Skip to content
GitLab
Commit 1969a156 authored by Matthieu Schaller's avatar Matthieu Schaller
Browse files

Merge branch 'cache_auto_vec' into 'master' 

Cache auto vec

Auto-vectorise populating the cache with hints to the compiler and `restrict` keyword.

See merge request !404
parents fca42236 58b06679
Hide whitespace changes
Inline Side-by-side
src/align.h
View file @ 1969a156
......@@ -58,6 +58,22 @@
#define swift_align_information(array, alignment) ;
#endif
/**
* @brief Macro to create a restrict pointer to an array and tell the compiler
* that the given array has the specified
* alignment.
*
* Note that this turns into a no-op but gives information to the compiler.
*
* @param array The array.
* @param ptr Pointer to array
* @param type Type of array
* @param alignment The alignment in bytes of the array.
*/
#define swift_declare_aligned_ptr(type, array, ptr, alignment) \
type *restrict array = ptr; \
swift_align_information(array, alignment);
/**
* @brief Macro to tell the compiler that a given number is 0 modulo a given
* size.
......
src/cache.h
View file @ 1969a156
......@@ -23,6 +23,7 @@
#include "../config.h"
/* Local headers */
#include "align.h"
#include "cell.h"
#include "error.h"
#include "part.h"
......@@ -30,9 +31,7 @@
#include "vector.h"
#define NUM_VEC_PROC 2
#define CACHE_ALIGN 64
#define C2_CACHE_SIZE (NUM_VEC_PROC * VEC_SIZE * 6) + (NUM_VEC_PROC * VEC_SIZE)
#define C2_CACHE_ALIGN sizeof(float) * VEC_SIZE
#ifdef WITH_VECTORIZATION
/* Cache struct to hold a local copy of a cells' particle
......@@ -40,31 +39,31 @@
struct cache {
/* Particle x position. */
float *restrict x __attribute__((aligned(CACHE_ALIGN)));
float *restrict x SWIFT_CACHE_ALIGN;
/* Particle y position. */
float *restrict y __attribute__((aligned(CACHE_ALIGN)));
float *restrict y SWIFT_CACHE_ALIGN;
/* Particle z position. */
float *restrict z __attribute__((aligned(CACHE_ALIGN)));
float *restrict z SWIFT_CACHE_ALIGN;
/* Particle smoothing length. */
float *restrict h __attribute__((aligned(CACHE_ALIGN)));
float *restrict h SWIFT_CACHE_ALIGN;
/* Particle mass. */
float *restrict m __attribute__((aligned(CACHE_ALIGN)));
float *restrict m SWIFT_CACHE_ALIGN;
/* Particle x velocity. */
float *restrict vx __attribute__((aligned(CACHE_ALIGN)));
float *restrict vx SWIFT_CACHE_ALIGN;
/* Particle y velocity. */
float *restrict vy __attribute__((aligned(CACHE_ALIGN)));
float *restrict vy SWIFT_CACHE_ALIGN;
/* Particle z velocity. */
float *restrict vz __attribute__((aligned(CACHE_ALIGN)));
float *restrict vz SWIFT_CACHE_ALIGN;
/* Maximum index into neighbouring cell for particles that are in range. */
int *restrict max_index __attribute__((aligned(CACHE_ALIGN)));
int *restrict max_index SWIFT_CACHE_ALIGN;
/* Cache size. */
int count;
......@@ -75,28 +74,28 @@ struct cache {
struct c2_cache {
/* Separation between two particles squared. */
float r2q[C2_CACHE_SIZE] __attribute__((aligned(C2_CACHE_ALIGN)));
float r2q[C2_CACHE_SIZE] SWIFT_CACHE_ALIGN;
/* x separation between two particles. */
float dxq[C2_CACHE_SIZE] __attribute__((aligned(C2_CACHE_ALIGN)));
float dxq[C2_CACHE_SIZE] SWIFT_CACHE_ALIGN;
/* y separation between two particles. */
float dyq[C2_CACHE_SIZE] __attribute__((aligned(C2_CACHE_ALIGN)));
float dyq[C2_CACHE_SIZE] SWIFT_CACHE_ALIGN;
/* z separation between two particles. */
float dzq[C2_CACHE_SIZE] __attribute__((aligned(C2_CACHE_ALIGN)));
float dzq[C2_CACHE_SIZE] SWIFT_CACHE_ALIGN;
/* Mass of particle pj. */
float mq[C2_CACHE_SIZE] __attribute__((aligned(C2_CACHE_ALIGN)));
float mq[C2_CACHE_SIZE] SWIFT_CACHE_ALIGN;
/* x velocity of particle pj. */
float vxq[C2_CACHE_SIZE] __attribute__((aligned(C2_CACHE_ALIGN)));
float vxq[C2_CACHE_SIZE] SWIFT_CACHE_ALIGN;
/* y velocity of particle pj. */
float vyq[C2_CACHE_SIZE] __attribute__((aligned(C2_CACHE_ALIGN)));
float vyq[C2_CACHE_SIZE] SWIFT_CACHE_ALIGN;
/* z velocity of particle pj. */
float vzq[C2_CACHE_SIZE] __attribute__((aligned(C2_CACHE_ALIGN)));
float vzq[C2_CACHE_SIZE] SWIFT_CACHE_ALIGN;
};
/**
......@@ -130,15 +129,16 @@ __attribute__((always_inline)) INLINE void cache_init(struct cache *c,
free(c->max_index);
}
error += posix_memalign((void **)&c->x, CACHE_ALIGN, sizeBytes);
error += posix_memalign((void **)&c->y, CACHE_ALIGN, sizeBytes);
error += posix_memalign((void **)&c->z, CACHE_ALIGN, sizeBytes);
error += posix_memalign((void **)&c->m, CACHE_ALIGN, sizeBytes);
error += posix_memalign((void **)&c->vx, CACHE_ALIGN, sizeBytes);
error += posix_memalign((void **)&c->vy, CACHE_ALIGN, sizeBytes);
error += posix_memalign((void **)&c->vz, CACHE_ALIGN, sizeBytes);
error += posix_memalign((void **)&c->h, CACHE_ALIGN, sizeBytes);
error += posix_memalign((void **)&c->max_index, CACHE_ALIGN, sizeIntBytes);
error += posix_memalign((void **)&c->x, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error += posix_memalign((void **)&c->y, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error += posix_memalign((void **)&c->z, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error += posix_memalign((void **)&c->m, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error += posix_memalign((void **)&c->vx, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error += posix_memalign((void **)&c->vy, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error += posix_memalign((void **)&c->vz, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error += posix_memalign((void **)&c->h, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error += posix_memalign((void **)&c->max_index, SWIFT_CACHE_ALIGNMENT,
sizeIntBytes);
if (error != 0)
error("Couldn't allocate cache, no. of particles: %d", (int)count);
......@@ -152,156 +152,43 @@ __attribute__((always_inline)) INLINE void cache_init(struct cache *c,
* @param ci_cache The cache.
*/
__attribute__((always_inline)) INLINE void cache_read_particles(
const struct cell *const ci, struct cache *const ci_cache) {
const struct cell *restrict const ci,
struct cache *restrict const ci_cache) {
#if defined(GADGET2_SPH)
/* Shift the particles positions to a local frame so single precision can be
* used instead of double precision. */
#if defined(WITH_VECTORIZATION) && defined(__ICC)
#pragma vector aligned
#endif
for (int i = 0; i < ci->count; i++) {
ci_cache->x[i] = ci->parts[i].x[0] - ci->loc[0];
ci_cache->y[i] = ci->parts[i].x[1] - ci->loc[1];
ci_cache->z[i] = ci->parts[i].x[2] - ci->loc[2];
ci_cache->h[i] = ci->parts[i].h;
ci_cache->m[i] = ci->parts[i].mass;
ci_cache->vx[i] = ci->parts[i].v[0];
ci_cache->vy[i] = ci->parts[i].v[1];
ci_cache->vz[i] = ci->parts[i].v[2];
}
#endif
}
/**
* @brief Populate cache by reading in the particles from two cells in unsorted
* order.
*
* @param ci The i #cell.
* @param cj The j #cell.
* @param ci_cache The cache for cell ci.
* @param cj_cache The cache for cell cj.
* @param shift The amount to shift the particle positions to account for BCs
*/
__attribute__((always_inline)) INLINE void cache_read_two_cells(
const struct cell *const ci, const struct cell *const cj,
struct cache *const ci_cache, struct cache *const cj_cache,
const double *const shift) {
/* Shift the particles positions to a local frame (ci frame) so single
* precision can be
* used instead of double precision. Also shift the cell ci, particles
* positions due to BCs but leave cell cj. */
for (int i = 0; i < ci->count; i++) {
ci_cache->x[i] = ci->parts[i].x[0] - ci->loc[0] - shift[0];
ci_cache->y[i] = ci->parts[i].x[1] - ci->loc[1] - shift[1];
ci_cache->z[i] = ci->parts[i].x[2] - ci->loc[2] - shift[2];
ci_cache->h[i] = ci->parts[i].h;
ci_cache->m[i] = ci->parts[i].mass;
ci_cache->vx[i] = ci->parts[i].v[0];
ci_cache->vy[i] = ci->parts[i].v[1];
ci_cache->vz[i] = ci->parts[i].v[2];
}
for (int i = 0; i < cj->count; i++) {
cj_cache->x[i] = cj->parts[i].x[0] - ci->loc[0];
cj_cache->y[i] = cj->parts[i].x[1] - ci->loc[1];
cj_cache->z[i] = cj->parts[i].x[2] - ci->loc[2];
cj_cache->h[i] = cj->parts[i].h;
cj_cache->m[i] = cj->parts[i].mass;
cj_cache->vx[i] = cj->parts[i].v[0];
cj_cache->vy[i] = cj->parts[i].v[1];
cj_cache->vz[i] = cj->parts[i].v[2];
}
}
__attribute__((always_inline)) INLINE void cache_read_cell_sorted(
const struct cell *const ci, struct cache *const ci_cache,
const struct entry *restrict sort_i, double *const loc,
double *const shift) {
int idx;
/* Shift the particles positions to a local frame (ci frame) so single precision
* can be
* used instead of double precision. Also shift the cell ci, particles positions
* due to BCs but leave cell cj. */
#if defined(WITH_VECTORIZATION) && defined(__ICC)
#pragma simd
#endif
for (int i = 0; i < ci->count; i++) {
idx = sort_i[i].i;
ci_cache->x[i] = ci->parts[idx].x[0] - loc[0] - shift[0];
ci_cache->y[i] = ci->parts[idx].x[1] - loc[1] - shift[1];
ci_cache->z[i] = ci->parts[idx].x[2] - loc[2] - shift[2];
ci_cache->h[i] = ci->parts[idx].h;
ci_cache->m[i] = ci->parts[idx].mass;
ci_cache->vx[i] = ci->parts[idx].v[0];
ci_cache->vy[i] = ci->parts[idx].v[1];
ci_cache->vz[i] = ci->parts[idx].v[2];
}
}
/**
* @brief Populate cache by reading in the particles from two cells in sorted
* order.
*
* @param ci The i #cell.
* @param cj The j #cell.
* @param ci_cache The #cache for cell ci.
* @param cj_cache The #cache for cell cj.
* @param sort_i The array of sorted particle indices for cell ci.
* @param sort_j The array of sorted particle indices for cell ci.
* @param shift The amount to shift the particle positions to account for BCs
*/
__attribute__((always_inline)) INLINE void cache_read_two_cells_sorted(
const struct cell *const ci, const struct cell *const cj,
struct cache *const ci_cache, struct cache *const cj_cache,
const struct entry *restrict sort_i, const struct entry *restrict sort_j,
const double *const shift) {
int idx;
/* Shift the particles positions to a local frame (ci frame) so single precision
* can be
* used instead of double precision. Also shift the cell ci, particles positions
* due to BCs but leave cell cj. */
#if defined(WITH_VECTORIZATION) && defined(__ICC)
#pragma simd
#endif
/* Let the compiler know that the data is aligned and create pointers to the
* arrays inside the cache. */
swift_declare_aligned_ptr(float, x, ci_cache->x, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, y, ci_cache->y, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, z, ci_cache->z, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, h, ci_cache->h, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, m, ci_cache->m, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, vx, ci_cache->vx, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, vy, ci_cache->vy, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, vz, ci_cache->vz, SWIFT_CACHE_ALIGNMENT);
const struct part *restrict parts = ci->parts;
double loc[3];
loc[0] = ci->loc[0];
loc[1] = ci->loc[1];
loc[2] = ci->loc[2];
/* Shift the particles positions to a local frame so single precision can be
* used instead of double precision. */
for (int i = 0; i < ci->count; i++) {
idx = sort_i[i].i;
ci_cache->x[i] = ci->parts[idx].x[0] - ci->loc[0] - shift[0];
ci_cache->y[i] = ci->parts[idx].x[1] - ci->loc[1] - shift[1];
ci_cache->z[i] = ci->parts[idx].x[2] - ci->loc[2] - shift[2];
ci_cache->h[i] = ci->parts[idx].h;
ci_cache->m[i] = ci->parts[idx].mass;
ci_cache->vx[i] = ci->parts[idx].v[0];
ci_cache->vy[i] = ci->parts[idx].v[1];
ci_cache->vz[i] = ci->parts[idx].v[2];
x[i] = (float)(parts[i].x[0] - loc[0]);
y[i] = (float)(parts[i].x[1] - loc[1]);
z[i] = (float)(parts[i].x[2] - loc[2]);
h[i] = parts[i].h;
m[i] = parts[i].mass;
vx[i] = parts[i].v[0];
vy[i] = parts[i].v[1];
vz[i] = parts[i].v[2];
}
#if defined(WITH_VECTORIZATION) && defined(__ICC)
#pragma simd
#endif
for (int i = 0; i < cj->count; i++) {
idx = sort_j[i].i;
cj_cache->x[i] = cj->parts[idx].x[0] - ci->loc[0];
cj_cache->y[i] = cj->parts[idx].x[1] - ci->loc[1];
cj_cache->z[i] = cj->parts[idx].x[2] - ci->loc[2];
cj_cache->h[i] = cj->parts[idx].h;
cj_cache->m[i] = cj->parts[idx].mass;
cj_cache->vx[i] = cj->parts[idx].v[0];
cj_cache->vy[i] = cj->parts[idx].v[1];
cj_cache->vz[i] = cj->parts[idx].v[2];
}
}
/**
......@@ -322,13 +209,13 @@ __attribute__((always_inline)) INLINE void cache_read_two_cells_sorted(
* interaction.
*/
__attribute__((always_inline)) INLINE void cache_read_two_partial_cells_sorted(
const struct cell *const ci, const struct cell *const cj,
struct cache *const ci_cache, struct cache *const cj_cache,
const struct entry *restrict sort_i, const struct entry *restrict sort_j,
const double *const shift, int *first_pi, int *last_pj,
const int num_vec_proc) {
const struct cell *restrict const ci, const struct cell *restrict const cj,
struct cache *restrict const ci_cache,
struct cache *restrict const cj_cache, const struct entry *restrict sort_i,
const struct entry *restrict sort_j, const double *restrict const shift,
int *first_pi, int *last_pj, const int num_vec_proc) {
int idx, ci_cache_idx;
int idx;
/* Pad number of particles read to the vector size. */
int rem = (ci->count - *first_pi) % (num_vec_proc * VEC_SIZE);
if (rem != 0) {
......@@ -346,74 +233,97 @@ __attribute__((always_inline)) INLINE void cache_read_two_partial_cells_sorted(
int first_pi_align = *first_pi;
int last_pj_align = *last_pj;
/* Shift the particles positions to a local frame (ci frame) so single precision
* can be
* used instead of double precision. Also shift the cell ci, particles positions
* due to BCs but leave cell cj. */
#if defined(WITH_VECTORIZATION) && defined(__ICC)
#pragma vector aligned
#endif
for (int i = first_pi_align; i < ci->count; i++) {
/* Make sure ci_cache is filled from the first element. */
ci_cache_idx = i - first_pi_align;
idx = sort_i[i].i;
ci_cache->x[ci_cache_idx] = ci->parts[idx].x[0] - ci->loc[0] - shift[0];
ci_cache->y[ci_cache_idx] = ci->parts[idx].x[1] - ci->loc[1] - shift[1];
ci_cache->z[ci_cache_idx] = ci->parts[idx].x[2] - ci->loc[2] - shift[2];
ci_cache->h[ci_cache_idx] = ci->parts[idx].h;
ci_cache->m[ci_cache_idx] = ci->parts[idx].mass;
ci_cache->vx[ci_cache_idx] = ci->parts[idx].v[0];
ci_cache->vy[ci_cache_idx] = ci->parts[idx].v[1];
ci_cache->vz[ci_cache_idx] = ci->parts[idx].v[2];
const struct part *restrict parts_i = ci->parts;
const struct part *restrict parts_j = cj->parts;
double loc[3];
loc[0] = ci->loc[0];
loc[1] = ci->loc[1];
loc[2] = ci->loc[2];
/* Let the compiler know that the data is aligned and create pointers to the
* arrays inside the cache. */
swift_declare_aligned_ptr(float, x, ci_cache->x, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, y, ci_cache->y, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, z, ci_cache->z, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, h, ci_cache->h, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, m, ci_cache->m, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, vx, ci_cache->vx, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, vy, ci_cache->vy, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, vz, ci_cache->vz, SWIFT_CACHE_ALIGNMENT);
int ci_cache_count = ci->count - first_pi_align;
/* Shift the particles positions to a local frame (ci frame) so single
* precision
* can be
* used instead of double precision. Also shift the cell ci, particles
* positions
* due to BCs but leave cell cj. */
for (int i = 0; i < ci_cache_count; i++) {
idx = sort_i[i + first_pi_align].i;
x[i] = (float)(parts_i[idx].x[0] - loc[0] - shift[0]);
y[i] = (float)(parts_i[idx].x[1] - loc[1] - shift[1]);
z[i] = (float)(parts_i[idx].x[2] - loc[2] - shift[2]);
h[i] = parts_i[idx].h;
m[i] = parts_i[idx].mass;
vx[i] = parts_i[idx].v[0];
vy[i] = parts_i[idx].v[1];
vz[i] = parts_i[idx].v[2];
}
/* Pad cache with fake particles that exist outside the cell so will not
* interact.*/
float fake_pix = 2.0f * ci->parts[sort_i[ci->count - 1].i].x[0];
float fake_pix = 2.0f * parts_i[sort_i[ci->count - 1].i].x[0];
for (int i = ci->count - first_pi_align;
i < ci->count - first_pi_align + VEC_SIZE; i++) {
ci_cache->x[i] = fake_pix;
ci_cache->y[i] = 1.f;
ci_cache->z[i] = 1.f;
ci_cache->h[i] = 1.f;
ci_cache->m[i] = 1.f;
ci_cache->vx[i] = 1.f;
ci_cache->vy[i] = 1.f;
ci_cache->vz[i] = 1.f;
x[i] = fake_pix;
y[i] = 1.f;
z[i] = 1.f;
h[i] = 1.f;
m[i] = 1.f;
vx[i] = 1.f;
vy[i] = 1.f;
vz[i] = 1.f;
}
#if defined(WITH_VECTORIZATION) && defined(__ICC)
#pragma vector aligned
#endif
/* Let the compiler know that the data is aligned and create pointers to the
* arrays inside the cache. */
swift_declare_aligned_ptr(float, xj, cj_cache->x, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, yj, cj_cache->y, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, zj, cj_cache->z, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, hj, cj_cache->h, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, mj, cj_cache->m, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, vxj, cj_cache->vx, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, vyj, cj_cache->vy, SWIFT_CACHE_ALIGNMENT);
swift_declare_aligned_ptr(float, vzj, cj_cache->vz, SWIFT_CACHE_ALIGNMENT);
for (int i = 0; i <= last_pj_align; i++) {
idx = sort_j[i].i;
cj_cache->x[i] = cj->parts[idx].x[0] - ci->loc[0];
cj_cache->y[i] = cj->parts[idx].x[1] - ci->loc[1];
cj_cache->z[i] = cj->parts[idx].x[2] - ci->loc[2];
cj_cache->h[i] = cj->parts[idx].h;
cj_cache->m[i] = cj->parts[idx].mass;
cj_cache->vx[i] = cj->parts[idx].v[0];
cj_cache->vy[i] = cj->parts[idx].v[1];
cj_cache->vz[i] = cj->parts[idx].v[2];
xj[i] = (float)(parts_j[idx].x[0] - loc[0]);
yj[i] = (float)(parts_j[idx].x[1] - loc[1]);
zj[i] = (float)(parts_j[idx].x[2] - loc[2]);
hj[i] = parts_j[idx].h;
mj[i] = parts_j[idx].mass;
vxj[i] = parts_j[idx].v[0];
vyj[i] = parts_j[idx].v[1];
vzj[i] = parts_j[idx].v[2];
}
/* Pad cache with fake particles that exist outside the cell so will not
* interact.*/
float fake_pjx = 2.0f * cj->parts[sort_j[cj->count - 1].i].x[0];
for (int i = last_pj_align + 1; i < last_pj_align + 1 + VEC_SIZE; i++) {
cj_cache->x[i] = fake_pjx;
cj_cache->y[i] = 1.f;
cj_cache->z[i] = 1.f;
cj_cache->h[i] = 1.f;
cj_cache->m[i] = 1.f;
cj_cache->vx[i] = 1.f;
cj_cache->vy[i] = 1.f;
cj_cache->vz[i] = 1.f;
xj[i] = fake_pjx;
yj[i] = 1.f;
zj[i] = 1.f;
hj[i] = 1.f;
mj[i] = 1.f;
vxj[i] = 1.f;
vyj[i] = 1.f;
vzj[i] = 1.f;
}
}
......
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