TREX/qmckl
1
0
Fork 0
mirror of https://github.com/TREX-CoE/qmckl.git synced 2024-12-22 20:36:01 +01:00
Code Issues Projects Releases Wiki Activity

Removed IVPDEP from SM

Browse Source
This commit is contained in:
Anthony Scemama 2023-09-14 09:53:17 +02:00
parent 2291103a9b
commit 10ee050050
2 changed files with 5 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
org/qmckl_memory.org
View File

@ -125,11 +125,10 @@ void* qmckl_malloc(qmckl_context context,
~qmckl_context~.
4. The function then allocates memory:
If the ~HAVE_HPC~ macro is defined, the memory allocation is done using
the ~aligned_alloc~ function with a 64-byte alignment, rounding up the
requested size to the nearest multiple of 64 bytes. If the ~HAVE_HPC~
macro is not defined, the memory allocation is done using the standard
~malloc~ function.
If the ~HAVE_HPC~ and ~HAVE_ALIGNED_ALLOC~ macros are defined, the memory
allocation is done using the ~aligned_alloc~ function with a 64-byte alignment,
rounding up the requested size to the nearest multiple of 64 bytes. Else, the
memory allocation is done using the standard ~malloc~ function.
5 If the allocation fails, the function returns ~NULL~.
@ -154,7 +153,7 @@ void* qmckl_malloc(qmckl_context context, const qmckl_memory_info_struct info) {
qmckl_context_struct* const ctx = (qmckl_context_struct*) context;
/* Allocate memory and zero it */
#ifdef HAVE_HPC
#if defined(HAVE_HPC) && defined(HAVE_ALIGNED_ALLOC)
assert( ((info.size+64) >> 6) << 6 >= info.size );
void * pointer = aligned_alloc(64, ((info.size+64) >> 6) << 6 );
#else

16
org/qmckl_sherman_morrison_woodbury.org
View File

@ -348,22 +348,6 @@ Common includes and macros used by all the Sherman-Morrison-Woodbury kernels.
#include "assert.h"
#include "stdio.h"
// Order important because
// __GNUC__ also set in ICC, ICX and CLANG
// __clang__ also set in ICX
#if defined(__INTEL_COMPILER)
#define IVDEP _Pragma("ivdep")
#define ALIGNED _Pragma("vector aligned")
#elif defined(__INTEL_LLVM_COMPILER)
#define IVDEP _Pragma("ivdep")
#define ALIGNED _Pragma("vector aligned")
#elif defined(__clang__)
#define IVDEP _Pragma("clang loop vectorize(enable)")
#define ALIGNED
#elif defined(__GNUC__)
#define IVDEP _Pragma("GCC ivdep")
#define ALIGNED
#endif
#+end_src
~qmckl_sm_naive_hpc~ is a high performance variation of