Skip to content
GitLab
Commit be7b9694 authored by Matthieu Schaller's avatar Matthieu Schaller
Browse files

Better definition of the doself_grav_pp function and better documentation.

parent cb0b3c77
Hide whitespace changes
Inline Side-by-side
src/gravity_cache.h
View file @ be7b9694
......@@ -175,8 +175,6 @@ __attribute__((always_inline)) INLINE static void gravity_cache_populate(
SWIFT_CACHE_ALIGNMENT);
swift_assume_size(gcount_padded, VEC_SIZE);
if (shift[0] != 0.) error("OO");
/* Fill the input caches */
for (int i = 0; i < gcount; ++i) {
x[i] = (float)(gparts[i].x[0] - shift[0]);
......@@ -191,6 +189,7 @@ __attribute__((always_inline)) INLINE static void gravity_cache_populate(
float dy = y[i] - CoM[1];
float dz = z[i] - CoM[2];
/* Apply periodic BC */
if (periodic) {
dx = nearestf(dx, dim[0]);
dy = nearestf(dy, dim[1]);
......
src/runner_doiact_grav.h
View file @ be7b9694
......@@ -241,8 +241,6 @@ static INLINE void runner_dopair_grav_pp_full(
const float dim[3], const struct engine *restrict e,
struct gpart *restrict gparts_i, struct gpart *restrict gparts_j) {
TIMER_TIC;
/* Loop over all particles in ci... */
for (int pid = 0; pid < gcount_i; pid++) {
......@@ -334,8 +332,6 @@ static INLINE void runner_dopair_grav_pp_full(
ci_cache->a_z[pid] = a_z;
ci_cache->pot[pid] = pot;
}
TIMER_TOC(timer_dopair_grav_pp_full);
}
/**
......@@ -368,8 +364,6 @@ static INLINE void runner_dopair_grav_pp_truncated(
const float r_s_inv, const struct engine *restrict e,
struct gpart *restrict gparts_i, struct gpart *restrict gparts_j) {
TIMER_TIC;
#ifdef SWIFT_DEBUG_CHECKS
if (!e->s->periodic)
error("Calling truncated PP function in non-periodic setup.");
......@@ -464,8 +458,6 @@ static INLINE void runner_dopair_grav_pp_truncated(
ci_cache->a_z[pid] = a_z;
ci_cache->pot[pid] = pot;
}
TIMER_TOC(timer_dopair_grav_pp_trunc);
}
/**
......@@ -496,8 +488,6 @@ static INLINE void runner_dopair_grav_pm_full(
struct gpart *restrict gparts_i, const int gcount_i,
const struct cell *restrict cj) {
TIMER_TIC;
/* Make the compiler understand we are in happy vectorization land */
swift_declare_aligned_ptr(float, x, ci_cache->x, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, y, ci_cache->y, SWIFT_CACHE_ALIGNMENT);
......@@ -580,8 +570,6 @@ static INLINE void runner_dopair_grav_pm_full(
gparts_i[pid].num_interacted += cj->multipole->m_pole.num_gpart;
#endif
}
TIMER_TOC(timer_dopair_grav_pm_full);
}
/**
......@@ -619,8 +607,6 @@ static INLINE void runner_dopair_grav_pm_truncated(
error("Calling truncated PP function in non-periodic setup.");
#endif
TIMER_TIC;
/* Make the compiler understand we are in happy vectorization land */
swift_declare_aligned_ptr(float, x, ci_cache->x, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, y, ci_cache->y, SWIFT_CACHE_ALIGNMENT);
......@@ -701,8 +687,6 @@ static INLINE void runner_dopair_grav_pm_truncated(
gparts_i[pid].num_interacted += cj->multipole->m_pole.num_gpart;
#endif
}
TIMER_TOC(timer_dopair_grav_pm_trunc);
}
/**
......@@ -906,52 +890,26 @@ static INLINE void runner_dopair_grav_pp(struct runner *r, struct cell *ci,
if (cj_active && symmetric)
gravity_cache_write_back(cj_cache, cj->gparts, gcount_j);
TIMER_TOC(timer_dopair_grav_branch);
TIMER_TOC(timer_dopair_grav_pp);
}
/**
* @brief Computes the interaction of all the particles in a cell using the
* full Newtonian potential.
* @brief Compute the non-truncated gravity interactions between all particles
* of a cell and the particles of the other cell.
*
* @param r The #runner.
* @param c The #cell.
* The calculation is performed non-symmetrically using the pre-filled
* #gravity_cache structures. The loop over the j cache should auto-vectorize.
*
* @todo Use a local cache for the particles.
*/
static INLINE void runner_doself_grav_pp_full(struct runner *r,
struct cell *c) {
error("bb");
/* Some constants */
const struct engine *const e = r->e;
struct gravity_cache *const ci_cache = &r->ci_gravity_cache;
/* Cell properties */
const int gcount = c->gcount;
struct gpart *restrict gparts = c->gparts;
const int c_active = cell_is_active_gravity(c, e);
const double loc[3] = {c->loc[0] + 0.5 * c->width[0],
c->loc[1] + 0.5 * c->width[1],
c->loc[2] + 0.5 * c->width[2]};
/* Anything to do here ?*/
if (!c_active) return;
#ifdef SWIFT_DEBUG_CHECKS
/* Check that we fit in cache */
if (gcount > ci_cache->count)
error("Not enough space in the cache! gcount=%d", gcount);
#endif
/* Computed the padded counts */
const int gcount_padded = gcount - (gcount % VEC_SIZE) + VEC_SIZE;
gravity_cache_populate_no_mpole(e->max_active_bin, ci_cache, gparts, gcount,
gcount_padded, loc, c, e->gravity_properties);
/* Ok... Here we go ! */
* @param ci_cache #gravity_cache contaning the particles to be updated.
* @param gcount The number of particles in the cell.
* @param gcount_padded The number of particles in the cell padded to the
* vector length.
*
* @param e The #engine (for debugging checks only).
* @param gparts The #gpart in the cell (for debugging checks only).
*/ static INLINE void runner_doself_grav_pp_full(
struct gravity_cache *restrict ci_cache, const int gcount,
const int gcount_padded, const struct engine *e, struct gpart *gparts) {
/* Loop over all particles in ci... */
for (int pid = 0; pid < gcount; pid++) {
......@@ -1031,56 +989,36 @@ static INLINE void runner_doself_grav_pp_full(struct runner *r,
ci_cache->a_z[pid] = a_z;
ci_cache->pot[pid] = pot;
}
/* Write back to the particles */
gravity_cache_write_back(ci_cache, gparts, gcount);
}
/**
* @brief Computes the interaction of all the particles in a cell using the
* truncated Newtonian potential.
* @brief Compute the truncated gravity interactions between all particles
* of a cell and the particles of the other cell.
*
* @param r The #runner.
* @param c The #cell.
* The calculation is performed non-symmetrically using the pre-filled
* #gravity_cache structures. The loop over the j cache should auto-vectorize.
*
* @todo Use a local cache for the particles.
* This function only makes sense in periodic BCs.
*
* @param ci_cache #gravity_cache contaning the particles to be updated.
* @param gcount The number of particles in the cell.
* @param gcount_padded The number of particles in the cell padded to the
* vector length.
* @param r_s_inv The inverse of the gravity-mesh smoothing-scale.
*
* @param e The #engine (for debugging checks only).
* @param gparts The #gpart in the cell (for debugging checks only).
*/
static INLINE void runner_doself_grav_pp_truncated(struct runner *r,
struct cell *c) {
/* Some constants */
const struct engine *const e = r->e;
const float r_s_inv = e->mesh->r_s_inv;
/* Caches to play with */
struct gravity_cache *const ci_cache = &r->ci_gravity_cache;
/* Cell properties */
const int gcount = c->gcount;
struct gpart *restrict gparts = c->gparts;
const int c_active = cell_is_active_gravity(c, e);
const double loc[3] = {0., 0., 0.};
//{c->loc[0] + 0.5 * c->width[0],
// c->loc[1] + 0.5 * c->width[1],
// c->loc[2] + 0.5 * c->width[2]};
/* Anything to do here ?*/
if (!c_active) return;
static INLINE void runner_doself_grav_pp_truncated(
struct gravity_cache *restrict ci_cache, const int gcount,
const int gcount_padded, const float r_s_inv, const struct engine *e,
struct gpart *gparts) {
#ifdef SWIFT_DEBUG_CHECKS
/* Check that we fit in cache */
if (gcount > ci_cache->count)
error("Not enough space in the caches! gcount=%d", gcount);
if (!e->s->periodic)
error("Calling truncated PP function in non-periodic setup.");
#endif
/* Computed the padded counts */
const int gcount_padded = gcount - (gcount % VEC_SIZE) + VEC_SIZE;
gravity_cache_populate_no_mpole(e->max_active_bin, ci_cache, gparts, gcount,
gcount_padded, loc, c, e->gravity_properties);
/* Ok... Here we go ! */
/* Loop over all particles in ci... */
for (int pid = 0; pid < gcount; pid++) {
......@@ -1161,24 +1099,29 @@ static INLINE void runner_doself_grav_pp_truncated(struct runner *r,
ci_cache->a_z[pid] = a_z;
ci_cache->pot[pid] = pot;
}
/* Write back to the particles */
gravity_cache_write_back(ci_cache, gparts, gcount);
}
/**
* @brief Computes the interaction of all the particles in a cell directly
* (Switching function between truncated and full)
* @brief Computes the interaction of all the particles in a cell with all the
* other ones.
*
* This function switches between the full potential and the truncated one
* depending on needs.
*
* This function starts by constructing the require #gravity_cache for the
* cell and then call the specialised functions doing the actual work on
* the cache. It then write the data back to the particles.
*
* @param r The #runner.
* @param c The #cell.
*/
static INLINE void runner_doself_grav_pp(struct runner *r, struct cell *c) {
/* Some properties of the space */
/* Recover some useful constants */
const struct engine *e = r->e;
const struct space *s = e->s;
const int periodic = s->periodic;
const float r_s_inv = e->mesh->r_s_inv;
const double r_s = e->mesh->r_s;
const double min_trunc = e->gravity_properties->r_cut_min * r_s;
......@@ -1197,21 +1140,57 @@ static INLINE void runner_doself_grav_pp(struct runner *r, struct cell *c) {
/* Do we need to start by drifting things ? */
if (!cell_are_gpart_drifted(c, e)) error("Un-drifted gparts");
/* Start by constructing a cache for the particles */
struct gravity_cache *const ci_cache = &r->ci_gravity_cache;
/* Shift to apply to the particles in the cell */
const double loc[3] = {c->loc[0] + 0.5 * c->width[0],
c->loc[1] + 0.5 * c->width[1],
c->loc[2] + 0.5 * c->width[2]};
/* Computed the padded counts */
const int gcount = c->gcount;
const int gcount_padded = gcount - (gcount % VEC_SIZE) + VEC_SIZE;
#ifdef SWIFT_DEBUG_CHECKS
/* Check that we fit in cache */
if (gcount > ci_cache->count)
error("Not enough space in the cache! gcount=%d", gcount);
#endif
/* Fill the cache */
gravity_cache_populate_no_mpole(e->max_active_bin, ci_cache, c->gparts,
gcount, gcount_padded, loc, c,
e->gravity_properties);
/* Can we use the Newtonian version or do we need the truncated one ? */
if (!periodic) {
runner_doself_grav_pp_full(r, c);
/* Not periodic -> Can always use Newtonian potential */
runner_doself_grav_pp_full(ci_cache, gcount, gcount_padded, e, c->gparts);
} else {
/* Get the maximal distance between any two particles */
const double max_r = 2. * c->multipole->r_max;
/* Do we need to use the truncated interactions ? */
if (max_r > min_trunc)
runner_doself_grav_pp_truncated(r, c);
else
runner_doself_grav_pp_full(r, c);
if (max_r > min_trunc) {
/* Periodic but far-away cells must use the truncated potential */
runner_doself_grav_pp_truncated(ci_cache, gcount, gcount_padded, r_s_inv,
e, c->gparts);
} else {
/* Periodic but close-by cells can use the full Newtonian potential */
runner_doself_grav_pp_full(ci_cache, gcount, gcount_padded, e, c->gparts);
}
}
/* Write back to the particles */
gravity_cache_write_back(ci_cache, c->gparts, gcount);
TIMER_TOC(timer_doself_grav_pp);
}
......
src/timers.c
View file @ be7b9694
......@@ -55,10 +55,7 @@ const char* timers_names[timer_count] = {
"dopair_gradient",
"dopair_force",
"dopair_grav_mm",
"dopair_grav_pm_full",
"dopair_grav_pm_trunc",
"dopair_grav_pp_full",
"dopair_grav_pp_trunc",
"dopair_grav_pp",
"dograv_external",
"dograv_down",
"dograv_mesh",
......
src/timers.h
View file @ be7b9694
......@@ -56,10 +56,7 @@ enum {
timer_dopair_gradient,
timer_dopair_force,
timer_dopair_grav_mm,
timer_dopair_grav_pm_full,
timer_dopair_grav_pm_trunc,
timer_dopair_grav_pp_full,
timer_dopair_grav_pp_trunc,
timer_dopair_grav_pp,
timer_dograv_external,
timer_dograv_down,
timer_dograv_mesh,
......
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