Redesigned SIMD module and unit tests.

[alexxy/gromacs.git] / src / gromacs / mdlib / nbnxn_kernels / nbnxn_kernel_simd_utils_x86_128d.h

diff --git a/src/gromacs/mdlib/nbnxn_kernels/nbnxn_kernel_simd_utils_x86_128d.h b/src/gromacs/mdlib/nbnxn_kernels/nbnxn_kernel_simd_utils_x86_128d.h
index 92be81d99f3b073c1ad201d2ab201e12ae047926..d7a3d037a2cc71e1d1a23a64a358e7c5f90a96e4 100644 (file)
--- a/src/gromacs/mdlib/nbnxn_kernels/nbnxn_kernel_simd_utils_x86_128d.h
+++ b/src/gromacs/mdlib/nbnxn_kernels/nbnxn_kernel_simd_utils_x86_128d.h
@@ -45,8 +45,13 @@
  *   energy group pair energy storage
  */
 
+#define gmx_mm_extract_epi32(x, imm) _mm_cvtsi128_si32(_mm_srli_si128((x), 4 * (imm)))
+
 typedef gmx_simd_int32_t gmx_exclfilter;
-static const int filter_stride = GMX_SIMD_INT32_WIDTH/GMX_SIMD_REAL_WIDTH;
+/* This is set to a constant for now, since the code does not adapt automatically just
+ * because we set the SIMD widths to other values.
+ */
+static const int filter_stride = 2;
 
 /* Transpose 2 double precision registers */
 static gmx_inline void
@@ -183,13 +188,14 @@ gmx_load1_exclfilter(int e)
 static gmx_inline gmx_exclfilter
 gmx_load_exclusion_filter(const unsigned *i)
 {
-    return _mm_load_si128((__m128i *) i);
+    /* For now this has to be an explicit-float load since we use stride==2 */
+    return gmx_simd_load_fi(i);
 }
 
 static gmx_inline gmx_simd_bool_t
 gmx_checkbitmask_pb(gmx_exclfilter m0, gmx_exclfilter m1)
 {
-    return gmx_mm_castsi128_pd(_mm_cmpeq_epi32(_mm_andnot_si128(m0, m1), _mm_setzero_si128()));
+    return _mm_castsi128_pd(_mm_cmpeq_epi32(_mm_andnot_si128(m0, m1), _mm_setzero_si128()));
 }
 
 #endif /* _nbnxn_kernel_simd_utils_x86_s128d_h_ */