abafar-dask-demo-20210702.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3a633470-3da0-477a-b83d-5fed6c797657",
   "metadata": {},
   "source": [
    "# Launching Dask clusters\n",
    "\n",
    "This notebook illustrates how to launch a Dask cluster on Abafar."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "9e9ef947-c9b1-4706-8d9e-f3226e9a583f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# gain access to the gateway\n",
    "from dask_gateway import Gateway\n",
    "gateway = Gateway()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "a4f6cdad-1926-4bfc-9920-a18bf23a702e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "579bf0d8253544c9a73fd2f33a5cc87a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(HTML(value='<h2>Cluster Options</h2>'), GridBox(children=(HTML(value=\"<p style='font-weight: bo…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# list available options\n",
    "options = gateway.cluster_options()\n",
    "options"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cf1b377b-93d6-449c-9731-444520c6b155",
   "metadata": {},
   "source": [
    "**N.B., when selecting the environment, make sure it matches the environment you're using to run this notebook!**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "0fca85b8-d196-48ee-84ef-77442d3df4e5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1ae10961e40449e5827342ee41cca47c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(HTML(value='<h2>GatewayCluster</h2>'), HBox(children=(HTML(value='\\n<div>\\n<style scoped>\\n    …"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# create a cluster with the options selected via the widgets above\n",
    "# ... may take a few minutes to provision a VM for the scheduler\n",
    "cluster = gateway.new_cluster(options)\n",
    "cluster"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "f991632c-560d-49a5-a6d5-396159f61809",
   "metadata": {},
   "outputs": [],
   "source": [
    "# alternatively, create a cluster by specifying environment and profile as arguments\n",
    "# cluster = gateway.new_cluster(environment=\"binder-v3.0.0\", profile=\"High Memory Worker\")\n",
    "# cluster"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "697a5cd3-34ec-41bc-828c-00aad0894f37",
   "metadata": {},
   "outputs": [],
   "source": [
    "# you can scale the cluster using the widget above, or via a manual API call\n",
    "# ... may take a few minutes to provision VMs for the workers\n",
    "cluster.scale(8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "a4e63eaf-ffbd-4c69-b32e-b179b7fdf95d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# get a client - this tells Dask to use the cluster you've just created\n",
    "client = cluster.get_client()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "b513edaa-3fef-48e0-8333-74732e51f1b4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1.00095105, 0.99624698, 1.00024171, ..., 1.00190085, 1.00039146,\n",
       "       0.99803349])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# now we can use Dask!\n",
    "import dask.array as da\n",
    "x = da.random.random((50000, 50000), chunks=(1000, 1000))\n",
    "y = x + x.T\n",
    "z = y[::2, 5000:].mean(axis=1)\n",
    "z.compute()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e23c294f-b6b7-4dff-95a0-c83a23bc02d3",
   "metadata": {},
   "source": [
    "**When you've finished using the cluster, scale it back down to zero.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "87ef4049-2c96-4256-bbb2-494dc84fa3d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "cluster.scale(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "de4c410b-dcc8-47ad-8b6b-2c05b58f22a7",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:abafar-20210701a]",
   "language": "python",
   "name": "conda-env-abafar-20210701a-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}