Skip to content
GitLab
Commit d71aa3df authored by Matthieu Schaller's avatar Matthieu Schaller
Browse files

Bring variable definitions closer to their actual use. Better readability.

parent 722cf879
Hide whitespace changes
Inline Side-by-side
src/runner_doiact_vec.c
View file @ d71aa3df
......@@ -532,8 +532,6 @@ __attribute__((always_inline)) INLINE void runner_doself1_density_vec(
struct part *restrict parts = c->parts;
const int count = c->count;
vector v_hi, v_vix, v_viy, v_viz, v_hig2, v_r2;
TIMER_TIC;
/* Anything to do here? */
......@@ -566,6 +564,7 @@ __attribute__((always_inline)) INLINE void runner_doself1_density_vec(
if (!part_is_active_no_debug(pi, max_active_bin)) continue;
vector v_pix, v_piy, v_piz;
vector v_hi, v_vix, v_viy, v_viz, v_hig2, v_r2;
const float hi = cell_cache->h[pid];
......@@ -586,9 +585,7 @@ __attribute__((always_inline)) INLINE void runner_doself1_density_vec(
v_curlvxSum, v_curlvySum, v_curlvzSum;
/* Get the inverse of hi. */
vector v_hi_inv;
v_hi_inv = vec_reciprocal(v_hi);
vector v_hi_inv = vec_reciprocal(v_hi);
v_rhoSum.v = vec_setzero();
v_rho_dhSum.v = vec_setzero();
......@@ -754,9 +751,6 @@ __attribute__((always_inline)) INLINE void runner_doself2_force_vec(
struct part *restrict parts = c->parts;
const int count = c->count;
vector v_hi, v_vix, v_viy, v_viz, v_hig2, v_r2;
vector v_rhoi, v_grad_hi, v_pOrhoi2, v_balsara_i, v_ci;
TIMER_TIC;
if (!cell_is_active(c, e)) return;
......@@ -793,6 +787,8 @@ __attribute__((always_inline)) INLINE void runner_doself2_force_vec(
if (!part_is_active_no_debug(pi, max_active_bin)) continue;
vector v_pix, v_piy, v_piz;
vector v_hi, v_vix, v_viy, v_viz, v_hig2;
vector v_rhoi, v_grad_hi, v_pOrhoi2, v_balsara_i, v_ci;
const float hi = cell_cache->h[pid];
......@@ -819,9 +815,7 @@ __attribute__((always_inline)) INLINE void runner_doself2_force_vec(
v_entropy_dtSum;
/* Get the inverse of hi. */
vector v_hi_inv;
v_hi_inv = vec_reciprocal(v_hi);
vector v_hi_inv = vec_reciprocal(v_hi);
v_a_hydro_xSum.v = vec_setzero();
v_a_hydro_ySum.v = vec_setzero();
......@@ -853,13 +847,12 @@ __attribute__((always_inline)) INLINE void runner_doself2_force_vec(
}
}
vector v_pjx, v_pjy, v_pjz, hj, hjg2;
/* Find all of particle pi's interacions and store needed values in the
* secondary cache.*/
for (int pjd = 0; pjd < count_align; pjd += (num_vec_proc * VEC_SIZE)) {
/* Load 1 set of vectors from the particle cache. */
vector v_pjx, v_pjy, v_pjz, hj, hjg2;
v_pjx.v = vec_load(&cell_cache->x[pjd]);
v_pjy.v = vec_load(&cell_cache->y[pjd]);
v_pjz.v = vec_load(&cell_cache->z[pjd]);
......@@ -867,8 +860,7 @@ __attribute__((always_inline)) INLINE void runner_doself2_force_vec(
hjg2.v = vec_mul(vec_mul(hj.v, hj.v), kernel_gamma2_vec.v);
/* Compute the pairwise distance. */
vector v_dx, v_dy, v_dz;
vector v_dx, v_dy, v_dz, v_r2;
v_dx.v = vec_sub(v_pix.v, v_pjx.v);
v_dy.v = vec_sub(v_piy.v, v_pjy.v);
v_dz.v = vec_sub(v_piz.v, v_pjz.v);
......@@ -895,12 +887,10 @@ __attribute__((always_inline)) INLINE void runner_doself2_force_vec(
/* If there are any interactions perform them. */
if (doi_mask) {
vector v_hj_inv;
v_hj_inv = vec_reciprocal(hj);
vector v_hj_inv = vec_reciprocal(hj);
/* To stop floating point exceptions for when particle separations are
* 0.
*/
* 0. */
v_r2.v = vec_add(v_r2.v, vec_set1(FLT_MIN));
runner_iact_nonsym_1_vec_force(
......@@ -947,8 +937,6 @@ void runner_dopair1_density_vec(struct runner *r, struct cell *ci,
const struct engine *restrict e = r->e;
const timebin_t max_active_bin = e->max_active_bin;
vector v_hi, v_vix, v_viy, v_viz, v_hig2;
TIMER_TIC;
/* Get the cutoff shift. */
......@@ -1064,6 +1052,7 @@ void runner_dopair1_density_vec(struct runner *r, struct cell *ci,
const int exit_iteration = max_index_i[pid];
vector v_pix, v_piy, v_piz;
vector v_hi, v_vix, v_viy, v_viz, v_hig2;
/* Fill particle pi vectors. */
v_pix.v = vec_set1(ci_cache->x[ci_cache_idx]);
......@@ -1082,9 +1071,7 @@ void runner_dopair1_density_vec(struct runner *r, struct cell *ci,
v_curlvxSum, v_curlvySum, v_curlvzSum;
/* Get the inverse of hi. */
vector v_hi_inv;
v_hi_inv = vec_reciprocal(v_hi);
vector v_hi_inv = vec_reciprocal(v_hi);
v_rhoSum.v = vec_setzero();
v_rho_dhSum.v = vec_setzero();
......@@ -1105,14 +1092,13 @@ void runner_dopair1_density_vec(struct runner *r, struct cell *ci,
exit_iteration_align += pad;
}
vector v_pjx, v_pjy, v_pjz;
/* Loop over the parts in cj. */
for (int pjd = 0; pjd <= exit_iteration_align; pjd += VEC_SIZE) {
/* Get the cache index to the jth particle. */
const int cj_cache_idx = pjd;
vector v_pjx, v_pjy, v_pjz;
vector v_dx, v_dy, v_dz, v_r2;
#ifdef SWIFT_DEBUG_CHECKS
......@@ -1211,9 +1197,7 @@ void runner_dopair1_density_vec(struct runner *r, struct cell *ci,
v_curlvxSum, v_curlvySum, v_curlvzSum;
/* Get the inverse of hj. */
vector v_hj_inv;
v_hj_inv = vec_reciprocal(v_hj);
vector v_hj_inv = vec_reciprocal(v_hj);
v_rhoSum.v = vec_setzero();
v_rho_dhSum.v = vec_setzero();
......@@ -1224,8 +1208,6 @@ void runner_dopair1_density_vec(struct runner *r, struct cell *ci,
v_curlvySum.v = vec_setzero();
v_curlvzSum.v = vec_setzero();
vector v_pix, v_piy, v_piz;
/* Convert exit iteration to cache indices. */
int exit_iteration_align = exit_iteration - first_pi;
......@@ -1250,6 +1232,7 @@ void runner_dopair1_density_vec(struct runner *r, struct cell *ci,
}
#endif
vector v_pix, v_piy, v_piz;
vector v_dx, v_dy, v_dz, v_r2;
/* Load 2 sets of vectors from the particle cache. */
......@@ -1323,9 +1306,6 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
const struct engine *restrict e = r->e;
const timebin_t max_active_bin = e->max_active_bin;
vector v_hi, v_vix, v_viy, v_viz, v_hig2, v_r2;
vector v_rhoi, v_grad_hi, v_pOrhoi2, v_balsara_i, v_ci;
TIMER_TIC;
/* Get the cutoff shift. */
......@@ -1448,6 +1428,8 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
const int exit_iteration = max_index_i[pid];
vector v_pix, v_piy, v_piz;
vector v_hi, v_vix, v_viy, v_viz, v_hig2;
vector v_rhoi, v_grad_hi, v_pOrhoi2, v_balsara_i, v_ci;
/* Fill particle pi vectors. */
v_pix.v = vec_set1(ci_cache->x[ci_cache_idx]);
......@@ -1472,8 +1454,7 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
v_sigSum, v_entropy_dtSum;
/* Get the inverse of hi. */
vector v_hi_inv;
v_hi_inv = vec_reciprocal(v_hi);
vector v_hi_inv = vec_reciprocal(v_hi);
v_a_hydro_xSum.v = vec_setzero();
v_a_hydro_ySum.v = vec_setzero();
......@@ -1492,8 +1473,6 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
exit_iteration_align += pad;
}
vector v_pjx, v_pjy, v_pjz, hj, hjg2;
/* Loop over the parts in cj. */
for (int pjd = 0; pjd <= exit_iteration_align; pjd += VEC_SIZE) {
......@@ -1501,6 +1480,7 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
const int cj_cache_idx = pjd;
vector v_dx, v_dy, v_dz;
vector v_pjx, v_pjy, v_pjz, v_hj, v_hjg2, v_r2;
#ifdef SWIFT_DEBUG_CHECKS
if (cj_cache_idx % VEC_SIZE != 0 || cj_cache_idx < 0 ||
......@@ -1514,8 +1494,8 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
v_pjx.v = vec_load(&cj_cache->x[cj_cache_idx]);
v_pjy.v = vec_load(&cj_cache->y[cj_cache_idx]);
v_pjz.v = vec_load(&cj_cache->z[cj_cache_idx]);
hj.v = vec_load(&cj_cache->h[cj_cache_idx]);
hjg2.v = vec_mul(vec_mul(hj.v, hj.v), kernel_gamma2_vec.v);
v_hj.v = vec_load(&cj_cache->h[cj_cache_idx]);
v_hjg2.v = vec_mul(vec_mul(v_hj.v, v_hj.v), kernel_gamma2_vec.v);
/* Compute the pairwise distance. */
v_dx.v = vec_sub(v_pix.v, v_pjx.v);
......@@ -1531,7 +1511,7 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
/* Form a mask from r2 < hig2 mask and r2 < hjg2 mask. */
vector v_h2;
v_h2.v = vec_fmax(v_hig2.v, hjg2.v);
v_h2.v = vec_fmax(v_hig2.v, v_hjg2.v);
vec_create_mask(v_doi_mask, vec_cmp_lt(v_r2.v, v_h2.v));
/* Form integer masks. */
......@@ -1539,8 +1519,7 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
/* If there are any interactions perform them. */
if (doi_mask) {
vector v_hj_inv;
v_hj_inv = vec_reciprocal(hj);
vector v_hj_inv = vec_reciprocal(v_hj);
runner_iact_nonsym_1_vec_force(
&v_r2, &v_dx, &v_dy, &v_dz, v_vix, v_viy, v_viz, v_rhoi,
......@@ -1616,9 +1595,7 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
v_sigSum, v_entropy_dtSum;
/* Get the inverse of hj. */
vector v_hj_inv;
v_hj_inv = vec_reciprocal(v_hj);
vector v_hj_inv = vec_reciprocal(v_hj);
v_a_hydro_xSum.v = vec_setzero();
v_a_hydro_ySum.v = vec_setzero();
......@@ -1637,8 +1614,6 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
} else
exit_iteration_align -= rem;
vector v_pix, v_piy, v_piz, hi, hig2;
/* Loop over the parts in ci. */
for (int ci_cache_idx = exit_iteration_align;
ci_cache_idx < ci_cache_count; ci_cache_idx += VEC_SIZE) {
......@@ -1649,14 +1624,15 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
}
#endif
vector v_dx, v_dy, v_dz;
vector v_pix, v_piy, v_piz, v_hi, v_hig2;
vector v_dx, v_dy, v_dz, v_r2;
/* Load 2 sets of vectors from the particle cache. */
v_pix.v = vec_load(&ci_cache->x[ci_cache_idx]);
v_piy.v = vec_load(&ci_cache->y[ci_cache_idx]);
v_piz.v = vec_load(&ci_cache->z[ci_cache_idx]);
hi.v = vec_load(&ci_cache->h[ci_cache_idx]);
hig2.v = vec_mul(vec_mul(hi.v, hi.v), kernel_gamma2_vec.v);
v_hi.v = vec_load(&ci_cache->h[ci_cache_idx]);
v_hig2.v = vec_mul(vec_mul(v_hi.v, v_hi.v), kernel_gamma2_vec.v);
/* Compute the pairwise distance. */
v_dx.v = vec_sub(v_pjx.v, v_pix.v);
......@@ -1672,7 +1648,7 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
/* Form a mask from r2 < hig2 mask and r2 < hjg2 mask. */
vector v_h2;
v_h2.v = vec_fmax(v_hjg2.v, hig2.v);
v_h2.v = vec_fmax(v_hjg2.v, v_hig2.v);
vec_create_mask(v_doj_mask, vec_cmp_lt(v_r2.v, v_h2.v));
/* Form integer masks. */
......@@ -1680,8 +1656,7 @@ void runner_dopair2_force_vec(struct runner *r, struct cell *ci,
/* If there are any interactions perform them. */
if (doj_mask) {
vector v_hi_inv;
v_hi_inv = vec_reciprocal(hi);
vector v_hi_inv = vec_reciprocal(v_hi);
runner_iact_nonsym_1_vec_force(
&v_r2, &v_dx, &v_dy, &v_dz, v_vjx, v_vjy, v_vjz, v_rhoj,
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment