test cuda trtrs

cuda-trtrs.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import unittest\n",
    "import torch\n",
    "torch.set_default_tensor_type(torch.cuda.float64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "class TEST(unittest.TestCase):\n",
    "    def test_trtrs(self):\n",
    "        a = torch.Tensor(((6.80, -2.11, 5.66, 5.97, 8.23),\n",
    "                          (-6.05, -3.30, 5.36, -4.44, 1.08),\n",
    "                          (-0.45, 2.58, -2.70, 0.27, 9.04),\n",
    "                          (8.32, 2.71, 4.35, -7.17, 2.14),\n",
    "                          (-9.67, -5.14, -7.26, 6.08, -6.87))).t()\n",
    "        b = torch.Tensor(((4.02, 6.19, -8.22, -7.57, -3.03),\n",
    "                          (-1.56, 4.00, -8.67, 1.75, 2.86),\n",
    "                          (9.81, -4.09, -4.57, -8.61, 8.99))).t()\n",
    "        \n",
    "        print(a.dtype)\n",
    "\n",
    "        U = torch.triu(a)\n",
    "        L = torch.tril(a)\n",
    "\n",
    "        # solve Ux = b\n",
    "        x = torch.trtrs(b, U)[0]\n",
    "        self.assertLessEqual(b.dist(torch.mm(U, x)), 1e-12)\n",
    "        x = torch.trtrs(b, U, True, False, False)[0]\n",
    "        self.assertLessEqual(b.dist(torch.mm(U, x)), 1e-12)\n",
    "\n",
    "        # solve Lx = b\n",
    "        x = torch.trtrs(b, L, False)[0]\n",
    "        self.assertLessEqual(b.dist(torch.mm(L, x)), 1e-12)\n",
    "        x = torch.trtrs(b, L, False, False, False)[0]\n",
    "        self.assertLessEqual(b.dist(torch.mm(L, x)), 1e-12)\n",
    "\n",
    "        # solve U'x = b\n",
    "        x = torch.trtrs(b, U, True, True)[0]\n",
    "        self.assertLessEqual(b.dist(torch.mm(U.t(), x)), 1e-12)\n",
    "        x = torch.trtrs(b, U, True, True, False)[0]\n",
    "        self.assertLessEqual(b.dist(torch.mm(U.t(), x)), 1e-12)\n",
    "\n",
    "        # solve U'x = b by manual transposition\n",
    "        y = torch.trtrs(b, U.t(), False, False)[0]\n",
    "        self.assertLessEqual(x.dist(y), 1e-12)\n",
    "\n",
    "        # solve L'x = b\n",
    "        x = torch.trtrs(b, L, False, True)[0]\n",
    "        self.assertLessEqual(b.dist(torch.mm(L.t(), x)), 1e-12)\n",
    "        x = torch.trtrs(b, L, False, True, False)[0]\n",
    "        self.assertLessEqual(b.dist(torch.mm(L.t(), x)), 1e-12)\n",
    "\n",
    "        # solve L'x = b by manual transposition\n",
    "        y = torch.trtrs(b, L.t(), True, False)[0]\n",
    "        self.assertLessEqual(x.dist(y), 1e-12)\n",
    "\n",
    "        # test reuse\n",
    "        res1 = torch.trtrs(b, a)[0]\n",
    "        ta = torch.Tensor()\n",
    "        tb = torch.Tensor()\n",
    "        torch.trtrs(b, a, out=(tb, ta))\n",
    "        self.assertEqual(res1.dist(tb), 0)\n",
    "        tb.zero_()\n",
    "        torch.trtrs(b, a, out=(tb, ta))\n",
    "        self.assertEqual(res1.dist(tb), 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.cuda.float64\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      ".\n",
      "----------------------------------------------------------------------\n",
      "Ran 1 test in 1.827s\n",
      "\n",
      "OK\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<unittest.main.TestProgram at 0x7f230e5e8f98>"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unittest.main(argv=['first-arg-is-ignored'], exit=False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python (pyro)",
   "language": "python",
   "name": "pyro"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}