use ARMCI not MA mem in GA_Transpose (#211)

GA_Transpose uses ga_malloc, which defaults to using
MA memory.  this means that GA requires MA to be initialized.
this changes GA_Transpose to use ARMCI memory.

if ARMCI_Malloc_local is much slower than ga_malloc, this might
be noticeable (because the block is called in the loop)
but that seems unlikely at this point.

it is trivial to revert to the old implementation by undefining
the macro (GA_TRANSPOSE_USE_ARMCI_MEM) that enables this.

fix issue #163

Signed-off-by: Jeff Hammond <jeff.science@gmail.com>

Signed-off-by: Jeff Hammond <jeff.science@gmail.com>

global/src/global.nalg.c

@@ -25,6 +25,10 @@
 #include "ga-wapi.h"
 #include "base.h"
 
+// workaround https://github.com/GlobalArrays/ga/issues/163
+// by using ARMCI_Malloc_local instead of ga_malloc -> MA
+#define GA_TRANSPOSE_USE_ARMCI_MEM 1
+
 #ifdef MSG_COMMS_MPI
 extern ARMCI_Group* ga_get_armci_group_(int);
 #endif
@@ -981,7 +985,14 @@ _iterator_hdl hdl;
       int i, size=GAsizeofM(atype);
 
       nelem = (hi[0]-lo[0]+1)*(hi[1]-lo[1]+1);
+#ifdef GA_TRANSPOSE_USE_ARMCI_MEM
+{
+      size_t bytes = nelem * size;
+      ptr_tmp = ARMCI_Malloc_local(bytes);
+}
+#else
       ptr_tmp = (char *) ga_malloc(nelem, atype, "transpose_tmp");
+#endif
 
       nrow   = hi[0] -lo[0]+1;
       ncol   = hi[1] -lo[1]+1; 
@@ -995,7 +1006,11 @@ _iterator_hdl hdl;
         ptr_a += ld[0]*size;
       }
       pnga_put(g_b, lob, hib, ptr_tmp ,&ncol);
+#ifdef GA_TRANSPOSE_USE_ARMCI_MEM
+      ARMCI_Free_local(ptr_tmp);
+#else
       ga_free(ptr_tmp);
+#endif
     }
 #else
     num_blocks_a = pnga_total_blocks(g_a);