From 0d12aad8fd38c746cf4fc15f2ce152734298a362 Mon Sep 17 00:00:00 2001
From: chenhao <8516202+chenhao_1209@user.noreply.gitee.com>
Date: Thu, 22 May 2025 11:54:11 +0800
Subject: [PATCH] add jupyter for slow rank
---
.../recipes/cluster_display.py | 23 +
.../recipes/p2p_pairing/stats.ipynb | 4068 +++++++++++++++++
.../recipes/slow_rank/slow_rank.py | 190 +
3 files changed, 4281 insertions(+)
create mode 100644 profiler/msprof_analyze/cluster_analyse/recipes/p2p_pairing/stats.ipynb
create mode 100644 profiler/msprof_analyze/cluster_analyse/recipes/slow_rank/slow_rank.py
diff --git a/profiler/msprof_analyze/cluster_analyse/recipes/cluster_display.py b/profiler/msprof_analyze/cluster_analyse/recipes/cluster_display.py
index 35e7ef63872..7e8913948f7 100644
--- a/profiler/msprof_analyze/cluster_analyse/recipes/cluster_display.py
+++ b/profiler/msprof_analyze/cluster_analyse/recipes/cluster_display.py
@@ -192,6 +192,29 @@ def display_graph(figs, x_axis, y_axes, title=None,
figs.append(fig)
+def display_bar(x_axis, y_axes, title=None, y_index=None):
+ if isinstance(y_axes, pd.DataFrame):
+ data = y_axes.set_index(x_axis)
+ elif isinstance(y_axes, dict):
+ data = pd.DataFrame(y_axes, index=x_axis)
+ elif isinstance(y_axes, pd.Series):
+ data = pd.DataFrame({"": y_axes}, index=x_axis)
+ elif isinstance(y_axes, np.ndarray):
+ data = pd.DataFrame({"": pd.Series(y_axes)}, index=x_axis)
+ else:
+ return
+
+ fig = data.plot.bar(title=title)
+ fig.bar_label(fig.containers[0])
+ if y_index is not None and y_index in y_axes:
+ # get index of the top1
+ top1_indices = data[y_index].nlargest(1).index
+ # change the color for the top1
+ for i, bar in enumerate(fig.patches):
+ if data.index[i] in top1_indices:
+ bar.set_color('#FFA500') # highlight in orange
+
+
def display_stats_per_rank_groups_combobox(rank_stats_gdf):
names = list(rank_stats_gdf.groups.keys())
if len(names) > 1:
diff --git a/profiler/msprof_analyze/cluster_analyse/recipes/p2p_pairing/stats.ipynb b/profiler/msprof_analyze/cluster_analyse/recipes/p2p_pairing/stats.ipynb
new file mode 100644
index 00000000000..0573ba03bb6
--- /dev/null
+++ b/profiler/msprof_analyze/cluster_analyse/recipes/p2p_pairing/stats.ipynb
@@ -0,0 +1,4068 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "jp-MarkdownHeadingCollapsed": true
+ },
+ "source": [
+ "## Slow Rank\n",
+ "集群场景通信算子快慢卡汇总分析\n",
+ "\n",
+ "1.根据卡粒度，统计每个Rank上的影响因子\n",
+ "\n",
+ "2.将统计的结果按柱状图呈现，TOP影响的极为慢卡候选"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## 数据准备"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "metadata": {
+ "vscode": {
+ "languageId": "plaintext"
+ }
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ ""
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ " \n",
+ " \n",
+ " "
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "import pandas as pd\n",
+ "import plotly.offline as pyo\n",
+ "\n",
+ "from IPython.display import display, HTML\n",
+ "\n",
+ "import cluster_display\n",
+ "\n",
+ "display(HTML(\"\"))\n",
+ "pd.set_option('display.max_columns', None)\n",
+ "pd.set_option('display.max_rows', None)\n",
+ "pyo.init_notebook_mode()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "展示各Rank受影响程度的统计表"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " slowAffectCount | \n",
+ "
\n",
+ " \n",
+ " | rankId | \n",
+ " | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " 10 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " 9 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " 15 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "