OpenMP for gravmag.polyprism

Benchmarks are very good! Sometimes more than 2x speedup over numpy version. Moved the vertices loop out of kernels to avoid segfaults. Had to change the way PolygonalPrism makes .x and .y attributes so that they are C contiguous arrays.
fatiando · Jul 3, 2014 · 5db1796 · 5db1796
1 parent a9913de
commit 5db1796
Show file tree

Hide file tree

Showing 5 changed files with 8,338 additions and 5,690 deletions.
diff --git a/benchmarks/gravmag_polyprism.ipynb b/benchmarks/gravmag_polyprism.ipynb
@@ -1,7 +1,7 @@
 {
  "metadata": {
   "name": "",
-  "signature": "sha256:35610cfe33d9fa8c60169f57cdf6a0a0eee501fc57a6a40ca8247020de5b93d1"
+  "signature": "sha256:1d07d7eba0f3040aaf4f3caad22fa3ef3411fcef439131fa90860f23caef4c7c"
  },
  "nbformat": 3,
  "nbformat_minor": 0,
@@ -52,7 +52,7 @@
        "output_type": "stream",
        "stream": "stdout",
        "text": [
-        "1 loops, best of 3: 1.6 s per loop\n"
+        "1 loops, best of 3: 1.59 s per loop\n"
        ]
       }
      ],
@@ -71,7 +71,7 @@
        "output_type": "stream",
        "stream": "stdout",
        "text": [
-        "1 loops, best of 3: 1.51 s per loop\n"
+        "1 loops, best of 3: 693 ms per loop\n"
        ]
       }
      ],
@@ -90,7 +90,7 @@
        "output_type": "stream",
        "stream": "stdout",
        "text": [
-        "1 loops, best of 3: 837 ms per loop\n"
+        "1 loops, best of 3: 846 ms per loop\n"
        ]
       }
      ],
@@ -109,7 +109,7 @@
        "output_type": "stream",
        "stream": "stdout",
        "text": [
-        "1 loops, best of 3: 613 ms per loop\n"
+        "1 loops, best of 3: 301 ms per loop\n"
        ]
       }
      ],
@@ -128,11 +128,11 @@
        "output_type": "stream",
        "stream": "stdout",
        "text": [
-        "1 loops, best of 3: 7.98 s per loop\n"
+        "1 loops, best of 3: 8.24 s per loop\n"
        ]
       }
      ],
-     "prompt_number": 4
+     "prompt_number": 7
     },
     {
      "cell_type": "code",
@@ -147,19 +147,11 @@
        "output_type": "stream",
        "stream": "stdout",
        "text": [
-        "1 loops, best of 3: 6.25 s per loop\n"
+        "1 loops, best of 3: 2.64 s per loop\n"
        ]
       }
      ],
-     "prompt_number": 3
-    },
-    {
-     "cell_type": "code",
-     "collapsed": false,
-     "input": [],
-     "language": "python",
-     "metadata": {},
-     "outputs": []
+     "prompt_number": 8
     }
    ],
    "metadata": {}