diff options
| author | Eric Anholt <[email protected]> | 2016-11-30 11:52:06 -0800 |
|---|---|---|
| committer | Eric Anholt <[email protected]> | 2016-11-30 19:58:09 -0800 |
| commit | d182740ac8bf2fc3aefe331235bfe72bbdd9987d (patch) | |
| tree | dea5aa69162b710cb1fc1d2ab7577e0e5fd4acc1 /src/gallium/drivers | |
| parent | 1f9daf7cd1cf3d31571776a0142e9e339168a8df (diff) | |
vc4: Improve interleaving of texture coordinates vs results.
The latency_between was trying to handle the delay between the coordinate
write ("before") and the corresponding sample read ("after"), but we were
handing in the two instructions swapped.
This meant that we tried to fit things between a tex_s and its *preceding*
tex_result. This made us only interleave normal texture coordinates by
accident, and pessimized UBO reads by pushing the tex_result collection
earlier until there was nothing but it (and then its preceding coordinate
setup) left.
In addition to latency reduction, things end up packing better (probably
due to reduced live ranges of the texture results):
total instructions in shared programs: 98121 -> 94775 (-3.41%)
instructions in affected programs: 91196 -> 87850 (-3.67%)
3DMMES performance +1.15569% +/- 0.124714% (n=8,10)
Diffstat (limited to 'src/gallium/drivers')
| -rw-r--r-- | src/gallium/drivers/vc4/vc4_qir_schedule.c | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/src/gallium/drivers/vc4/vc4_qir_schedule.c b/src/gallium/drivers/vc4/vc4_qir_schedule.c index a8ef189e583..ea48a858337 100644 --- a/src/gallium/drivers/vc4/vc4_qir_schedule.c +++ b/src/gallium/drivers/vc4/vc4_qir_schedule.c @@ -593,7 +593,7 @@ compute_delay(struct schedule_node *n) compute_delay(n->children[i]); n->delay = MAX2(n->delay, n->children[i]->delay + - latency_between(n, n->children[i])); + latency_between(n->children[i], n)); } } } @@ -644,8 +644,8 @@ schedule_instructions(struct vc4_compile *c, child->unblocked_time = MAX2(child->unblocked_time, state->time + - latency_between(chosen, - child)); + latency_between(child, + chosen)); child->parent_count--; if (child->parent_count == 0) list_add(&child->link, &state->worklist); |