notebooks/run_job_on_NERSC.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Submit job on Nersc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook is used for submitting jobs on NERSC.\n",
    "\n",
    "[**Common slurm commands**](https://github.com/NERSC/slurm-magic), such as **_%sbatch, %squeue, %scancel, %salloc, %srun, %scontrol_**\n",
    "\n",
    "Any command that works at the command-line can be used in Jupyter Notebook by prefixing it with the **!** character. For example, the **ls, pwd, and echo** commands can be run as follows: **!ls, !pwd, !echo** ([link](https://jakevdp.github.io/PythonDataScienceHandbook/01.05-ipython-and-shell-commands.html))\n",
    "\n",
    "Besides **%cd**, other available shell-like magic functions are **%cat, %cp, %env, %ls, %man, %mkdir, %more, %mv, %pwd, %rm, and %rmdir**, any of which can be used without the % sign if `automagic` is on."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-12T20:59:20.425706Z",
     "start_time": "2018-07-12T20:59:20.416768Z"
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import subprocess\n",
    "import pandas as pd\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext slurm_magic # must load this first"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Sample batch script"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**on Edison**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```bash\n",
    "#!/bin/bash -l\n",
    "\n",
    "#SBATCH -A m1800\n",
    "#SBATCH -N 171\n",
    "#SBATCH -t 48:00:00\n",
    "#SBATCH -L SCRATCH  \n",
    "#SBATCH -J 100x2_1w\n",
    "#SBATCH --qos regular\n",
    "#SBATCH --mail-type ALL\n",
    "#SBATCH --mail-user pin.shuai@pnnl.gov\n",
    "\n",
    "cd $SLURM_SUBMIT_DIR\n",
    "\n",
    "srun -n 4096 /global/project/projectdirs/pflotran/pflotran-edison/src/pflotran/pflotran -pflotranin pflotran_100x100x2_1w_bc.in\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**on Cori**\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```bash\n",
    "#!/bin/bash -l\n",
    "\n",
    "#SBATCH -A m1800\n",
    "#SBATCH -N 171\n",
    "#SBATCH -t 48:00:00\n",
    "#SBATCH -L SCRATCH  \n",
    "#SBATCH -J 100x2_1w\n",
    "#SBATCH --qos regular\n",
    "#SBATCH --mail-type ALL\n",
    "#SBATCH --mail-user pin.shuai@pnnl.gov\n",
    "#SBATCH -C haswell\n",
    "\n",
    "cd $SLURM_SUBMIT_DIR\n",
    "\n",
    "srun -n 4096 /global/project/projectdirs/pflotran/pflotran-cori/src/pflotran/pflotran -pflotranin pflotran_100x100x2_1w_bc.in\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## execute batch job"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "import os\n",
    "os.chdir(\"/global/cscratch1/sd/pshuai/TH_cp_high\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Submitted batch job 21376851\\n'"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%sbatch\n",
    "#!/bin/bash -l\n",
    "\n",
    "#SBATCH -A m1800\n",
    "#SBATCH -N 1\n",
    "#SBATCH -t 00:30:00\n",
    "#SBATCH -L SCRATCH  \n",
    "#SBATCH -J TH-cp-high-test\n",
    "#SBATCH --qos debug\n",
    "#SBATCH --mail-type ALL\n",
    "#SBATCH --mail-user pin.shuai@pnnl.gov\n",
    "#SBATCH -C haswell\n",
    "\n",
    "cd $SLURM_SUBMIT_DIR\n",
    "\n",
    "srun -n 2 /global/project/projectdirs/pflotran/pflotran-cori/src/pflotran/pflotran -pflotranin TH_cp_high.in"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "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>JOBID</th>\n",
       "      <th>PARTITION</th>\n",
       "      <th>NAME</th>\n",
       "      <th>USER</th>\n",
       "      <th>ST</th>\n",
       "      <th>TIME</th>\n",
       "      <th>NODES</th>\n",
       "      <th>NODELIST(REASON)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>21250229</td>\n",
       "      <td>regular</td>\n",
       "      <td>TH-cont</td>\n",
       "      <td>pshuai</td>\n",
       "      <td>PD</td>\n",
       "      <td>0:00</td>\n",
       "      <td>128</td>\n",
       "      <td>(Priority)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>21250633</td>\n",
       "      <td>regular</td>\n",
       "      <td>TH-smoot</td>\n",
       "      <td>pshuai</td>\n",
       "      <td>PD</td>\n",
       "      <td>0:00</td>\n",
       "      <td>128</td>\n",
       "      <td>(Priority)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>21217356</td>\n",
       "      <td>regular</td>\n",
       "      <td>TH-cp-hi</td>\n",
       "      <td>pshuai</td>\n",
       "      <td>PD</td>\n",
       "      <td>0:00</td>\n",
       "      <td>128</td>\n",
       "      <td>(JobHeldUser)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>21217399</td>\n",
       "      <td>regular</td>\n",
       "      <td>TH-cp-lo</td>\n",
       "      <td>pshuai</td>\n",
       "      <td>PD</td>\n",
       "      <td>0:00</td>\n",
       "      <td>128</td>\n",
       "      <td>(JobHeldUser)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      JOBID PARTITION      NAME    USER  ST  TIME  NODES NODELIST(REASON)\n",
       "0  21250229   regular   TH-cont  pshuai  PD  0:00    128       (Priority)\n",
       "1  21250633   regular  TH-smoot  pshuai  PD  0:00    128       (Priority)\n",
       "2  21217356   regular  TH-cp-hi  pshuai  PD  0:00    128    (JobHeldUser)\n",
       "3  21217399   regular  TH-cp-lo  pshuai  PD  0:00    128    (JobHeldUser)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%squeue -u pshuai"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "''"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%scancel 13162214"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## execute interactive job"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('',\n",
       " 'Launched in background. Redirecting stdin to /dev/null\\nsalloc: Pending job allocation 12123294\\nsalloc: job 12123294 queued and waiting for resources\\nsalloc: job 12123294 has been allocated resources\\nsalloc: Granted job allocation 12123294\\nsrun: error: ioctl(TIOCGWINSZ): Inappropriate ioctl for device\\nsrun: error: Not using a pseudo-terminal, disregarding --pty option\\nsalloc: Relinquishing job allocation 12123294\\n')"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%salloc -N 1 -q debug -C haswell -t 00:30:00 -L SCRATCH"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('',\n",
       " 'Launched in background. Redirecting stdin to /dev/null\\nsrun: error: No hardware architecture specified (-C)!\\nsrun: error: Unable to allocate resources: Unspecified error\\n')"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%srun -n 32 ../pflotran-cori-flux -pflotranin pflotran.in"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "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>JOBID</th>\n",
       "      <th>PARTITION</th>\n",
       "      <th>NAME</th>\n",
       "      <th>USER</th>\n",
       "      <th>ST</th>\n",
       "      <th>TIME</th>\n",
       "      <th>NODES</th>\n",
       "      <th>NODELIST(REASON)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [JOBID, PARTITION, NAME, USER, ST, TIME, NODES, NODELIST(REASON)]\n",
       "Index: []"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%squeue -u pshuai"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "hello world! \n"
     ]
    }
   ],
   "source": [
    "!sh /global/cscratch1/sd/pshuai/test/test.sh"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#!/bin/bash -l\n",
      "\n",
      " printf \"hello world! \\n\"\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "!cat /global/cscratch1/sd/pshuai/test/test.sh"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# scaling test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Scaling test is run as follows:\n",
    "* increase the # of cores while specifying `MAX_STEPS = 50` in PFLOTRAN input deck. \n",
    "* run interactive job using sample code below\n",
    "* plot `FLOW TS BE SNES time` and `TRANSPORT TS BE SNES time` over the # of cores"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```bash\n",
    "switch ( \"$NERSC_HOST\" )\n",
    "# used 'srun -n # ./a.out'\n",
    "case cori:\n",
    "  alias q1    'salloc -N 1 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 1 ../pflotran-cori -pflotranin pflotran_100x100x05_6h_bc.in\n",
    "  alias q2    'salloc -N 1 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "  alias q4    'salloc -N 1 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "  alias q8    'salloc -N 1 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "  alias q16    'salloc -N 1 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "  alias q32    'salloc -N 1 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 32 ../pflotran-cori -pflotranin pflotran_100x100x05_6h_bc.in\n",
    "  alias q64    'salloc -N 2 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 64 ../pflotran-cori -pflotranin pflotran_100x100x05_6h_bc.in\n",
    "  alias q128    'salloc -N 4 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "  alias q256    'salloc -N 8 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "  alias q512    'salloc -N 16 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "  alias q1024    'salloc -N 32 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "  alias q2048    'salloc -N 64 -q debug -C haswell -t 00:30:00 -L SCRATCH'\n",
    "  breaksw\n",
    "    \n",
    "case edison:\n",
    "  alias q1    'salloc -N 1 -q debug -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 1 ../pflotran-edison -pflotranin pflotran_100x100x2_6h_bc.in\n",
    "  alias q4    'salloc -N 1 -q debug -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 4 ../pflotran-edison -pflotranin pflotran_200x200x2_6h_bc_new_iniH.in\n",
    "  alias q8    'salloc -N 1 -q debug -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 8 ../pflotran-edison -pflotranin pflotran_1000x1000x5_6h_bc.in\n",
    "  alias q16   'salloc -N 1 -q debug -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 16 ../pflotran-edison -pflotranin pflotran_100x100x05_6h_bc.in\n",
    "  alias q32   'salloc -N 2 -q debug -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 32 ../pflotran-edison -pflotranin pflotran_1000x1000x5_6h_bc.in\n",
    "  alias q64   'salloc -N 3 -q debug -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 64 ../pflotran-edison -pflotranin pflotran_1000x1000x5_6h_bc.in\n",
    "  alias q128  'salloc -N 6 -q debug -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 128 ../pflotran-edison -pflotranin pflotran_200x200x2_6h_bc_new_iniH.in\n",
    "  alias q256  'salloc -N 11 -q debug -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 256 ../pflotran-edison -pflotranin pflotran_1000x1000x5_6h_bc.in\n",
    "  alias q512  'salloc -N 22 -q debug -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 512 ../pflotran-edison -pflotranin pflotran_1000x1000x5_6h_bc.in\n",
    "  alias q1024  'salloc -N 44 -q debug -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 1024 ../pflotran-edison -pflotranin pflotran_1000x1000x5_6h_bc.in\n",
    "  alias q2048 'salloc -N 86 -q debug -t 00:30:00 -L SCRATCH'\n",
    "                srun -n 2048 ../pflotran-edison -pflotranin pflotran_1000x1000x5_6h_bc.in\n",
    "  alias q4096 'salloc -N 171 -q debug -t 00:30:00 -L SCRATCH'\n",
    "  alias q8192 'salloc -N 342 -q debug -t 00:30:00 -L SCRATCH'\n",
    "  breaksw\n",
    "endsw\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## scaling test for HFR_model_1000x1000x5_6h_bc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Scaling test results for model resolution of 1000x1000x5 m with a total of 144,000 grids and `Max step = 500`. **\n",
    "\n",
    "|# of cores|Flow solution time (sec)|Transport solution time (sec)|Wall clock time (sec)|\n",
    "|----------|------------------|-----------------------|---------------|\n",
    "|4|556|98.9|802|\n",
    "|8|720.3|67.9|900.3|\n",
    "|16|411.3|38.5|556.4|\n",
    "|32|223.1|17.3|324.7|\n",
    "|64|165.9|10.4|257.4|\n",
    "|128|133.9|5.1|215.1|\n",
    "|**256**|**90.6**|**2.4**|**168.9**|\n",
    "|512|136.5|1.5|214.3|\n",
    "|1024|112.3|1.6|200.4|\n",
    "|2048|408.1|2.4|502|"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "plot simulation time vs. # of cores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-13T23:34:21.861217Z",
     "start_time": "2018-07-13T23:34:21.566557Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x11d985438>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11d807128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "d = {'cores' : [4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048],\n",
    "    'flow_time': [556, 720.3, 411.3, 223.1, 165.9, 133.9, 90.6, 136.5, 112.3, 408.1],\n",
    "    'trans_time': [98.9, 67.9, 38.5, 17.3, 10.4, 5.1, 2.4, 1.5, 1.6, 2.4],\n",
    "    'wall_time' : [802, 900.3, 556.4, 324.7, 257.4, 215.1, 168.9, 214.3, 200.4, 502]}\n",
    "df = pd.DataFrame(data= d)\n",
    "\n",
    "matplotlib.style.use('ggplot')\n",
    "\n",
    "fig, ax = plt.subplots(1, 1)\n",
    "df.plot(x='cores', y= ['flow_time', 'trans_time', 'wall_time'], ax=ax, figsize=(8, 8), logx=True)\n",
    "ax.set_ylabel('Time (sec)')\n",
    "ax.set_title('scaling test for model 1000x1000x5 m')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## scaling test for 200x200x2_head_bc_new_iniH model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`MAX_STEPS 100` for testing model of 200x200x2m with 9 million grids\n",
    "\n",
    "|dof per core|# of cores|Flow solution time (sec)|Transport solution time (sec)|Wall clock time (sec)|Total Wall clock time (hr)$^*$|\n",
    "|---|----------|------------------|-----------------------|---------------|---|\n",
    "|70312|128|841|37.1|1260|\n",
    "|35156|256|359.4|22.4|733.6|\n",
    "|17578|512|220.3|12|585.4|\n",
    "|8789|1024|86.9|6|445|\n",
    "|**4394**|**2048**|**47.8**|**2.5**|**520**|**67.25**|\n",
    "|2197|4096|42.1|1.6|465|\n",
    "\n",
    "*total wall clock time is the time required to finish one simulation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-13T23:36:07.847105Z",
     "start_time": "2018-07-13T23:36:07.640071Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x11de35d30>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11bde1f28>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "d = {'cores' : [ 128, 256, 512, 1024, 2048, 4096],\n",
    "    'flow_time': [841, 359, 220, 86.9, 47, 42],\n",
    "    'trans_time': [37.1, 22.4, 12, 6, 2.5, 1.6],\n",
    "    'wall_time' : [1260, 733, 585, 445, 520, 465]}\n",
    "df = pd.DataFrame(data= d)\n",
    "\n",
    "matplotlib.style.use('ggplot')\n",
    "\n",
    "fig, ax = plt.subplots(1, 1)\n",
    "df.plot(x='cores', y= ['flow_time', 'trans_time', 'wall_time'], ax=ax, figsize=(8, 8), logx=True)\n",
    "ax.set_ylabel('Time (sec)')\n",
    "ax.set_title('scaling test for model 200x200x2 m')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## scaling test for 100x100x2 m"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-20T23:51:04.957959Z",
     "start_time": "2018-07-20T23:51:04.954010Z"
    }
   },
   "source": [
    "`MAX_STEPS 100` for testing model of 100x100x2 m with 18 million grids\n",
    "\n",
    "|dof per core|# of cores|Flow solution time (sec)|Transport solution time (sec)|Wall clock time (sec)|Total wall time (hr)|\n",
    "|---|----------|------------------|-----------------------|---------------|---|\n",
    "|17578|1024|206.2|21.2|668|\n",
    "|8789|2048|104.1|11.2|474.3|\n",
    "|**4394**|**4096**|**59.4**|**4.6**|**450.5**|**109.2**|\n",
    "|2197|8192|68.9|3.6|548.1|"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## scaling test for 100x100x1 m\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`MAX_STEPS 100` for testing model of 100x100x1 m with 36 million grids\n",
    "\n",
    "|dof per core|# of cores|Flow solution time (sec)|Transport solution time (sec)|Wall clock time (sec)|\n",
    "|---|----------|------------------|-----------------------|---------------|\n",
    "|35156|1024|729|43|1130|\n",
    "|17578|2048| 315 | 21 | 700 |\n",
    "|8789|4096| 189 | 10 | 645 |\n",
    "|4394|8192|115|5.6|649|"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## scaling test for 100x100x0.5 m "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`MAX_STEPS 100` for testing model of 100x100x0.5 m with 72 million grids\n",
    "\n",
    "|dof per core|# of cores|Flow solution time (sec)|Transport solution time (sec)|Wall clock time (sec)|\n",
    "|---|----------|------------------|-----------------------|---------------|\n",
    "|35156|2048|765.7|42.7|1228.2|\n",
    "|17578|4096|327.1|19.8|787.5|\n",
    "|8789|8192|189.7|10.2|739|\n",
    "|4394|16384|"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# job logs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "|Completed jobs|Description|Total run time (hr)\n",
    "|:---|:---:|:---:|\n",
    "|HFR_model_200x200x2_6h_bc|200x200x2-m grid, **6h-river-bc**, no-flow-inland-bc, 200-m-thick|\n",
    "HFR_model_200x200x2_1w_bc|200x200x2-m grid, **1w-river-bc**, no-flow-inland-bc, 200-m-thick\n",
    "HFR_model_200x200x5_6h_1w_bc|200x200x5-m grid, **both 6h- and 1w-river-bc**, no-flow-inland-bc, 200-m-thick\n",
    "HFR_model_100x100x5_6h_bc|100x100x5-m grid, 6h-river-bc, no-flow-inland-bc, 200-m-thick\n",
    "HFR_model_1000x1000x5_6h_bc|1000x1000x5-m grid, 6h-river-bc, no-flow-inland-bc, 200-m-thick\n",
    "HFR_model_200x200x2_head_bc_new_iniH|200x200x2-m grid, 6h-river-bc, **flow-inland-bc**, 200-m-thick\n",
    "HFR_model_100x100x2|100x100x2-m grid, 6h-river-bc, no-flow-inland-bc, **100-m-thick**, iniH from MASS1 pts\n",
    "HFR_model_200x200x2_6h_bc_new_iniH|iniH from MASS1 pts\n",
    "HFR_model_100x100x2_new_iniH|**iniH from 200m-model output**, has basalt unit, 6h-river-bc\n",
    "HFR_model_100x100x2_1w_bc|iniH from 200m-model output, has basalt unit, 1w-river-bc\n",
    "HFR_model_100x100x2_cyclic|use restart file from **HFR_model_100x100x2_new_iniH** and run it for another 8 yrs| 105 \n",
    "HFR_model_100x100x2_1w_cyclic|use restart file from **HFR_model_100x100x2_1w_bc** and run it for another 8 yrs\n",
    "\n",
    "\\* indicates model in progress"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.4"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}