{ "cells": [ { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "FileUpload doesn't work with vscode, please copy resulting database into `/tmp/jnw` directory." ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import os\n", "import cub \n", "\n", "from ipywidgets import widgets\n", "from IPython.display import display, clear_output\n", "\n", "if not os.path.exists('/tmp/jnw'):\n", " os.mkdir('/tmp/jnw')\n", "os.chdir('/tmp/jnw')" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "To analyze results, let's select one algorithm first:" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "9d392f48d3d34f5b9c69e040c246ccf1", "version_major": 2, "version_minor": 0 }, "text/plain": [ "Dropdown(description='Algorithm:', options=('cub.bench.merge_sort.keys', 'cub.bench.merge_sort.pairs', 'cub.be…" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "37521c7a22c34a2c803b2c661f378834", "version_major": 2, "version_minor": 0 }, "text/plain": [ "Output()" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "storage = cub.bench.Storage()\n", "\n", "# Create a list of options for the dropdown menu\n", "algorithm_list = storage.algnames()\n", "\n", "# Create an output widget to display the dataframe\n", "output = widgets.Output()\n", "\n", "# Create the dropdown menu widget\n", "dropdown = widgets.Dropdown(\n", " options=algorithm_list,\n", " value=algorithm_list[0],\n", " description='Algorithm:',\n", " disabled=False,\n", ")\n", "\n", "global df\n", "\n", "df = storage.alg_to_df(algorithm_list[0])\n", "def on_dropdown_change(change):\n", " global df\n", " if change['type'] == 'change' and change['name'] == 'value':\n", " df = storage.alg_to_df(change['new'])\n", " with output:\n", " clear_output()\n", " # Display the selected dataframe column\n", " display(df)\n", "\n", "# Attach the event handler to the dropdown menu\n", "dropdown.observe(on_dropdown_change)\n", "\n", "display(dropdown)\n", "\n", "with output:\n", " display(df)\n", "\n", "display(output)" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Now we should select a particular CTK, GPU and CUB:" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "2e23b0f8960b4831bfd3ee7ffe343b07", "version_major": 2, "version_minor": 0 }, "text/plain": [ "HBox(children=(Dropdown(description='GPU:', options=('GeForce GTX 1070 (256)',), value='GeForce GTX 1070 (256)…" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "gpus_list = df['gpu'].unique()\n", "ctks_list = df['ctk'].unique()\n", "cubs_list = df['cub'].unique()\n", "\n", "global gpu \n", "global ctk\n", "global cub\n", "\n", "gpu = gpus_list[0]\n", "ctk = ctks_list[0]\n", "cub = cubs_list[0]\n", "\n", "def create_dropdown_menu(name, options, callback):\n", " dropdown = widgets.Dropdown(\n", " options=options,\n", " value=options[0],\n", " description=name,\n", " disabled=False)\n", " dropdown.observe(callback)\n", " return dropdown\n", "\n", "def on_gpu_change(change):\n", " global gpu\n", " if change['type'] == 'change' and change['name'] == 'value':\n", " gpu = change['new']\n", "\n", "def on_ctk_change(change):\n", " global ctk\n", " if change['type'] == 'change' and change['name'] == 'value':\n", " ctk = change['new']\n", "\n", "def on_cub_change(change):\n", " global cub\n", " if change['type'] == 'change' and change['name'] == 'value':\n", " cub = change['new']\n", "\n", "gpu_dropdown = create_dropdown_menu('GPU:', gpus_list, on_gpu_change)\n", "ctk_dropdown = create_dropdown_menu('CTK:', ctks_list, on_ctk_change)\n", "cub_dropdown = create_dropdown_menu('CUB:', cubs_list, on_cub_change)\n", "\n", "dropdown_row = widgets.HBox([gpu_dropdown, ctk_dropdown, cub_dropdown])\n", "display(dropdown_row)" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "Let's look at the data now" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
variantelapsedcentersamplesT{ct}OffsetT{ct}Elements{io}[pow2]Entropy
0base1.6441180.000040[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...I8I32161.000
1ipt_19.tpb_1601.6177690.000046[3.8912e-05, 3.8912e-05, 3.8912e-05, 3.8912e-0...I8I32161.000
2base1.6720300.000039[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...I8I32160.811
3ipt_19.tpb_1601.6165260.000046[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...I8I32160.811
4base1.6646730.000040[3.6864003e-05, 3.7888003e-05, 3.7888003e-05, ...I8I32160.544
...........................
3331ipt_15.tpb_5761.4166530.017443[0.017391616, 0.017420288, 0.01742848, 0.01743...I8I64280.811
3332ipt_15.tpb_5761.7676520.022247[0.022086656, 0.022111233, 0.022162432, 0.0221...I8I64280.544
3333ipt_15.tpb_5762.0669450.026129[0.025906175, 0.0259072, 0.026003456, 0.026004...I8I64280.337
3334ipt_15.tpb_5762.3672390.028263[0.028086271, 0.02810163, 0.028107777, 0.02821...I8I64280.201
3335ipt_15.tpb_5761.8162690.024484[0.024112128, 0.024114177, 0.024117248, 0.0241...I8I64280.000
\n", "

3336 rows × 8 columns

\n", "
" ], "text/plain": [ " variant elapsed center \\\n", "0 base 1.644118 0.000040 \n", "1 ipt_19.tpb_160 1.617769 0.000046 \n", "2 base 1.672030 0.000039 \n", "3 ipt_19.tpb_160 1.616526 0.000046 \n", "4 base 1.664673 0.000040 \n", "... ... ... ... \n", "3331 ipt_15.tpb_576 1.416653 0.017443 \n", "3332 ipt_15.tpb_576 1.767652 0.022247 \n", "3333 ipt_15.tpb_576 2.066945 0.026129 \n", "3334 ipt_15.tpb_576 2.367239 0.028263 \n", "3335 ipt_15.tpb_576 1.816269 0.024484 \n", "\n", " samples T{ct} OffsetT{ct} \\\n", "0 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... I8 I32 \n", "1 [3.8912e-05, 3.8912e-05, 3.8912e-05, 3.8912e-0... I8 I32 \n", "2 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... I8 I32 \n", "3 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... I8 I32 \n", "4 [3.6864003e-05, 3.7888003e-05, 3.7888003e-05, ... I8 I32 \n", "... ... ... ... \n", "3331 [0.017391616, 0.017420288, 0.01742848, 0.01743... I8 I64 \n", "3332 [0.022086656, 0.022111233, 0.022162432, 0.0221... I8 I64 \n", "3333 [0.025906175, 0.0259072, 0.026003456, 0.026004... I8 I64 \n", "3334 [0.028086271, 0.02810163, 0.028107777, 0.02821... I8 I64 \n", "3335 [0.024112128, 0.024114177, 0.024117248, 0.0241... I8 I64 \n", "\n", " Elements{io}[pow2] Entropy \n", "0 16 1.000 \n", "1 16 1.000 \n", "2 16 0.811 \n", "3 16 0.811 \n", "4 16 0.544 \n", "... ... ... \n", "3331 28 0.811 \n", "3332 28 0.544 \n", "3333 28 0.337 \n", "3334 28 0.201 \n", "3335 28 0.000 \n", "\n", "[3336 rows x 8 columns]" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "alg_df = df.loc[(df['gpu'] == gpu) & (df['ctk'] == ctk) & (df['cub'] == cub) & (df['center'] != float('inf'))].copy()\n", "alg_df.drop(columns=['gpu', 'ctk', 'cub'], inplace=True)\n", "alg_df" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
variantelapsedcentersamplesT{ct}OffsetT{ct}Elements{io}[pow2]Entropybase_samplesspeedup
0base1.6441180.000040[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...I8I32161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...1.000000
1ipt_19.tpb_1601.6177690.000046[3.8912e-05, 3.8912e-05, 3.8912e-05, 3.8912e-0...I8I32161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.866667
2ipt_8.tpb_9601.6170640.000043[3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3...I8I32161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.928571
3ipt_14.tpb_3201.6671540.000042[3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0...I8I32161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.951220
4ipt_10.tpb_5761.6672080.000042[3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0...I8I32161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.951220
.................................
3331ipt_16.tpb_2561.6165770.022057[0.022006784, 0.022012929, 0.022017023, 0.0220...I8I64280.000[0.009255935, 0.009261056, 0.009263105, 0.0092...0.420706
3332ipt_20.tpb_4481.6661830.022471[0.022202367, 0.022203391, 0.022276096, 0.0222...I8I64280.000[0.009255935, 0.009261056, 0.009263105, 0.0092...0.412960
3333ipt_17.tpb_2561.6668300.022688[0.022679552, 0.022679552, 0.022679552, 0.0226...I8I64280.000[0.009255935, 0.009261056, 0.009263105, 0.0092...0.409009
3334ipt_15.tpb_5441.6664950.022356[0.022345727, 0.022346752, 0.022349825, 0.0223...I8I64280.000[0.009255935, 0.009261056, 0.009263105, 0.0092...0.415079
3335ipt_15.tpb_5761.8162690.024484[0.024112128, 0.024114177, 0.024117248, 0.0241...I8I64280.000[0.009255935, 0.009261056, 0.009263105, 0.0092...0.379005
\n", "

3336 rows × 10 columns

\n", "
" ], "text/plain": [ " variant elapsed center \\\n", "0 base 1.644118 0.000040 \n", "1 ipt_19.tpb_160 1.617769 0.000046 \n", "2 ipt_8.tpb_960 1.617064 0.000043 \n", "3 ipt_14.tpb_320 1.667154 0.000042 \n", "4 ipt_10.tpb_576 1.667208 0.000042 \n", "... ... ... ... \n", "3331 ipt_16.tpb_256 1.616577 0.022057 \n", "3332 ipt_20.tpb_448 1.666183 0.022471 \n", "3333 ipt_17.tpb_256 1.666830 0.022688 \n", "3334 ipt_15.tpb_544 1.666495 0.022356 \n", "3335 ipt_15.tpb_576 1.816269 0.024484 \n", "\n", " samples T{ct} OffsetT{ct} \\\n", "0 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... I8 I32 \n", "1 [3.8912e-05, 3.8912e-05, 3.8912e-05, 3.8912e-0... I8 I32 \n", "2 [3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3... I8 I32 \n", "3 [3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0... I8 I32 \n", "4 [3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0... I8 I32 \n", "... ... ... ... \n", "3331 [0.022006784, 0.022012929, 0.022017023, 0.0220... I8 I64 \n", "3332 [0.022202367, 0.022203391, 0.022276096, 0.0222... I8 I64 \n", "3333 [0.022679552, 0.022679552, 0.022679552, 0.0226... I8 I64 \n", "3334 [0.022345727, 0.022346752, 0.022349825, 0.0223... I8 I64 \n", "3335 [0.024112128, 0.024114177, 0.024117248, 0.0241... I8 I64 \n", "\n", " Elements{io}[pow2] Entropy \\\n", "0 16 1.000 \n", "1 16 1.000 \n", "2 16 1.000 \n", "3 16 1.000 \n", "4 16 1.000 \n", "... ... ... \n", "3331 28 0.000 \n", "3332 28 0.000 \n", "3333 28 0.000 \n", "3334 28 0.000 \n", "3335 28 0.000 \n", "\n", " base_samples speedup \n", "0 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 1.000000 \n", "1 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.866667 \n", "2 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.928571 \n", "3 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.951220 \n", "4 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.951220 \n", "... ... ... \n", "3331 [0.009255935, 0.009261056, 0.009263105, 0.0092... 0.420706 \n", "3332 [0.009255935, 0.009261056, 0.009263105, 0.0092... 0.412960 \n", "3333 [0.009255935, 0.009261056, 0.009263105, 0.0092... 0.409009 \n", "3334 [0.009255935, 0.009261056, 0.009263105, 0.0092... 0.415079 \n", "3335 [0.009255935, 0.009261056, 0.009263105, 0.0092... 0.379005 \n", "\n", "[3336 rows x 10 columns]" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "bench_columns = ['variant', 'elapsed', 'center', 'samples']\n", "workload_columns = [col for col in alg_df.columns if col not in bench_columns]\n", "# variants = alg_df['variant'].unique()\n", "\n", "base_df = alg_df[alg_df['variant'] == 'base'].drop(columns=['variant']).rename(columns={'center': 'base_center', 'samples': 'base_samples'})\n", "base_df.drop(columns=['elapsed'], inplace=True)\n", "\n", "merged_df = alg_df.merge(base_df, on=[col for col in alg_df.columns if col in workload_columns])\n", "merged_df['speedup'] = merged_df['base_center'] / merged_df['center']\n", "merged_df = merged_df.drop(columns=['base_center'])\n", "bench_columns.append('speedup')\n", "bench_columns.append('base_samples')\n", "merged_df" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "5cf0a28e9cd142b09be7d77c5b58fb6d", "version_major": 2, "version_minor": 0 }, "text/plain": [ "HBox(children=(Dropdown(description='T', options=('I8',), value='I8'), Dropdown(description='OffsetT', options…" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import functools\n", "\n", "ct_axes_values = {}\n", "for col in merged_df.columns:\n", " if '{ct}' in col:\n", " ct_axes_values[col] = merged_df[col].unique()\n", "\n", "global ct_axes_chosen_values\n", "ct_axes_chosen_values = {}\n", "\n", "for ct_axis in ct_axes_values:\n", " ct_axes_chosen_values[ct_axis] = ct_axes_values[ct_axis][0]\n", "\n", "def on_ct_value_change(ct_axis, change):\n", " global ct_axes_chosen_values\n", " if change['type'] == 'change' and change['name'] == 'value':\n", " ct_axes_chosen_values[ct_axis] = change['new']\n", "\n", "ct_dropdowns = []\n", "for ct_axis in ct_axes_values:\n", " ct_dropdowns.append(create_dropdown_menu(ct_axis.replace('{ct}', ''), ct_axes_values[ct_axis], functools.partial(on_ct_value_change, ct_axis)))\n", "\n", "ct_dropdown_row = widgets.HBox(ct_dropdowns)\n", "display(ct_dropdown_row)" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
variantelapsedcentersamplesElements{io}[pow2]Entropybase_samplesspeedup
0base1.6441180.000040[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...1.000000
1ipt_19.tpb_1601.6177690.000046[3.8912e-05, 3.8912e-05, 3.8912e-05, 3.8912e-0...161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.866667
2ipt_8.tpb_9601.6170640.000043[3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3...161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.928571
3ipt_14.tpb_3201.6671540.000042[3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0...161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.951220
4ipt_10.tpb_5761.6672080.000042[3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0...161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.951220
...........................
1819ipt_24.tpb_10241.6168160.022079[0.022068225, 0.022069247, 0.022069247, 0.0220...280.000[0.009991168, 0.009993216, 0.009995264, 0.0099...0.453437
1820ipt_24.tpb_5761.7166440.022847[0.02249728, 0.022499328, 0.022675455, 0.02268...280.000[0.009991168, 0.009993216, 0.009995264, 0.0099...0.438195
1821ipt_22.tpb_1281.8166210.024490[0.024477696, 0.024480768, 0.024481792, 0.0244...280.000[0.009991168, 0.009993216, 0.009995264, 0.0099...0.408806
1822ipt_23.tpb_5121.6665520.022298[0.02228736, 0.022290433, 0.022291455, 0.02229...280.000[0.009991168, 0.009993216, 0.009995264, 0.0099...0.449001
1823ipt_24.tpb_3201.6666860.022458[0.022228992, 0.022247422, 0.022256639, 0.0222...280.000[0.009991168, 0.009993216, 0.009995264, 0.0099...0.445787
\n", "

1824 rows × 8 columns

\n", "
" ], "text/plain": [ " variant elapsed center \\\n", "0 base 1.644118 0.000040 \n", "1 ipt_19.tpb_160 1.617769 0.000046 \n", "2 ipt_8.tpb_960 1.617064 0.000043 \n", "3 ipt_14.tpb_320 1.667154 0.000042 \n", "4 ipt_10.tpb_576 1.667208 0.000042 \n", "... ... ... ... \n", "1819 ipt_24.tpb_1024 1.616816 0.022079 \n", "1820 ipt_24.tpb_576 1.716644 0.022847 \n", "1821 ipt_22.tpb_128 1.816621 0.024490 \n", "1822 ipt_23.tpb_512 1.666552 0.022298 \n", "1823 ipt_24.tpb_320 1.666686 0.022458 \n", "\n", " samples Elements{io}[pow2] \\\n", "0 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 16 \n", "1 [3.8912e-05, 3.8912e-05, 3.8912e-05, 3.8912e-0... 16 \n", "2 [3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3... 16 \n", "3 [3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0... 16 \n", "4 [3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0... 16 \n", "... ... ... \n", "1819 [0.022068225, 0.022069247, 0.022069247, 0.0220... 28 \n", "1820 [0.02249728, 0.022499328, 0.022675455, 0.02268... 28 \n", "1821 [0.024477696, 0.024480768, 0.024481792, 0.0244... 28 \n", "1822 [0.02228736, 0.022290433, 0.022291455, 0.02229... 28 \n", "1823 [0.022228992, 0.022247422, 0.022256639, 0.0222... 28 \n", "\n", " Entropy base_samples speedup \n", "0 1.000 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 1.000000 \n", "1 1.000 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.866667 \n", "2 1.000 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.928571 \n", "3 1.000 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.951220 \n", "4 1.000 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.951220 \n", "... ... ... ... \n", "1819 0.000 [0.009991168, 0.009993216, 0.009995264, 0.0099... 0.453437 \n", "1820 0.000 [0.009991168, 0.009993216, 0.009995264, 0.0099... 0.438195 \n", "1821 0.000 [0.009991168, 0.009993216, 0.009995264, 0.0099... 0.408806 \n", "1822 0.000 [0.009991168, 0.009993216, 0.009995264, 0.0099... 0.449001 \n", "1823 0.000 [0.009991168, 0.009993216, 0.009995264, 0.0099... 0.445787 \n", "\n", "[1824 rows x 8 columns]" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "tuning_df_loc = None \n", "\n", "for ct_axis in ct_axes_values:\n", " if tuning_df_loc is None:\n", " tuning_df_loc = (merged_df[ct_axis] == ct_axes_chosen_values[ct_axis])\n", " else:\n", " tuning_df_loc = tuning_df_loc & (merged_df[ct_axis] == ct_axes_chosen_values[ct_axis])\n", "\n", "tuning_df = merged_df.loc[tuning_df_loc].copy()\n", "\n", "for ct_axis in ct_axes_values:\n", " tuning_df.drop(columns=[ct_axis], inplace=True)\n", "\n", "tuning_df" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
variantelapsedcentersamplesElements{io}[pow2]Entropybase_samplesspeedup
43ipt_14.tpb_3521.6670970.000042[3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0...161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.951220
46ipt_17.tpb_2561.6660770.000042[3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0...161.000[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.951220
119ipt_14.tpb_3521.6661180.000043[3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0...160.811[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.904762
122ipt_17.tpb_2561.6167900.000044[3.584e-05, 3.6864003e-05, 3.6864003e-05, 3.68...160.811[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.883721
195ipt_14.tpb_3521.6663900.000043[3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3...160.544[3.6864003e-05, 3.7888003e-05, 3.7888003e-05, ...0.928571
198ipt_17.tpb_2561.6662350.000044[3.6864003e-05, 3.6864003e-05, 3.6864003e-05, ...160.544[3.6864003e-05, 3.7888003e-05, 3.7888003e-05, ...0.906977
271ipt_14.tpb_3521.6666750.000043[3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3...160.337[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.928571
274ipt_17.tpb_2561.6161540.000044[3.6864003e-05, 3.6864003e-05, 3.6864003e-05, ...160.337[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.906977
347ipt_14.tpb_3521.6664530.000043[3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3...160.201[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.928571
350ipt_17.tpb_2561.6164550.000045[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...160.201[3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ...0.886364
423ipt_14.tpb_3521.6662180.000043[3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0...160.000[2.4576e-05, 2.4576e-05, 2.4576e-05, 2.4576e-0...0.595238
426ipt_17.tpb_2561.6168590.000043[3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3...160.000[2.4576e-05, 2.4576e-05, 2.4576e-05, 2.4576e-0...0.595238
499ipt_14.tpb_3521.4158100.000078[6.9632e-05, 6.9632e-05, 6.9632e-05, 6.9632e-0...201.000[0.000110592, 0.000110592, 0.000110592, 0.0001...1.578947
502ipt_17.tpb_2561.4160470.000076[6.8608e-05, 6.8608e-05, 6.9632e-05, 6.9632e-0...201.000[0.000110592, 0.000110592, 0.000110592, 0.0001...1.621621
575ipt_14.tpb_3521.4163570.000086[7.8847996e-05, 7.8847996e-05, 7.8847996e-05, ...200.811[0.000118784, 0.000119808006, 0.000119808006, ...1.404762
578ipt_17.tpb_2561.4157710.000085[7.7824e-05, 7.8847996e-05, 7.8847996e-05, 7.8...200.811[0.000118784, 0.000119808006, 0.000119808006, ...1.421687
651ipt_14.tpb_3521.3672510.000104[9.6256e-05, 9.7280004e-05, 9.7280004e-05, 9.7...200.544[0.000119808006, 0.000119808006, 0.00011980800...1.166667
654ipt_17.tpb_2561.3668590.000102[9.5232004e-05, 9.5232004e-05, 9.5232004e-05, ...200.544[0.000119808006, 0.000119808006, 0.00011980800...1.190000
727ipt_14.tpb_3521.3660810.000123[0.000110592, 0.000111616, 0.000111616, 0.0001...200.337[0.000117760006, 0.000118784, 0.000118784, 0.0...0.983333
730ipt_17.tpb_2561.3666570.000120[0.000109568, 0.000109568, 0.000110592, 0.0001...200.337[0.000117760006, 0.000118784, 0.000118784, 0.0...1.008547
803ipt_14.tpb_3521.3655970.000127[0.000119808006, 0.000120832, 0.000120832, 0.0...200.201[0.000119808006, 0.000119808006, 0.00011980800...0.959677
806ipt_17.tpb_2561.3655720.000126[0.000118784, 0.000118784, 0.000118784, 0.0001...200.201[0.000119808006, 0.000119808006, 0.00011980800...0.967480
879ipt_14.tpb_3521.3163590.000120[0.00011264, 0.00011264, 0.00011264, 0.0001126...200.000[4.9152e-05, 4.9152e-05, 4.9152e-05, 4.9152e-0...0.427350
882ipt_17.tpb_2561.3165600.000119[0.000111616, 0.000111616, 0.000111616, 0.0001...200.000[4.9152e-05, 4.9152e-05, 4.9152e-05, 4.9152e-0...0.431034
955ipt_14.tpb_3521.2656530.000742[0.00071168, 0.00071168, 0.000713728, 0.000714...241.000[0.00142848, 0.001429504, 0.001431552, 0.00143...1.962759
958ipt_17.tpb_2561.2654220.000750[0.00072499196, 0.000726016, 0.00072704, 0.000...241.000[0.00142848, 0.001429504, 0.001431552, 0.00143...1.943989
1031ipt_14.tpb_3521.2666420.000843[0.00082841597, 0.00082944, 0.00082944, 0.0008...240.811[0.001435648, 0.0014366719, 0.0014366719, 0.00...1.730255
1034ipt_17.tpb_2561.2658980.000853[0.00083968, 0.00084172795, 0.00084172795, 0.0...240.811[0.001435648, 0.0014366719, 0.0014366719, 0.00...1.709484
1107ipt_14.tpb_3521.2655620.001140[0.001132544, 0.001133568, 0.001133568, 0.0011...240.544[0.001430528, 0.001432576, 0.001434624, 0.0014...1.279874
1110ipt_17.tpb_2561.2663330.001140[0.001132544, 0.001132544, 0.001132544, 0.0011...240.544[0.001430528, 0.001432576, 0.001434624, 0.0014...1.279874
1183ipt_14.tpb_3521.2657580.001382[0.001375232, 0.001375232, 0.001375232, 0.0013...240.337[0.0014172159, 0.001419264, 0.001421312, 0.001...1.047407
1186ipt_17.tpb_2561.2656170.001369[0.001362944, 0.001362944, 0.001362944, 0.0013...240.337[0.0014172159, 0.001419264, 0.001421312, 0.001...1.057592
1259ipt_14.tpb_3521.2659350.001530[0.0015226881, 0.0015226881, 0.0015226881, 0.0...240.201[0.001415168, 0.0014172159, 0.001419264, 0.001...0.947791
1262ipt_17.tpb_2561.2654640.001510[0.001504256, 0.001504256, 0.001504256, 0.0015...240.201[0.001415168, 0.0014172159, 0.001419264, 0.001...0.960000
1335ipt_14.tpb_3521.2154660.001463[0.001455104, 0.001456128, 0.001456128, 0.0014...240.000[0.00064716797, 0.00064716797, 0.00064716797, ...0.445766
1338ipt_17.tpb_2561.2155400.001443[0.001434624, 0.001435648, 0.001435648, 0.0014...240.000[0.00064716797, 0.00064716797, 0.00064716797, ...0.452094
1411ipt_14.tpb_3521.3164230.011471[0.011004928, 0.0110090235, 0.01118208, 0.0111...281.000[0.022526976, 0.022532096, 0.02254131, 0.02254...1.966970
1414ipt_17.tpb_2561.3163850.011559[0.011276288, 0.011408384, 0.011437056, 0.0114...281.000[0.022526976, 0.022532096, 0.02254131, 0.02254...1.951984
1487ipt_14.tpb_3521.3166860.013171[0.012859392, 0.012862464, 0.01286656, 0.01288...280.811[0.022540288, 0.022542335, 0.022546433, 0.0225...1.714352
1490ipt_17.tpb_2561.3657260.013216[0.013073408, 0.013073408, 0.0130744325, 0.013...280.811[0.022540288, 0.022542335, 0.022546433, 0.0225...1.708442
1563ipt_14.tpb_3521.4658890.017779[0.017767424, 0.017768448, 0.017772544, 0.0177...280.544[0.022519808, 0.022535168, 0.022537217, 0.0225...1.269554
1566ipt_17.tpb_2561.5161410.017783[0.017772544, 0.017774591, 0.017774591, 0.0177...280.544[0.022519808, 0.022535168, 0.022537217, 0.0225...1.269262
1639ipt_14.tpb_3521.8181290.021678[0.021664768, 0.021665793, 0.021666817, 0.0216...280.337[0.022402048, 0.022403073, 0.022406144, 0.0224...1.034908
1642ipt_17.tpb_2561.8164410.021494[0.021485567, 0.021486592, 0.021486592, 0.0214...280.337[0.022402048, 0.022403073, 0.022406144, 0.0224...1.043783
1715ipt_14.tpb_3522.1169000.024031[0.024020992, 0.024022017, 0.024023041, 0.0240...280.201[0.022389758, 0.022416383, 0.022421503, 0.0224...0.935401
1718ipt_17.tpb_2562.0667890.023726[0.023716863, 0.023718912, 0.023718912, 0.0237...280.201[0.022389758, 0.022416383, 0.022421503, 0.0224...0.947432
1791ipt_14.tpb_3521.7169930.022998[0.022985727, 0.022989824, 0.022989824, 0.0229...280.000[0.009991168, 0.009993216, 0.009995264, 0.0099...0.435327
1794ipt_17.tpb_2561.6667280.022663[0.022649856, 0.022649856, 0.022652928, 0.0226...280.000[0.009991168, 0.009993216, 0.009995264, 0.0099...0.441759
\n", "
" ], "text/plain": [ " variant elapsed center \\\n", "43 ipt_14.tpb_352 1.667097 0.000042 \n", "46 ipt_17.tpb_256 1.666077 0.000042 \n", "119 ipt_14.tpb_352 1.666118 0.000043 \n", "122 ipt_17.tpb_256 1.616790 0.000044 \n", "195 ipt_14.tpb_352 1.666390 0.000043 \n", "198 ipt_17.tpb_256 1.666235 0.000044 \n", "271 ipt_14.tpb_352 1.666675 0.000043 \n", "274 ipt_17.tpb_256 1.616154 0.000044 \n", "347 ipt_14.tpb_352 1.666453 0.000043 \n", "350 ipt_17.tpb_256 1.616455 0.000045 \n", "423 ipt_14.tpb_352 1.666218 0.000043 \n", "426 ipt_17.tpb_256 1.616859 0.000043 \n", "499 ipt_14.tpb_352 1.415810 0.000078 \n", "502 ipt_17.tpb_256 1.416047 0.000076 \n", "575 ipt_14.tpb_352 1.416357 0.000086 \n", "578 ipt_17.tpb_256 1.415771 0.000085 \n", "651 ipt_14.tpb_352 1.367251 0.000104 \n", "654 ipt_17.tpb_256 1.366859 0.000102 \n", "727 ipt_14.tpb_352 1.366081 0.000123 \n", "730 ipt_17.tpb_256 1.366657 0.000120 \n", "803 ipt_14.tpb_352 1.365597 0.000127 \n", "806 ipt_17.tpb_256 1.365572 0.000126 \n", "879 ipt_14.tpb_352 1.316359 0.000120 \n", "882 ipt_17.tpb_256 1.316560 0.000119 \n", "955 ipt_14.tpb_352 1.265653 0.000742 \n", "958 ipt_17.tpb_256 1.265422 0.000750 \n", "1031 ipt_14.tpb_352 1.266642 0.000843 \n", "1034 ipt_17.tpb_256 1.265898 0.000853 \n", "1107 ipt_14.tpb_352 1.265562 0.001140 \n", "1110 ipt_17.tpb_256 1.266333 0.001140 \n", "1183 ipt_14.tpb_352 1.265758 0.001382 \n", "1186 ipt_17.tpb_256 1.265617 0.001369 \n", "1259 ipt_14.tpb_352 1.265935 0.001530 \n", "1262 ipt_17.tpb_256 1.265464 0.001510 \n", "1335 ipt_14.tpb_352 1.215466 0.001463 \n", "1338 ipt_17.tpb_256 1.215540 0.001443 \n", "1411 ipt_14.tpb_352 1.316423 0.011471 \n", "1414 ipt_17.tpb_256 1.316385 0.011559 \n", "1487 ipt_14.tpb_352 1.316686 0.013171 \n", "1490 ipt_17.tpb_256 1.365726 0.013216 \n", "1563 ipt_14.tpb_352 1.465889 0.017779 \n", "1566 ipt_17.tpb_256 1.516141 0.017783 \n", "1639 ipt_14.tpb_352 1.818129 0.021678 \n", "1642 ipt_17.tpb_256 1.816441 0.021494 \n", "1715 ipt_14.tpb_352 2.116900 0.024031 \n", "1718 ipt_17.tpb_256 2.066789 0.023726 \n", "1791 ipt_14.tpb_352 1.716993 0.022998 \n", "1794 ipt_17.tpb_256 1.666728 0.022663 \n", "\n", " samples Elements{io}[pow2] \\\n", "43 [3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0... 16 \n", "46 [3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0... 16 \n", "119 [3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0... 16 \n", "122 [3.584e-05, 3.6864003e-05, 3.6864003e-05, 3.68... 16 \n", "195 [3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3... 16 \n", "198 [3.6864003e-05, 3.6864003e-05, 3.6864003e-05, ... 16 \n", "271 [3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3... 16 \n", "274 [3.6864003e-05, 3.6864003e-05, 3.6864003e-05, ... 16 \n", "347 [3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3... 16 \n", "350 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 16 \n", "423 [3.4816e-05, 3.4816e-05, 3.4816e-05, 3.4816e-0... 16 \n", "426 [3.584e-05, 3.584e-05, 3.584e-05, 3.584e-05, 3... 16 \n", "499 [6.9632e-05, 6.9632e-05, 6.9632e-05, 6.9632e-0... 20 \n", "502 [6.8608e-05, 6.8608e-05, 6.9632e-05, 6.9632e-0... 20 \n", "575 [7.8847996e-05, 7.8847996e-05, 7.8847996e-05, ... 20 \n", "578 [7.7824e-05, 7.8847996e-05, 7.8847996e-05, 7.8... 20 \n", "651 [9.6256e-05, 9.7280004e-05, 9.7280004e-05, 9.7... 20 \n", "654 [9.5232004e-05, 9.5232004e-05, 9.5232004e-05, ... 20 \n", "727 [0.000110592, 0.000111616, 0.000111616, 0.0001... 20 \n", "730 [0.000109568, 0.000109568, 0.000110592, 0.0001... 20 \n", "803 [0.000119808006, 0.000120832, 0.000120832, 0.0... 20 \n", "806 [0.000118784, 0.000118784, 0.000118784, 0.0001... 20 \n", "879 [0.00011264, 0.00011264, 0.00011264, 0.0001126... 20 \n", "882 [0.000111616, 0.000111616, 0.000111616, 0.0001... 20 \n", "955 [0.00071168, 0.00071168, 0.000713728, 0.000714... 24 \n", "958 [0.00072499196, 0.000726016, 0.00072704, 0.000... 24 \n", "1031 [0.00082841597, 0.00082944, 0.00082944, 0.0008... 24 \n", "1034 [0.00083968, 0.00084172795, 0.00084172795, 0.0... 24 \n", "1107 [0.001132544, 0.001133568, 0.001133568, 0.0011... 24 \n", "1110 [0.001132544, 0.001132544, 0.001132544, 0.0011... 24 \n", "1183 [0.001375232, 0.001375232, 0.001375232, 0.0013... 24 \n", "1186 [0.001362944, 0.001362944, 0.001362944, 0.0013... 24 \n", "1259 [0.0015226881, 0.0015226881, 0.0015226881, 0.0... 24 \n", "1262 [0.001504256, 0.001504256, 0.001504256, 0.0015... 24 \n", "1335 [0.001455104, 0.001456128, 0.001456128, 0.0014... 24 \n", "1338 [0.001434624, 0.001435648, 0.001435648, 0.0014... 24 \n", "1411 [0.011004928, 0.0110090235, 0.01118208, 0.0111... 28 \n", "1414 [0.011276288, 0.011408384, 0.011437056, 0.0114... 28 \n", "1487 [0.012859392, 0.012862464, 0.01286656, 0.01288... 28 \n", "1490 [0.013073408, 0.013073408, 0.0130744325, 0.013... 28 \n", "1563 [0.017767424, 0.017768448, 0.017772544, 0.0177... 28 \n", "1566 [0.017772544, 0.017774591, 0.017774591, 0.0177... 28 \n", "1639 [0.021664768, 0.021665793, 0.021666817, 0.0216... 28 \n", "1642 [0.021485567, 0.021486592, 0.021486592, 0.0214... 28 \n", "1715 [0.024020992, 0.024022017, 0.024023041, 0.0240... 28 \n", "1718 [0.023716863, 0.023718912, 0.023718912, 0.0237... 28 \n", "1791 [0.022985727, 0.022989824, 0.022989824, 0.0229... 28 \n", "1794 [0.022649856, 0.022649856, 0.022652928, 0.0226... 28 \n", "\n", " Entropy base_samples speedup \n", "43 1.000 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.951220 \n", "46 1.000 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.951220 \n", "119 0.811 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.904762 \n", "122 0.811 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.883721 \n", "195 0.544 [3.6864003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.928571 \n", "198 0.544 [3.6864003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.906977 \n", "271 0.337 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.928571 \n", "274 0.337 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.906977 \n", "347 0.201 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.928571 \n", "350 0.201 [3.7888003e-05, 3.7888003e-05, 3.7888003e-05, ... 0.886364 \n", "423 0.000 [2.4576e-05, 2.4576e-05, 2.4576e-05, 2.4576e-0... 0.595238 \n", "426 0.000 [2.4576e-05, 2.4576e-05, 2.4576e-05, 2.4576e-0... 0.595238 \n", "499 1.000 [0.000110592, 0.000110592, 0.000110592, 0.0001... 1.578947 \n", "502 1.000 [0.000110592, 0.000110592, 0.000110592, 0.0001... 1.621621 \n", "575 0.811 [0.000118784, 0.000119808006, 0.000119808006, ... 1.404762 \n", "578 0.811 [0.000118784, 0.000119808006, 0.000119808006, ... 1.421687 \n", "651 0.544 [0.000119808006, 0.000119808006, 0.00011980800... 1.166667 \n", "654 0.544 [0.000119808006, 0.000119808006, 0.00011980800... 1.190000 \n", "727 0.337 [0.000117760006, 0.000118784, 0.000118784, 0.0... 0.983333 \n", "730 0.337 [0.000117760006, 0.000118784, 0.000118784, 0.0... 1.008547 \n", "803 0.201 [0.000119808006, 0.000119808006, 0.00011980800... 0.959677 \n", "806 0.201 [0.000119808006, 0.000119808006, 0.00011980800... 0.967480 \n", "879 0.000 [4.9152e-05, 4.9152e-05, 4.9152e-05, 4.9152e-0... 0.427350 \n", "882 0.000 [4.9152e-05, 4.9152e-05, 4.9152e-05, 4.9152e-0... 0.431034 \n", "955 1.000 [0.00142848, 0.001429504, 0.001431552, 0.00143... 1.962759 \n", "958 1.000 [0.00142848, 0.001429504, 0.001431552, 0.00143... 1.943989 \n", "1031 0.811 [0.001435648, 0.0014366719, 0.0014366719, 0.00... 1.730255 \n", "1034 0.811 [0.001435648, 0.0014366719, 0.0014366719, 0.00... 1.709484 \n", "1107 0.544 [0.001430528, 0.001432576, 0.001434624, 0.0014... 1.279874 \n", "1110 0.544 [0.001430528, 0.001432576, 0.001434624, 0.0014... 1.279874 \n", "1183 0.337 [0.0014172159, 0.001419264, 0.001421312, 0.001... 1.047407 \n", "1186 0.337 [0.0014172159, 0.001419264, 0.001421312, 0.001... 1.057592 \n", "1259 0.201 [0.001415168, 0.0014172159, 0.001419264, 0.001... 0.947791 \n", "1262 0.201 [0.001415168, 0.0014172159, 0.001419264, 0.001... 0.960000 \n", "1335 0.000 [0.00064716797, 0.00064716797, 0.00064716797, ... 0.445766 \n", "1338 0.000 [0.00064716797, 0.00064716797, 0.00064716797, ... 0.452094 \n", "1411 1.000 [0.022526976, 0.022532096, 0.02254131, 0.02254... 1.966970 \n", "1414 1.000 [0.022526976, 0.022532096, 0.02254131, 0.02254... 1.951984 \n", "1487 0.811 [0.022540288, 0.022542335, 0.022546433, 0.0225... 1.714352 \n", "1490 0.811 [0.022540288, 0.022542335, 0.022546433, 0.0225... 1.708442 \n", "1563 0.544 [0.022519808, 0.022535168, 0.022537217, 0.0225... 1.269554 \n", "1566 0.544 [0.022519808, 0.022535168, 0.022537217, 0.0225... 1.269262 \n", "1639 0.337 [0.022402048, 0.022403073, 0.022406144, 0.0224... 1.034908 \n", "1642 0.337 [0.022402048, 0.022403073, 0.022406144, 0.0224... 1.043783 \n", "1715 0.201 [0.022389758, 0.022416383, 0.022421503, 0.0224... 0.935401 \n", "1718 0.201 [0.022389758, 0.022416383, 0.022421503, 0.0224... 0.947432 \n", "1791 0.000 [0.009991168, 0.009993216, 0.009995264, 0.0099... 0.435327 \n", "1794 0.000 [0.009991168, 0.009993216, 0.009995264, 0.0099... 0.441759 " ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "mean_df = tuning_df.groupby(['variant'])['speedup'].mean()\n", "mean_df = mean_df.reset_index()\n", "mean_df.columns = ['variant', 'mean_speedup']\n", "mean_df.sort_values(by=['mean_speedup'], ascending=False, inplace=True)\n", "top_avg_variants = list(mean_df[mean_df['variant'] != 'base'].head(2)['variant'])\n", "\n", "result = tuning_df[tuning_df['variant'].isin(top_avg_variants)]\n", "result" ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "8837a4869d554085a448621b944d9f3c", "version_major": 2, "version_minor": 0 }, "text/plain": [ "VBox(children=(HBox(children=(Dropdown(description='Elements{io}[pow2]', options=('16', '20', '24', '28'), val…" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "%matplotlib inline\n", "import seaborn as sns\n", "import matplotlib.pyplot as plt\n", "\n", "rt_axes_values = {}\n", "for col in merged_df.columns:\n", " if '{ct}' not in col:\n", " if col not in bench_columns:\n", " rt_axes_values[col] = result[col].unique()\n", "\n", "rt_axes_values['variant'] = result['variant'].unique()\n", "\n", "global rt_axes_chosen_values\n", "rt_axes_chosen_values = {}\n", "\n", "for rt_axis in rt_axes_values:\n", " rt_axes_chosen_values[rt_axis] = rt_axes_values[rt_axis][0]\n", "\n", "\n", "def displot():\n", " display_df_loc = None \n", "\n", " for rt_axis in rt_axes_values:\n", " if display_df_loc is None:\n", " display_df_loc = (result[rt_axis] == rt_axes_chosen_values[rt_axis])\n", " else:\n", " display_df_loc = display_df_loc & (result[rt_axis] == rt_axes_chosen_values[rt_axis])\n", "\n", " display_df = result.loc[display_df_loc].copy()\n", " samples = list(display_df['samples'])[0]\n", " base_samples = list(display_df['base_samples'])[0]\n", "\n", " sns.displot({'base': base_samples, rt_axes_chosen_values['variant']: samples}, aspect=2.)\n", " plt.show()\n", "\n", "\n", "def on_rt_value_change(rt_axis, change):\n", " global rt_axes_chosen_values\n", " if change['type'] == 'change' and change['name'] == 'value':\n", " rt_axes_chosen_values[rt_axis] = change['new']\n", " with plot_output:\n", " clear_output(wait=True)\n", " displot()\n", "\n", "rt_dropdowns = []\n", "for rt_axis in rt_axes_values:\n", " rt_dropdowns.append(create_dropdown_menu(rt_axis, rt_axes_values[rt_axis], functools.partial(on_rt_value_change, rt_axis)))\n", "\n", "plot_output = widgets.Output()\n", "with plot_output:\n", " displot()\n", "\n", "rt_dropdown_row = widgets.HBox(rt_dropdowns)\n", "display(widgets.VBox([rt_dropdown_row, plot_output]))\n" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.10.6" }, "orig_nbformat": 4 }, "nbformat": 4, "nbformat_minor": 2 }