define NDEBUG in field_descriptor.cpp

src/field_descriptor.cpp

@@ -18,6 +18,8 @@
 *
 ************************************************************************/
 
+#define NDEBUG
+
 #include <stdlib.h>
 #include <algorithm>
 #include <iostream>

src/slab_field_particles.cpp

@@ -170,18 +170,15 @@ void slab_field_particles<rnumber>::rFFTW_to_buffered(rnumber *src, rnumber *dst
     std::copy(src,
               src + this->fs->rd->local_size,
               dst + bsize);
-    //DEBUG_MSG("send tag is %d\n", MOD(this->fs->rd->starts[0]-1, this->fs->rd->sizes[0]));
-    //DEBUG_MSG("recv tag is %d\n", MOD(this->fs->rd->starts[0]+this->fs->rd->subsizes[0]-1, this->fs->rd->sizes[0]));
-    //DEBUG_MSG("destination cpu is %d\n",
-    //        this->fs->rd->rank[MOD(this->fs->rd->starts[0]-1, this->fs->rd->sizes[0])]);
-    //DEBUG_MSG("source cpu is %d\n",
-    //        this->fs->rd->rank[MOD(this->fs->rd->starts[0]+this->fs->rd->subsizes[0], this->fs->rd->sizes[0])]
-    //        );
     /* take care of buffer regions.
      * I could make the code use blocking sends and receives, but it seems cleaner this way.
      * (alternative is to have a couple of loops).
      * */
     // 1. send lower slices
+    //DEBUG_MSG(
+    //        "destination rank is %d, message is %d\n",
+    //        this->fs->rd->rank[MOD(this->fs->rd->starts[0]+this->fs->rd->subsizes[0], this->fs->rd->sizes[0])],
+    //        MOD(this->fs->rd->starts[0]-1, this->fs->rd->sizes[0]));
     MPI_Isend(
             (void*)(src),
             bsize,
@@ -191,6 +188,10 @@ void slab_field_particles<rnumber>::rFFTW_to_buffered(rnumber *src, rnumber *dst
             this->fs->rd->comm,
             mpirequest);
     // 2. receive higher slices
+    //DEBUG_MSG(
+    //        "source rank is %d, message is %d\n",
+    //        this->fs->rd->rank[MOD(this->fs->rd->starts[0]-1, this->fs->rd->sizes[0])],
+    //        MOD(this->fs->rd->starts[0]+this->fs->rd->subsizes[0]-1, this->fs->rd->sizes[0]));
     MPI_Irecv(
             (void*)(dst + bsize + this->fs->rd->local_size),
             bsize,
@@ -204,7 +205,7 @@ void slab_field_particles<rnumber>::rFFTW_to_buffered(rnumber *src, rnumber *dst
     //DEBUG_MSG(
     //        "destination rank is %d, message is %d\n",
     //        this->fs->rd->rank[MOD(this->fs->rd->starts[0]+this->fs->rd->subsizes[0], this->fs->rd->sizes[0])],
-    //        this->fs->rd->starts[0]+this->fs->rd->subsizes[0]);
+    //        MOD(this->fs->rd->starts[0]+this->fs->rd->subsizes[0], this->fs->rd->sizes[0]));
     MPI_Isend(
             (void*)(src + this->fs->rd->local_size - bsize),
             bsize,

test.py

@@ -40,7 +40,8 @@ def main(opt):
     c.parameters['niter_todo'] = opt.nsteps
     c.parameters['famplitude'] = 0.0
     if opt.run:
-        subprocess.call(['rm', 'test1_*', 'test2_*'])
+        subprocess.call(['rm test1_*'], shell = True)
+        subprocess.call(['rm test2_*'], shell = True)
         #subprocess.call(['make', 'clean'])
         c.execute(ncpu = opt.ncpu)
     dtype = pickle.load(open(c.name + '_dtype.pickle'))