diff --git a/examples_cuda/mandelbrot_tasks3d/mandelbrot_tasks3d.cpp b/examples_cuda/mandelbrot_tasks3d/mandelbrot_tasks3d.cpp
index f6741cbd..6cfabf6f 100644
--- a/examples_cuda/mandelbrot_tasks3d/mandelbrot_tasks3d.cpp
+++ b/examples_cuda/mandelbrot_tasks3d/mandelbrot_tasks3d.cpp
@@ -131,7 +131,7 @@ int main(int argc, char *argv[]) {
         minISPC = std::min(minISPC, dt);
     }
 
-    printf("[mandelbrot ispc+tasks]:\t[%.3f] million cycles\n", minISPC);
+    fprintf(stderr,"[mandelbrot ispc+tasks]:\t[%.3f] million cycles\n", minISPC);
     writePPM(buf, width, height, "mandelbrot-ispc.ppm");
 
 
diff --git a/examples_cuda/mandelbrot_tasks3d/mandelbrot_tasks3d.ispc b/examples_cuda/mandelbrot_tasks3d/mandelbrot_tasks3d.ispc
index aeff4773..83b2f27f 100644
--- a/examples_cuda/mandelbrot_tasks3d/mandelbrot_tasks3d.ispc
+++ b/examples_cuda/mandelbrot_tasks3d/mandelbrot_tasks3d.ispc
@@ -96,8 +96,8 @@ mandelbrot_ispc(uniform float x0, uniform float y0,
                 uniform int maxIterations, uniform int output[]) {
     uniform float dx = (x1 - x0) / width;
     uniform float dy = (y1 - y0) / height;
-    const uniform int xspan = 32;  /* make sure it is big enough to avoid false-sharing */
-    const uniform int yspan = 4; 
+    const uniform int xspan = 64;  /* make sure it is big enough to avoid false-sharing */
+    const uniform int yspan = 8; 
 
 
 #if 1