Fixed a bug in determining default integer size.

Details:
- Fixed a bug that would cause configurations to inadvertantly define
  their integers to be 32 bits when those environments actually call for
  64-bit integers. While either BLIS_ARCH_64 or BLIS_ARCH_32 is defined
  in bli_system.h (based on whether preprocessor macros such as __x86_64
  or __aarch64__ are defined by the environment), bli_system.h was being
  #included *after* bli_config_macro_defs.h, in which the BLIS_ARCH_64
  macro was used to choose an integer type size in the event that
  BLIS_INT_TYPE_SIZE was not already defined by configure via
  bli_config.h. And due to the structure of the cpp code in that file,
  the 32-bit integer case was being chosen. Thanks to Francisco Igual
  and Devangi Parikh for their help in isolating this bug.
- Moved the #include of hbwmalloc.h and related preprocessor code to
  bli_kernel_macro_defs.h to facilitate the reshuffling of the #include
  for bli_system.h in blis.h.

frame/include/bli_kernel_macro_defs.h

@@ -57,6 +57,16 @@
 
 // -- Memory allocation --------------------------------------------------------
 
+// hbwmalloc.h provides hbw_malloc() and hbw_free() on systems with
+// libmemkind. But disable use of libmemkind if BLIS_DISABLE_MEMKIND
+// was explicitly defined.
+#ifdef BLIS_DISABLE_MEMKIND
+  #undef BLIS_ENABLE_MEMKIND
+#endif
+#ifdef BLIS_ENABLE_MEMKIND
+  #include <hbwmalloc.h>
+#endif
+
 // Memory allocation functions. These macros define the three types of
 // malloc()-style functions, and their free() counterparts: one for each
 // type of memory to be allocated.

frame/include/bli_system.h

@@ -43,15 +43,15 @@
 #include <float.h>
 #include <errno.h>
 
-// Determine if we are on a 64-bit or 32-bit architecture
+// Determine if we are on a 64-bit or 32-bit architecture.
 #if defined(_M_X64) || defined(__x86_64) || defined(__aarch64__) || \
     defined(_ARCH_PPC64)
 #define BLIS_ARCH_64
 #else
 #define BLIS_ARCH_32
 #endif
 
-// Determine the target operating system
+// Determine the target operating system.
 #if defined(_WIN32) || defined(__CYGWIN__)
 #define BLIS_OS_WINDOWS 1
 #elif defined(__APPLE__) || defined(__MACH__)
@@ -99,15 +99,5 @@
   #include <time.h>
 #endif
 
-// hbwmalloc.h provides hbw_malloc() and hbw_free() on systems with
-// libmemkind. But disable use of libmemkind if BLIS_DISABLE_MEMKIND
-// was explicitly defined.
-#ifdef BLIS_DISABLE_MEMKIND
-  #undef BLIS_ENABLE_MEMKIND
-#endif
-#ifdef BLIS_ENABLE_MEMKIND
-  #include <hbwmalloc.h>
-#endif
-
 
 #endif

frame/include/blis.h

@@ -44,23 +44,19 @@
 extern "C" {
 #endif
 
+// NOTE: PLEASE DON'T CHANGE THE ORDER IN WHICH HEADERS ARE INCLUDED UNLESS
+// YOU ARE SURE THAT IT DOESN'T BREAK INTER-HEADER MACRO DEPENDENCIES.
 
-// -- BLIS configuration definition --
+// -- System headers --
+// NOTE: This header must be included before bli_config_macro_defs.h.
 
-// NOTE: We include bli_config.h first because there might be something
-// defined there that is needed within one of the system headers. A good
-// example: posix_memalign() needs _GNU_SOURCE on GNU systems (I think).
-// 
-// PLEASE DON'T CHANGE THE ORDER IN WHICH HEADERS ARE INCLUDED UNLESS YOU
-// KNOW WHAT YOU ARE DOING.
+#include "bli_system.h"
 
-#include "bli_config.h"
-#include "bli_config_macro_defs.h"
 
+// -- configure definitions --
 
-// -- System headers --
-
-#include "bli_system.h"
+#include "bli_config.h"
+#include "bli_config_macro_defs.h"
 
 
 // -- Common BLIS definitions --