Merge remote-tracking branch 'gerrit/release-4-6'

[alexxy/gromacs.git] / src / gromacs / gmxlib / nonbonded / nb_kernel_sse2_single / sse_common_single.h

diff --git a/src/gromacs/gmxlib/nonbonded/nb_kernel_sse2_single/sse_common_single.h b/src/gromacs/gmxlib/nonbonded/nb_kernel_sse2_single/sse_common_single.h
index aa8ac2927133c2dbdcdc9583a203512624d55c62..2c70d5a5096f87b96349527f5fdfe23d8c30fb9a 100644 (file)
--- a/src/gromacs/gmxlib/nonbonded/nb_kernel_sse2_single/sse_common_single.h
+++ b/src/gromacs/gmxlib/nonbonded/nb_kernel_sse2_single/sse_common_single.h
@@ -31,17 +31,7 @@
 #include <config.h>
 #endif
 
-#include <xmmintrin.h> /* SSE */
-#ifdef GMX_SSE2
-#  include <emmintrin.h> /* SSE2 */
-#endif
-#ifdef GMX_SSE3
-#  include <pmmintrin.h> /* SSE3 */
-#endif
-#ifdef GMX_SSE4
-#  include <smmintrin.h> /* SSE4.1 */
-#endif
-
+#include <emmintrin.h>
 
 #include <stdio.h>
   
@@ -1029,11 +1019,7 @@ gmx_mm_int_lj_potonly_ps(__m128 rinvsq, __m128 c6, __m128 c12, __m128 *vvdwtot)
        return;
 }
 
-#ifdef GMX_SSE4
-#  define gmx_mm_extract_epi32(x, imm) _mm_extract_epi32(x,imm)
-#else
-#  define gmx_mm_extract_epi32(x, imm) _mm_cvtsi128_si32(_mm_srli_si128((x), 4 * (imm)))
-#endif
+#define gmx_mm_extract_epi32(x, imm) _mm_cvtsi128_si32(_mm_srli_si128((x), 4 * (imm)))
 
 
 /* Return force should be multiplied by -rinv to get fscal */
@@ -1902,18 +1888,11 @@ gmx_mm_update_iforce_1atom_ps(__m128 fix1, __m128 fiy1, __m128 fiz1,
 {
        __m128 t1,t2,t3;
        
-#ifdef GMX_SSE3
-       fix1 = _mm_hadd_ps(fix1,fix1);   
-       fiy1 = _mm_hadd_ps(fiy1,fiz1);
-       
-       fix1 = _mm_hadd_ps(fix1,fiy1); /* fiz1 fiy1 fix1 fix1 */ 
-#else
-       /* SSE2 */
        /* transpose data */
        t1 = fix1;
        _MM_TRANSPOSE4_PS(fix1,t1,fiy1,fiz1);  
        fix1 = _mm_add_ps(_mm_add_ps(fix1,t1), _mm_add_ps(fiy1,fiz1));
-#endif
+
        t2 = _mm_load_ss(fptr);
        t2 = _mm_loadh_pi(t2,(__m64 *)(fptr+1));
        t3 = _mm_load_ss(fshiftptr);
@@ -1936,15 +1915,6 @@ gmx_mm_update_iforce_2atoms_ps(__m128 fix1, __m128 fiy1, __m128 fiz1,
 {
        __m128 t1,t2,t3,t4;
        
-#ifdef GMX_SSE3
-       fix1 = _mm_hadd_ps(fix1,fiy1);   
-       fiz1 = _mm_hadd_ps(fiz1,fix2);
-       fiy2 = _mm_hadd_ps(fiy2,fiz2);
-       
-       fix1 = _mm_hadd_ps(fix1,fiz1); /* fix2 fiz1 fiy1 fix1 */ 
-       fiy2 = _mm_hadd_ps(fiy2,fiy2); /*  -    -   fiz2 fiy2 */
-#else
-       /* SSE2 */
        /* transpose data */
        _MM_TRANSPOSE4_PS(fix1,fiy1,fiz1,fix2);  
        t1 = _mm_unpacklo_ps(fiy2,fiz2);
@@ -1954,7 +1924,7 @@ gmx_mm_update_iforce_2atoms_ps(__m128 fix1, __m128 fiy1, __m128 fiz1,
        t1   = _mm_add_ps(t1,t2);
        t2   = _mm_movehl_ps(t2,t1);
        fiy2 = _mm_add_ps(t1,t2);
-#endif
+
        _mm_storeu_ps(fptr,   _mm_add_ps(fix1,_mm_loadu_ps(fptr)  ));
        t1 = _mm_loadl_pi(t1,(__m64 *)(fptr+4));
        _mm_storel_pi((__m64 *)(fptr+4), _mm_add_ps(fiy2,t1));
@@ -1985,18 +1955,6 @@ gmx_mm_update_iforce_3atoms_ps(__m128 fix1, __m128 fiy1, __m128 fiz1,
 {
        __m128 t1,t2,t3,t4;
        
-#ifdef GMX_SSE3
-       fix1 = _mm_hadd_ps(fix1,fiy1);   
-       fiz1 = _mm_hadd_ps(fiz1,fix2);
-       fiy2 = _mm_hadd_ps(fiy2,fiz2);
-       fix3 = _mm_hadd_ps(fix3,fiy3);
-       fiz3 = _mm_hadd_ps(fiz3,fiz3);
-       
-       fix1 = _mm_hadd_ps(fix1,fiz1); /* fix2 fiz1 fiy1 fix1 */ 
-       fiy2 = _mm_hadd_ps(fiy2,fix3); /* fiy3 fix3 fiz2 fiy2 */
-       fiz3 = _mm_hadd_ps(fiz3,fiz3); /*  -    -    -   fiz3 */
-#else
-       /* SSE2 */
        /* transpose data */
        _MM_TRANSPOSE4_PS(fix1,fiy1,fiz1,fix2);  
        _MM_TRANSPOSE4_PS(fiy2,fiz2,fix3,fiy3);
@@ -2007,7 +1965,7 @@ gmx_mm_update_iforce_3atoms_ps(__m128 fix1, __m128 fiy1, __m128 fiz1,
        fix1 = _mm_add_ps(_mm_add_ps(fix1,fiy1), _mm_add_ps(fiz1,fix2));
        fiy2 = _mm_add_ps(_mm_add_ps(fiy2,fiz2), _mm_add_ps(fix3,fiy3));
        fiz3 = _mm_add_ss(_mm_add_ps(fiz3,t1)  , _mm_add_ps(t2,t3));
-#endif
+
        _mm_storeu_ps(fptr,  _mm_add_ps(fix1,_mm_loadu_ps(fptr)  ));
        _mm_storeu_ps(fptr+4,_mm_add_ps(fiy2,_mm_loadu_ps(fptr+4)));
        _mm_store_ss (fptr+8,_mm_add_ss(fiz3,_mm_load_ss(fptr+8) ));
@@ -2038,19 +1996,6 @@ gmx_mm_update_iforce_4atoms_ps(__m128 fix1, __m128 fiy1, __m128 fiz1,
 {
        __m128 t1,t2,t3,t4,t5;
        
-#ifdef GMX_SSE3
-       fix1 = _mm_hadd_ps(fix1,fiy1);   
-       fiz1 = _mm_hadd_ps(fiz1,fix2);
-       fiy2 = _mm_hadd_ps(fiy2,fiz2);
-       fix3 = _mm_hadd_ps(fix3,fiy3);
-       fiz3 = _mm_hadd_ps(fiz3,fix4);
-       fiy4 = _mm_hadd_ps(fiy4,fiz4);
-       
-       fix1 = _mm_hadd_ps(fix1,fiz1); /* fix2 fiz1 fiy1 fix1 */ 
-       fiy2 = _mm_hadd_ps(fiy2,fix3); /* fiy3 fix3 fiz2 fiy2 */
-       fiz3 = _mm_hadd_ps(fiz3,fiy4); /* fiz4 fiy4 fix4 fiz3 */
-#else
-       /* SSE2 */
        /* transpose data */
        _MM_TRANSPOSE4_PS(fix1,fiy1,fiz1,fix2);  
        _MM_TRANSPOSE4_PS(fiy2,fiz2,fix3,fiy3);
@@ -2059,7 +2004,7 @@ gmx_mm_update_iforce_4atoms_ps(__m128 fix1, __m128 fiy1, __m128 fiz1,
        fix1 = _mm_add_ps(_mm_add_ps(fix1,fiy1), _mm_add_ps(fiz1,fix2));
        fiy2 = _mm_add_ps(_mm_add_ps(fiy2,fiz2), _mm_add_ps(fix3,fiy3));
        fiz3 = _mm_add_ps(_mm_add_ps(fiz3,fix4), _mm_add_ps(fiy4,fiz4));
-#endif
+
        _mm_storeu_ps(fptr,  _mm_add_ps(fix1,_mm_loadu_ps(fptr)  ));
        _mm_storeu_ps(fptr+4,_mm_add_ps(fiy2,_mm_loadu_ps(fptr+4)));
        _mm_storeu_ps(fptr+8,_mm_add_ps(fiz3,_mm_loadu_ps(fptr+8)));
@@ -2086,26 +2031,15 @@ gmx_mm_update_iforce_4atoms_ps(__m128 fix1, __m128 fiy1, __m128 fiz1,
 static inline void
 gmx_mm_update_1pot_ps(__m128 pot1, float *ptr1)
 {
-#ifdef GMX_SSE3
-       pot1 = _mm_hadd_ps(pot1,pot1);
-       pot1 = _mm_hadd_ps(pot1,pot1);
-#else
-       /* SSE2 */
        pot1 = _mm_add_ps(pot1,_mm_movehl_ps(pot1,pot1));
        pot1 = _mm_add_ps(pot1,_mm_shuffle_ps(pot1,pot1,_MM_SHUFFLE(0,0,0,1)));
-#endif
+
        _mm_store_ss(ptr1,_mm_add_ss(pot1,_mm_load_ss(ptr1)));
 }
                                   
 static inline void
 gmx_mm_update_2pot_ps(__m128 pot1, float *ptr1, __m128 pot2, float *ptr2)
 {
-#ifdef GMX_SSE3
-       pot1 = _mm_hadd_ps(pot1,pot2); 
-       pot1 = _mm_hadd_ps(pot1,pot1); 
-       pot2 = _mm_shuffle_ps(pot1,pot1,_MM_SHUFFLE(0,0,0,1));
-#else
-       /* SSE2 */
        __m128 t1,t2;
        t1   = _mm_movehl_ps(pot2,pot1); /* 2d 2c 1d 1c */
        t2   = _mm_movelh_ps(pot1,pot2); /* 2b 2a 1b 1a */
@@ -2113,7 +2047,7 @@ gmx_mm_update_2pot_ps(__m128 pot1, float *ptr1, __m128 pot2, float *ptr2)
        t2   = _mm_shuffle_ps(t1,t1,_MM_SHUFFLE(3,3,1,1));
        pot1 = _mm_add_ps(t1,t2);       /* -  2  -  1  */
        pot2 = _mm_movehl_ps(t2,t1);    /* -  -  -  2  */
-#endif
+
        _mm_store_ss(ptr1,_mm_add_ss(pot1,_mm_load_ss(ptr1)));
        _mm_store_ss(ptr2,_mm_add_ss(pot2,_mm_load_ss(ptr2)));
 }