mirror of
https://gitlab.freedesktop.org/mesa/mesa.git
synced 2025-12-22 15:40:11 +01:00
freedreno/gmem: rework gmem layout algo
And try a bit harder to find an optimal layout. Improves on a sub- optimal layout we arrive at in the 4 MRT pass in manhattan, picking up a bit more than 3%. Signed-off-by: Rob Clark <robdclark@chromium.org> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4976>
This commit is contained in:
parent
c46f46befe
commit
1bd38746d5
1 changed files with 48 additions and 29 deletions
|
|
@ -172,13 +172,30 @@ div_align(unsigned num, unsigned denom, unsigned al)
|
||||||
return align(DIV_ROUND_UP(num, denom), al);
|
return align(DIV_ROUND_UP(num, denom), al);
|
||||||
}
|
}
|
||||||
|
|
||||||
static uint32_t
|
static bool
|
||||||
total_size(struct gmem_key *key, uint32_t bin_w, uint32_t bin_h,
|
layout_gmem(struct gmem_key *key, uint32_t nbins_x, uint32_t nbins_y,
|
||||||
struct fd_gmem_stateobj *gmem)
|
struct fd_gmem_stateobj *gmem)
|
||||||
{
|
{
|
||||||
|
struct fd_screen *screen = gmem->screen;
|
||||||
uint32_t gmem_align = key->gmem_page_align * 0x1000;
|
uint32_t gmem_align = key->gmem_page_align * 0x1000;
|
||||||
uint32_t total = 0, i;
|
uint32_t total = 0, i;
|
||||||
|
|
||||||
|
if ((nbins_x == 0) || (nbins_y == 0))
|
||||||
|
return false;
|
||||||
|
|
||||||
|
uint32_t bin_w, bin_h;
|
||||||
|
bin_w = div_align(key->width, nbins_x, screen->gmem_alignw);
|
||||||
|
bin_h = div_align(key->height, nbins_y, screen->gmem_alignh);
|
||||||
|
|
||||||
|
gmem->bin_w = bin_w;
|
||||||
|
gmem->bin_h = bin_h;
|
||||||
|
|
||||||
|
/* due to aligning bin_w/h, we could end up with one too
|
||||||
|
* many bins in either dimension, so recalculate:
|
||||||
|
*/
|
||||||
|
gmem->nbins_x = DIV_ROUND_UP(key->width, bin_w);
|
||||||
|
gmem->nbins_y = DIV_ROUND_UP(key->height, bin_h);
|
||||||
|
|
||||||
for (i = 0; i < MAX_RENDER_TARGETS; i++) {
|
for (i = 0; i < MAX_RENDER_TARGETS; i++) {
|
||||||
if (key->cbuf_cpp[i]) {
|
if (key->cbuf_cpp[i]) {
|
||||||
gmem->cbuf_base[i] = align(total, gmem_align);
|
gmem->cbuf_base[i] = align(total, gmem_align);
|
||||||
|
|
@ -196,7 +213,7 @@ total_size(struct gmem_key *key, uint32_t bin_w, uint32_t bin_h,
|
||||||
total = gmem->zsbuf_base[1] + key->zsbuf_cpp[1] * bin_w * bin_h;
|
total = gmem->zsbuf_base[1] + key->zsbuf_cpp[1] * bin_w * bin_h;
|
||||||
}
|
}
|
||||||
|
|
||||||
return total;
|
return total <= screen->gmemsize_bytes;
|
||||||
}
|
}
|
||||||
|
|
||||||
static struct fd_gmem_stateobj *
|
static struct fd_gmem_stateobj *
|
||||||
|
|
@ -209,28 +226,13 @@ gmem_stateobj_init(struct fd_screen *screen, struct gmem_key *key)
|
||||||
gmem->key = key;
|
gmem->key = key;
|
||||||
list_inithead(&gmem->node);
|
list_inithead(&gmem->node);
|
||||||
|
|
||||||
const uint32_t gmem_alignw = screen->gmem_alignw;
|
|
||||||
const uint32_t gmem_alignh = screen->gmem_alignh;
|
|
||||||
const unsigned npipes = screen->num_vsc_pipes;
|
const unsigned npipes = screen->num_vsc_pipes;
|
||||||
const uint32_t gmem_size = screen->gmemsize_bytes;
|
|
||||||
uint32_t nbins_x = 1, nbins_y = 1;
|
uint32_t nbins_x = 1, nbins_y = 1;
|
||||||
uint32_t bin_w, bin_h;
|
|
||||||
uint32_t max_width = bin_width(screen);
|
uint32_t max_width = bin_width(screen);
|
||||||
uint32_t i, j, t, xoff, yoff;
|
uint32_t i, j, t, xoff, yoff;
|
||||||
uint32_t tpp_x, tpp_y;
|
uint32_t tpp_x, tpp_y;
|
||||||
int tile_n[npipes];
|
int tile_n[npipes];
|
||||||
|
|
||||||
bin_w = div_align(key->width, 1, gmem_alignw);
|
|
||||||
bin_h = div_align(key->height, 1, gmem_alignh);
|
|
||||||
|
|
||||||
/* first, find a bin width that satisfies the maximum width
|
|
||||||
* restrictions:
|
|
||||||
*/
|
|
||||||
while (bin_w > max_width) {
|
|
||||||
nbins_x++;
|
|
||||||
bin_w = div_align(key->width, nbins_x, gmem_alignw);
|
|
||||||
}
|
|
||||||
|
|
||||||
if (fd_mesa_debug & FD_DBG_MSGS) {
|
if (fd_mesa_debug & FD_DBG_MSGS) {
|
||||||
debug_printf("binning input: cbuf cpp:");
|
debug_printf("binning input: cbuf cpp:");
|
||||||
for (i = 0; i < key->nr_cbufs; i++)
|
for (i = 0; i < key->nr_cbufs; i++)
|
||||||
|
|
@ -239,27 +241,44 @@ gmem_stateobj_init(struct fd_screen *screen, struct gmem_key *key)
|
||||||
key->zsbuf_cpp[0], key->width, key->height);
|
key->zsbuf_cpp[0], key->width, key->height);
|
||||||
}
|
}
|
||||||
|
|
||||||
|
/* first, find a bin width that satisfies the maximum width
|
||||||
|
* restrictions:
|
||||||
|
*/
|
||||||
|
while (div_align(key->width, nbins_x, screen->gmem_alignw) > max_width) {
|
||||||
|
nbins_x++;
|
||||||
|
}
|
||||||
|
|
||||||
/* then find a bin width/height that satisfies the memory
|
/* then find a bin width/height that satisfies the memory
|
||||||
* constraints:
|
* constraints:
|
||||||
*/
|
*/
|
||||||
while (total_size(key, bin_w, bin_h, gmem) > gmem_size) {
|
while (!layout_gmem(key, nbins_x, nbins_y, gmem)) {
|
||||||
if (bin_w > bin_h) {
|
if (nbins_y > nbins_x) {
|
||||||
nbins_x++;
|
nbins_x++;
|
||||||
bin_w = div_align(key->width, nbins_x, gmem_alignw);
|
|
||||||
} else {
|
} else {
|
||||||
nbins_y++;
|
nbins_y++;
|
||||||
bin_h = div_align(key->height, nbins_y, gmem_alignh);
|
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
DBG("using %d bins of size %dx%d", nbins_x*nbins_y, bin_w, bin_h);
|
/* Lets see if we can tweak the layout a bit and come up with
|
||||||
|
* something better:
|
||||||
|
*/
|
||||||
|
if ((((nbins_x - 1) * (nbins_y + 1)) < (nbins_x * nbins_y)) &&
|
||||||
|
layout_gmem(key, nbins_x - 1, nbins_y + 1, gmem)) {
|
||||||
|
nbins_x--;
|
||||||
|
nbins_y++;
|
||||||
|
} else if ((((nbins_x + 1) * (nbins_y - 1)) < (nbins_x * nbins_y)) &&
|
||||||
|
layout_gmem(key, nbins_x + 1, nbins_y - 1, gmem)) {
|
||||||
|
nbins_x++;
|
||||||
|
nbins_y--;
|
||||||
|
}
|
||||||
|
|
||||||
|
layout_gmem(key, nbins_x, nbins_y, gmem);
|
||||||
|
|
||||||
|
DBG("using %d bins of size %dx%d", gmem->nbins_x * gmem->nbins_y,
|
||||||
|
gmem->bin_w, gmem->bin_h);
|
||||||
|
|
||||||
memcpy(gmem->cbuf_cpp, key->cbuf_cpp, sizeof(key->cbuf_cpp));
|
memcpy(gmem->cbuf_cpp, key->cbuf_cpp, sizeof(key->cbuf_cpp));
|
||||||
memcpy(gmem->zsbuf_cpp, key->zsbuf_cpp, sizeof(key->zsbuf_cpp));
|
memcpy(gmem->zsbuf_cpp, key->zsbuf_cpp, sizeof(key->zsbuf_cpp));
|
||||||
gmem->bin_h = bin_h;
|
|
||||||
gmem->bin_w = bin_w;
|
|
||||||
gmem->nbins_x = nbins_x;
|
|
||||||
gmem->nbins_y = nbins_y;
|
|
||||||
gmem->minx = key->minx;
|
gmem->minx = key->minx;
|
||||||
gmem->miny = key->miny;
|
gmem->miny = key->miny;
|
||||||
gmem->width = key->width;
|
gmem->width = key->width;
|
||||||
|
|
@ -348,7 +367,7 @@ gmem_stateobj_init(struct fd_screen *screen, struct gmem_key *key)
|
||||||
xoff = key->minx;
|
xoff = key->minx;
|
||||||
|
|
||||||
/* clip bin height: */
|
/* clip bin height: */
|
||||||
bh = MIN2(bin_h, key->miny + key->height - yoff);
|
bh = MIN2(gmem->bin_h, key->miny + key->height - yoff);
|
||||||
|
|
||||||
for (j = 0; j < nbins_x; j++) {
|
for (j = 0; j < nbins_x; j++) {
|
||||||
struct fd_tile *tile = &gmem->tile[t];
|
struct fd_tile *tile = &gmem->tile[t];
|
||||||
|
|
@ -361,7 +380,7 @@ gmem_stateobj_init(struct fd_screen *screen, struct gmem_key *key)
|
||||||
assert(p < gmem->num_vsc_pipes);
|
assert(p < gmem->num_vsc_pipes);
|
||||||
|
|
||||||
/* clip bin width: */
|
/* clip bin width: */
|
||||||
bw = MIN2(bin_w, key->minx + key->width - xoff);
|
bw = MIN2(gmem->bin_w, key->minx + key->width - xoff);
|
||||||
tile->n = !is_a20x(screen) ? tile_n[p]++ :
|
tile->n = !is_a20x(screen) ? tile_n[p]++ :
|
||||||
((i % tpp_y + 1) << 3 | (j % tpp_x + 1));
|
((i % tpp_y + 1) << 3 | (j % tpp_x + 1));
|
||||||
tile->p = p;
|
tile->p = p;
|
||||||
|
|
|
||||||
Loading…
Add table
Reference in a new issue