From a86d214c73e1ed46a4aef2cfb6d20447f835d9e5 Mon Sep 17 00:00:00 2001 From: Rafael Ravedutti Date: Tue, 28 Mar 2023 02:19:46 +0200 Subject: [PATCH] Add working version with old masking Signed-off-by: Rafael Ravedutti --- gromacs/atom.c | 6 +-- gromacs/force_lj.c | 95 +++++++++++++++++++++++++++++----------------- 2 files changed, 63 insertions(+), 38 deletions(-) diff --git a/gromacs/atom.c b/gromacs/atom.c index 2131b69..bfb2ab7 100644 --- a/gromacs/atom.c +++ b/gromacs/atom.c @@ -62,12 +62,12 @@ void createAtom(Atom *atom, Parameter *param) { } for(int j = 0; j < MAX(CLUSTER_M, VECTOR_WIDTH); j++) { - atom->diagonal_4xn_j_minus_i[j] = j - 0.5; + atom->diagonal_4xn_j_minus_i[j] = (MD_FLOAT)(j) - 0.5; } for(int j = 0; j < VECTOR_WIDTH / 2; j++) { - atom->diagonal_2xnn_j_minus_i[j] = j - 0.5; - atom->diagonal_2xnn_j_minus_i[VECTOR_WIDTH / 2 + j] = j - 1 - 0.5; + atom->diagonal_2xnn_j_minus_i[j] = (MD_FLOAT)(j) - 0.5; + atom->diagonal_2xnn_j_minus_i[VECTOR_WIDTH / 2 + j] = (MD_FLOAT)(j - 1) - 0.5; } for(int i = 0; i < CLUSTER_M * VECTOR_WIDTH; i++) { diff --git a/gromacs/force_lj.c b/gromacs/force_lj.c index 7d79014..923d4ef 100644 --- a/gromacs/force_lj.c +++ b/gromacs/force_lj.c @@ -190,15 +190,13 @@ double computeForceLJ_2xnn_half(Parameter *param, Atom *atom, Neighbor *neighbor MD_SIMD_FLOAT zero_S = simd_broadcast(0.0); MD_SIMD_FLOAT one_S = simd_broadcast(1.0); - /* - #if UNROLL_I == UNROLL_J - MD_SIMD_MASK diagonal_mask_S0, diagonal_mask_S2; - diagonal_mask0 = (zero_S < diagonal_jmi_S); + #if CLUSTER_M <= CLUSTER_N + MD_SIMD_MASK diagonal_mask0, diagonal_mask2; + diagonal_mask0 = simd_mask_cond_lt(zero_S, diagonal_jmi_S); diagonal_jmi_S = diagonal_jmi_S - one_S; diagonal_jmi_S = diagonal_jmi_S - one_S; - diagonal_mask2 = (zero_S < diagonal_jmi_S); - #elif 2 * UNROLL_I == UNROLL_J - */ + diagonal_mask2 = simd_mask_cond_lt(zero_S, diagonal_jmi_S); + #else MD_SIMD_MASK diagonal_mask00, diagonal_mask02, diagonal_mask10, diagonal_mask12; diagonal_mask00 = simd_mask_cond_lt(zero_S, diagonal_jmi_S); diagonal_jmi_S = diagonal_jmi_S - one_S; @@ -210,7 +208,7 @@ double computeForceLJ_2xnn_half(Parameter *param, Atom *atom, Neighbor *neighbor diagonal_jmi_S = diagonal_jmi_S - one_S; diagonal_jmi_S = diagonal_jmi_S - one_S; diagonal_mask12 = simd_mask_cond_lt(zero_S, diagonal_jmi_S); - //#endif + #endif #pragma omp for for(int ci = 0; ci < atom->Nclusters_local; ci++) { @@ -238,6 +236,7 @@ double computeForceLJ_2xnn_half(Parameter *param, Atom *atom, Neighbor *neighbor MD_SIMD_FLOAT fiz2 = simd_zero(); for(int k = 0; k < numneighs; k++) { + unsigned int mask0, mask1, mask2, mask3; int cj = neighs[k].cj; int cj_vec_base = CJ_VECTOR_BASE_INDEX(cj); int imask = neighs[k].imask; @@ -246,7 +245,7 @@ double computeForceLJ_2xnn_half(Parameter *param, Atom *atom, Neighbor *neighbor MD_SIMD_MASK interact0; MD_SIMD_MASK interact2; - gmx_load_simd_2xnn_interactions((int)imask, filter0, filter2, &interact0, &interact2); + //gmx_load_simd_2xnn_interactions((int)imask, filter0, filter2, &interact0, &interact2); MD_SIMD_FLOAT xj_tmp = simd_load_h_duplicate(&cj_x[CL_X_OFFSET]); MD_SIMD_FLOAT yj_tmp = simd_load_h_duplicate(&cj_x[CL_Y_OFFSET]); @@ -260,48 +259,74 @@ double computeForceLJ_2xnn_half(Parameter *param, Atom *atom, Neighbor *neighbor MD_SIMD_FLOAT rsq0 = simd_fma(delx0, delx0, simd_fma(dely0, dely0, simd_mul(delz0, delz0))); MD_SIMD_FLOAT rsq2 = simd_fma(delx2, delx2, simd_fma(dely2, dely2, simd_mul(delz2, delz2))); + #if CLUSTER_M == CLUSTER_N + unsigned int cond0 = (unsigned int)(cj == ci_cj0); + mask0 = (unsigned int)(0xf - 0x1 * cond0); + mask1 = (unsigned int)(0xf - 0x3 * cond0); + mask2 = (unsigned int)(0xf - 0x7 * cond0); + mask3 = (unsigned int)(0xf - 0xf * cond0); + #elif CLUSTER_M < CLUSTER_N + unsigned int cond0 = (unsigned int)((cj << 1) + 0 == ci); + unsigned int cond1 = (unsigned int)((cj << 1) + 1 == ci); + mask0 = (unsigned int)(0xff - 0x1 * cond0 - 0x1f * cond1); + mask1 = (unsigned int)(0xff - 0x3 * cond0 - 0x3f * cond1); + mask2 = (unsigned int)(0xff - 0x7 * cond0 - 0x7f * cond1); + mask3 = (unsigned int)(0xff - 0xf * cond0 - 0xff * cond1); + #else + unsigned int cond0 = (unsigned int)(cj == ci_cj0); + unsigned int cond1 = (unsigned int)(cj == ci_cj1); + mask0 = (unsigned int)(0x3 - 0x1 * cond0); + mask1 = (unsigned int)(0x3 - 0x3 * cond0); + mask2 = (unsigned int)(0x3 - cond0 * 0x3 - 0x1 * cond1); + mask3 = (unsigned int)(0x3 - cond0 * 0x3 - 0x3 * cond1); + #endif + + MD_SIMD_MASK excl_mask0 = simd_mask_from_u32((mask1 << half_mask_bits) | mask0); + MD_SIMD_MASK excl_mask2 = simd_mask_from_u32((mask3 << half_mask_bits) | mask2); MD_SIMD_MASK cutoff_mask0 = simd_mask_cond_lt(rsq0, cutforcesq_vec); MD_SIMD_MASK cutoff_mask2 = simd_mask_cond_lt(rsq2, cutforcesq_vec); + cutoff_mask0 = simd_mask_and(cutoff_mask0, excl_mask0); + cutoff_mask2 = simd_mask_and(cutoff_mask2, excl_mask2); - /*#if UNROLL_J == UNROLL_I - if(cj == ci) { - cutoff_mask0 = cutoff_mask0 && diagonal_mask0; - cutoff_mask2 = cutoff_mask2 && diagonal_mask2; + /* + #if CLUSTER_M <= CLUSTER_N + if(ci == ci_cj0) { + cutoff_mask0 = simd_mask_and(cutoff_mask0, diagonal_mask0); + cutoff_mask2 = simd_mask_and(cutoff_mask2, diagonal_mask2); } - #elif UNROLL_J == 2 * UNROLL_I*/ - if(cj * 2 == ci) { + #else + if(ci == ci_cj0) { cutoff_mask0 = cutoff_mask0 && diagonal_mask00; cutoff_mask2 = cutoff_mask2 && diagonal_mask02; - } else if (cj * 2 + 1 == ci) { + } else if(ci == ci_cj1) { cutoff_mask0 = cutoff_mask0 && diagonal_mask10; cutoff_mask2 = cutoff_mask2 && diagonal_mask12; } - /*else - # error "Invalid cluster configuration!" - #endif*/ + #endif + */ MD_SIMD_FLOAT sr2_0 = simd_reciprocal(rsq0); MD_SIMD_FLOAT sr2_2 = simd_reciprocal(rsq2); - MD_SIMD_FLOAT sr6_0 = simd_mul(sr2_0, simd_mul(sr2_0, simd_mul(sr2_0, sigma6_vec))); - MD_SIMD_FLOAT sr6_2 = simd_mul(sr2_2, simd_mul(sr2_2, simd_mul(sr2_2, sigma6_vec))); + MD_SIMD_FLOAT sr6_0 = sr2_0 * sr2_0 * sr2_0; + MD_SIMD_FLOAT sr6_2 = sr2_2 * sr2_2 * sr2_2; - MD_SIMD_FLOAT force0 = simd_mul(c48_vec, simd_mul(sr6_0, simd_mul(simd_sub(sr6_0, c05_vec), simd_mul(sr2_0, eps_vec)))); - MD_SIMD_FLOAT force2 = simd_mul(c48_vec, simd_mul(sr6_2, simd_mul(simd_sub(sr6_2, c05_vec), simd_mul(sr2_2, eps_vec)))); + MD_SIMD_FLOAT force0 = c48_vec * sr6_0 * (sr6_0 - c05_vec) * sr2_0 * eps_vec; + MD_SIMD_FLOAT force2 = c48_vec * sr6_2 * (sr6_2 - c05_vec) * sr2_2 * eps_vec; - MD_SIMD_FLOAT tx0 = select_by_mask(simd_mul(delx0, force0), cutoff_mask0); - MD_SIMD_FLOAT ty0 = select_by_mask(simd_mul(dely0, force0), cutoff_mask0); - MD_SIMD_FLOAT tz0 = select_by_mask(simd_mul(delz0, force0), cutoff_mask0); - MD_SIMD_FLOAT tx2 = select_by_mask(simd_mul(delx2, force2), cutoff_mask2); - MD_SIMD_FLOAT ty2 = select_by_mask(simd_mul(dely2, force2), cutoff_mask2); - MD_SIMD_FLOAT tz2 = select_by_mask(simd_mul(delz2, force2), cutoff_mask2); + MD_SIMD_FLOAT tx0 = select_by_mask(delx0 * force0, cutoff_mask0); + MD_SIMD_FLOAT ty0 = select_by_mask(dely0 * force0, cutoff_mask0); + MD_SIMD_FLOAT tz0 = select_by_mask(delz0 * force0, cutoff_mask0); + MD_SIMD_FLOAT tx2 = select_by_mask(delx2 * force2, cutoff_mask2); + MD_SIMD_FLOAT ty2 = select_by_mask(dely2 * force2, cutoff_mask2); + MD_SIMD_FLOAT tz2 = select_by_mask(delz2 * force2, cutoff_mask2); - fix0 = simd_add(fix0, tx0); - fiy0 = simd_add(fiy0, ty0); - fiz0 = simd_add(fiz0, tz0); - fix2 = simd_add(fix2, tx2); - fiy2 = simd_add(fiy2, ty2); - fiz2 = simd_add(fiz2, tz2); + fix0 += tx0; + fiy0 += ty0; + fiz0 += tz0; + fix2 += tx2; + fiy2 += ty2; + fiz2 += tz2; #ifdef HALF_NEIGHBOR_LISTS_CHECK_CJ if(cj < CJ1_FROM_CI(atom->Nlocal)) {