From 69f2911e04ffb1b19eef1fafb8c040af271f656e Mon Sep 17 00:00:00 2001
From: Tor Aamodt <aamodt@ece.ubc.ca>
Date: Thu, 15 Jul 2010 18:09:46 -0800
Subject: creating branch for adding support for CUDA 3.x and Fermi

[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 6829]
---
 benchmarks/CUDA/LPS/laplace3d_gold.cpp | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)
 create mode 100644 benchmarks/CUDA/LPS/laplace3d_gold.cpp

(limited to 'benchmarks/CUDA/LPS/laplace3d_gold.cpp')

diff --git a/benchmarks/CUDA/LPS/laplace3d_gold.cpp b/benchmarks/CUDA/LPS/laplace3d_gold.cpp
new file mode 100644
index 0000000..966c4aa
--- /dev/null
+++ b/benchmarks/CUDA/LPS/laplace3d_gold.cpp
@@ -0,0 +1,27 @@
+
+////////////////////////////////////////////////////////////////////////////////
+// export C interface
+extern "C" 
+void Gold_laplace3d(int NX, int NY, int NZ, float* u1, float* u2) 
+{
+  int   i, j, k, ind;
+  float sixth=1.0f/6.0f;  // predefining this improves performance more than 10%
+
+  for (k=0; k<NZ; k++) {
+    for (j=0; j<NY; j++) {
+      for (i=0; i<NX; i++) {   // i loop innermost for sequential memory access
+	ind = i + j*NX + k*NX*NY;
+
+        if (i==0 || i==NX-1 || j==0 || j==NY-1|| k==0 || k==NZ-1) {
+          u2[ind] = u1[ind];          // Dirichlet b.c.'s
+        }
+        else {
+          u2[ind] = ( u1[ind-1    ] + u1[ind+1    ]
+                    + u1[ind-NX   ] + u1[ind+NX   ]
+                    + u1[ind-NX*NY] + u1[ind+NX*NY] ) * sixth;
+        }
+      }
+    }
+  }
+}
+
-- 
cgit v1.3