Skip to content
GitLab
Commit b2b4a66f authored by Matthieu Schaller's avatar Matthieu Schaller
Browse files

Documentation. Make sure the code compiles and runs without vectorization switched on.

parent e7f2ff8c
Hide whitespace changes
Inline Side-by-side
src/align.h
View file @ b2b4a66f
......@@ -54,6 +54,8 @@
#elif defined(__GNUC__)
#define swift_align_information(array, alignment) \
array = __builtin_assume_aligned(array, alignment);
#else
#define swift_align_information(array, alignment) ;
#endif
/**
......
src/gravity_cache.h
View file @ b2b4a66f
......@@ -28,8 +28,11 @@
#include "gravity.h"
#include "vector.h"
#ifdef WITH_VECTORIZATION
/**
* @brief A SoA object for the #gpart of a cell.
*
* This is used to help vectorize the leaf-leaf gravity interactions.
*/
struct gravity_cache {
/*! #gpart x position. */
......@@ -62,12 +65,12 @@ struct gravity_cache {
/**
* @brief Frees the memory allocated in a #gravity_cache
*
* @param c The #grvity_cache to free.
*
* @param c The #gravity_cache to free.
*/
static INLINE void gravity_cache_clean(struct gravity_cache *c) {
if(c->count > 0){
if (c->count > 0) {
free(c->x);
free(c->y);
free(c->z);
......@@ -81,19 +84,21 @@ static INLINE void gravity_cache_clean(struct gravity_cache *c) {
}
/**
* @brief Allocates memory for the #gpart caches used in the leaf-leaf interactions.
* @brief Allocates memory for the #gpart caches used in the leaf-leaf
* interactions.
*
* The cache is padded for the vector size and aligned properly
*
* @param c The #gravity_cache to allocate.
* @param count The number of #gpart to allocated for (space_splitsize is a good choice).
* @param count The number of #gpart to allocated for (space_splitsize is a good
* choice).
*/
static INLINE void gravity_cache_init(struct gravity_cache *c, int count) {
/* Size of the gravity cache */
const int padded_count = count - (count % VEC_SIZE) + VEC_SIZE;
const size_t sizeBytes = padded_count * sizeof(float);
/* Delete old stuff if any */
gravity_cache_clean(c);
......@@ -101,7 +106,8 @@ static INLINE void gravity_cache_init(struct gravity_cache *c, int count) {
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->epsilon, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error +=
posix_memalign((void **)&c->epsilon, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error += posix_memalign((void **)&c->m, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error += posix_memalign((void **)&c->a_x, SWIFT_CACHE_ALIGNMENT, sizeBytes);
error += posix_memalign((void **)&c->a_y, SWIFT_CACHE_ALIGNMENT, sizeBytes);
......@@ -113,8 +119,20 @@ static INLINE void gravity_cache_init(struct gravity_cache *c, int count) {
c->count = padded_count;
}
__attribute__((always_inline)) INLINE void gravity_cache_populate(struct gravity_cache *c, const struct gpart *restrict gparts, int gcount, int gcount_padded, int zero_output, double shift[3]) {
/**
* @brief Fills a #gravity_cache structure with some #gpart and shift them.
*
* @param c The #gravity_cache to fill.
* @param gparts The #gpart array to read from.
* @param gcount The number of particles to read.
* @param gcount_padded The number of particle to read padded to the next
* multiple of the vector length.
* @param zero_output Do we need to zero the output caches ?
* @param shift A shift to apply to all the particles.
*/
__attribute__((always_inline)) INLINE void gravity_cache_populate(
struct gravity_cache *c, const struct gpart *restrict gparts, int gcount,
int gcount_padded, int zero_output, double shift[3]) {
/* Make the compiler understand we are in happy vectorization land */
float *restrict x = c->x;
......@@ -128,7 +146,7 @@ __attribute__((always_inline)) INLINE void gravity_cache_populate(struct gravity
swift_align_information(x, SWIFT_CACHE_ALIGNMENT);
swift_align_information(y, SWIFT_CACHE_ALIGNMENT);
swift_align_information(z, SWIFT_CACHE_ALIGNMENT);
swift_align_information(epsilon, SWIFT_CACHE_ALIGNMENT);
swift_align_information(epsilon, SWIFT_CACHE_ALIGNMENT);
swift_align_information(m, SWIFT_CACHE_ALIGNMENT);
swift_align_information(a_x, SWIFT_CACHE_ALIGNMENT);
swift_align_information(a_y, SWIFT_CACHE_ALIGNMENT);
......@@ -136,7 +154,7 @@ __attribute__((always_inline)) INLINE void gravity_cache_populate(struct gravity
swift_assume_size(gcount_padded, VEC_SIZE);
/* Fill the input caches */
for(int i = 0; i < gcount; ++i) {
for (int i = 0; i < gcount; ++i) {
x[i] = (float)(gparts[i].x[0] - shift[0]);
y[i] = (float)(gparts[i].x[1] - shift[1]);
z[i] = (float)(gparts[i].x[2] - shift[2]);
......@@ -144,24 +162,40 @@ __attribute__((always_inline)) INLINE void gravity_cache_populate(struct gravity
m[i] = gparts[i].mass;
}
#ifdef SWIFT_DEBUG_CHECKS
if (gcount_padded < gcount) error("Padded counter smaller than counter");
#endif
/* Pad the caches */
for(int i = gcount; i < gcount_padded; ++i) {
for (int i = gcount; i < gcount_padded; ++i) {
x[i] = 0.f;
y[i] = 0.f;
y[i] = 0.f;
z[i] = 0.f;
epsilon[i] = 0.f;
m[i] = 0.f;
}
/* Zero the output caches */
if(zero_output) {
if (zero_output) {
bzero(a_x, gcount_padded * sizeof(float));
bzero(a_y, gcount_padded * sizeof(float));
bzero(a_z, gcount_padded * sizeof(float));
}
}
__attribute__((always_inline)) INLINE void gravity_cache_populate_no_shift(struct gravity_cache *c, const struct gpart *restrict gparts, int gcount, int gcount_padded, int zero_output) {
/**
* @brief Fills a #gravity_cache structure with some #gpart.
*
* @param c The #gravity_cache to fill.
* @param gparts The #gpart array to read from.
* @param gcount The number of particles to read.
* @param gcount_padded The number of particle to read padded to the next
* multiple of the vector length.
* @param zero_output Do we need to zero the output caches ?
*/
__attribute__((always_inline)) INLINE void gravity_cache_populate_no_shift(
struct gravity_cache *c, const struct gpart *restrict gparts, int gcount,
int gcount_padded, int zero_output) {
/* Make the compiler understand we are in happy vectorization land */
float *restrict x = c->x;
......@@ -175,14 +209,14 @@ __attribute__((always_inline)) INLINE void gravity_cache_populate_no_shift(struc
swift_align_information(x, SWIFT_CACHE_ALIGNMENT);
swift_align_information(y, SWIFT_CACHE_ALIGNMENT);
swift_align_information(z, SWIFT_CACHE_ALIGNMENT);
swift_align_information(epsilon, SWIFT_CACHE_ALIGNMENT);
swift_align_information(epsilon, SWIFT_CACHE_ALIGNMENT);
swift_align_information(m, SWIFT_CACHE_ALIGNMENT);
swift_align_information(a_x, SWIFT_CACHE_ALIGNMENT);
swift_align_information(a_y, SWIFT_CACHE_ALIGNMENT);
swift_align_information(a_z, SWIFT_CACHE_ALIGNMENT);
/* Fill the input caches */
for(int i = 0; i < gcount; ++i) {
for (int i = 0; i < gcount; ++i) {
x[i] = (float)(gparts[i].x[0]);
y[i] = (float)(gparts[i].x[1]);
z[i] = (float)(gparts[i].x[2]);
......@@ -190,26 +224,36 @@ __attribute__((always_inline)) INLINE void gravity_cache_populate_no_shift(struc
m[i] = gparts[i].mass;
}
#ifdef SWIFT_DEBUG_CHECKS
if (gcount_padded < gcount) error("Padded counter smaller than counter");
#endif
/* Pad the caches */
for(int i = gcount; i < gcount_padded; ++i) {
for (int i = gcount; i < gcount_padded; ++i) {
x[i] = 0.f;
y[i] = 0.f;
y[i] = 0.f;
z[i] = 0.f;
epsilon[i] = 0.f;
m[i] = 0.f;
}
/* Zero the output caches */
if(zero_output) {
if (zero_output) {
bzero(a_x, gcount_padded * sizeof(float));
bzero(a_y, gcount_padded * sizeof(float));
bzero(a_z, gcount_padded * sizeof(float));
}
}
__attribute__((always_inline)) INLINE void gravity_cache_write_back(const struct gravity_cache *c,
struct gpart *restrict gparts, int gcount) {
/**
* @brief Write the output cache values back to the #gpart.
*
* @param c The #gravity_cache to read from.
* @param gparts The #gpart array to write to.
* @param gcount The number of particles to write.
*/
__attribute__((always_inline)) INLINE void gravity_cache_write_back(
const struct gravity_cache *c, struct gpart *restrict gparts, int gcount) {
/* Make the compiler understand we are in happy vectorization land */
float *restrict a_x = c->a_x;
......@@ -220,12 +264,11 @@ __attribute__((always_inline)) INLINE void gravity_cache_write_back(const struct
swift_align_information(a_z, SWIFT_CACHE_ALIGNMENT);
/* Write stuff back to the particles */
for(int i = 0; i < gcount; ++i) {
for (int i = 0; i < gcount; ++i) {
gparts[i].a_grav[0] += a_x[i];
gparts[i].a_grav[1] += a_y[i];
gparts[i].a_grav[2] += a_z[i];
}
}
#endif
#endif /* SWIFT_GRAVITY_CACHE_H */
src/runner.h
View file @ b2b4a66f
......@@ -50,13 +50,14 @@ struct runner {
/*! The engine owing this runner. */
struct engine *e;
#ifdef WITH_VECTORIZATION
/*! The particle gravity_cache of cell ci. */
struct gravity_cache ci_gravity_cache;
/*! The particle gravity_cache of cell cj. */
struct gravity_cache cj_gravity_cache;
#ifdef WITH_VECTORIZATION
/*! The particle cache of cell ci. */
struct cache ci_cache;
......
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