Use cntl trees that block in n dimension.

Details:
- Updated _cntl.c files for each level-3 operation to induce blocked
  algorithms that first paritition in the n dimension with a blocksize
  of NC. Typically this is not an issue since only very large problems
  exceed that of NC. But developers often run very large problems, and
  so this extra blocking should be the default.
- Removed some recently introduced but now unused macros from
  bli_param_macro_defs.h.

frame/3/gemm/bli_gemm_cntl.c

@@ -219,8 +219,8 @@ void bli_gemm_cntl_init()
 	                          NULL );
 
 	// Alias the "master" gemm control tree to a shorter name.
-	gemm_cntl = gemm_cntl_mm_op;
-
+	//gemm_cntl = gemm_cntl_mm_op;
+	gemm_cntl = gemm_cntl_vl_mm;
 }
 
 void bli_gemm_cntl_finalize()

frame/3/hemm/bli_hemm_cntl.c

@@ -219,8 +219,8 @@ void bli_hemm_cntl_init()
 	                          NULL );
 
 	// Alias the "master" hemm control tree to a shorter name.
-	hemm_cntl = hemm_cntl_mm_op;
-
+	//hemm_cntl = hemm_cntl_mm_op;
+	hemm_cntl = hemm_cntl_vl_mm;
 }
 
 void bli_hemm_cntl_finalize()

frame/3/her2k/bli_her2k_cntl.c

@@ -222,8 +222,8 @@ void bli_her2k_cntl_init()
 	                           NULL );
 
 	// Alias the "master" her2k control tree to a shorter name.
-	her2k_cntl = her2k_cntl_mm_op;
-
+	//her2k_cntl = her2k_cntl_mm_op;
+	her2k_cntl = her2k_cntl_vl_mm;
 }
 
 void bli_her2k_cntl_finalize()

frame/3/herk/bli_herk_cntl.c

@@ -219,8 +219,8 @@ void bli_herk_cntl_init()
 	                          NULL );
 
 	// Alias the "master" herk control tree to a shorter name.
-	herk_cntl = herk_cntl_mm_op;
-
+	//herk_cntl = herk_cntl_mm_op;
+	herk_cntl = herk_cntl_vl_mm;
 }
 
 void bli_herk_cntl_finalize()

frame/3/trmm/bli_trmm_cntl.c

@@ -227,8 +227,8 @@ void bli_trmm_cntl_init()
 	                          NULL );
 
 	// Alias the "master" trmm control tree to a shorter name.
-	trmm_cntl = trmm_cntl_mm_op;
-
+	//trmm_cntl = trmm_cntl_mm_op;
+	trmm_cntl = trmm_cntl_vl_mm;
 }
 
 void bli_trmm_cntl_finalize()

frame/3/trmm3/bli_trmm3_cntl.c

@@ -227,8 +227,8 @@ void bli_trmm3_cntl_init()
 	                          NULL );
 
 	// Alias the "master" trmm3 control tree to a shorter name.
-	trmm3_cntl = trmm3_cntl_mm_op;
-
+	//trmm3_cntl = trmm3_cntl_mm_op;
+	trmm3_cntl = trmm3_cntl_vl_mm;
 }
 
 void bli_trmm3_cntl_finalize()

frame/3/trsm/bli_trsm_cntl.c

@@ -227,8 +227,8 @@ void bli_trsm_cntl_init()
 	                          NULL );
 
 	// Alias the "master" trsm control tree to a shorter name.
-	trsm_cntl = trsm_cntl_mm_op;
-
+	//trsm_cntl = trsm_cntl_mm_op;
+	trsm_cntl = trsm_cntl_vl_mm;
 }
 
 void bli_trsm_cntl_finalize()

frame/include/bli_param_macro_defs.h

@@ -424,10 +424,6 @@
 \
 	( i1 != iter - 1 || left == 0 )
 
-#define bli_is_last_iter_f( i1, iter, left ) \
-\
-	( i1 == iter - 1 && left == 0 )
-
 #define bli_is_edge_b( i1, iter, left ) \
 \
 	( i1 == 0 && left != 0 )
@@ -436,10 +432,6 @@
 \
 	( i1 != 0 || left == 0 )
 
-#define bli_is_last_iter_b( i1, iter, left ) \
-\
-	( i1 == 0 && left == 0 )
-
 
 // packbuf_t-related