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/CMakeFiles/devi...</td>\n",
+       "      <td>228</td>\n",
+       "      <td>9471</td>\n",
+       "      <td>b37d14fd75e2d29c</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>9243</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>_deps/gtest-build/googletest/CMakeFiles/gtest.dir/src/gtest-all.cc.o</td>\n",
+       "      <td>103</td>\n",
+       "      <td>9527</td>\n",
+       "      <td>330bec9abe978cde</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>9424</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/CMakeFiles/devi...</td>\n",
+       "      <td>243</td>\n",
+       "      <td>9687</td>\n",
+       "      <td>6a5b0a35a99473e8</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>9444</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                                            target  \\\n",
+       "0   /home/AMD/jshumway/composable_kernel/build-trace/CMakeFiles/cmake.verify_globs   \n",
+       "1                                                                      build.ninja   \n",
+       "2  library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/CMakeFiles/devi...   \n",
+       "3             _deps/gtest-build/googletest/CMakeFiles/gtest.dir/src/gtest-all.cc.o   \n",
+       "4  library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/CMakeFiles/devi...   \n",
+       "\n",
+       "   start_ms  end_ms          cmd_hash  worker_id  duration_ms  \n",
+       "0       487     511  20b4a9dba5446149         -1           24  \n",
+       "1       512  159966  80f2b673d9d1b995         -1       159454  \n",
+       "2       228    9471  b37d14fd75e2d29c         -1         9243  \n",
+       "3       103    9527  330bec9abe978cde         -1         9424  \n",
+       "4       243    9687  6a5b0a35a99473e8         -1         9444  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
    "source": [
     "from trace_analysis import NinjaLogParser\n",
     "\n",