diff options
author | Iago Toral Quiroga <[email protected]> | 2017-09-27 11:36:31 +0200 |
---|---|---|
committer | Iago Toral Quiroga <[email protected]> | 2017-09-29 08:18:25 +0200 |
commit | 47e527bd81f21cd18333fa7923d8ae8fd3a74d2c (patch) | |
tree | 65f1f2f88f1c8c1dbad02de21d8532bf7733f2da /src/intel/compiler/brw_fs.cpp | |
parent | 2df897cf1ffd5ae01ecdbb66195d292a2a15df91 (diff) |
i965/fs: force pull model for 64-bit GS inputs
Triggering the push model when 64-bit inputs are involved is not easy due to
the constrains on the maximum number of registers that we allow for this mode,
however, for GS with 'points' primitive type and just a couple of double
varyings we can trigger this and it just doesn't work because the
implementation is not 64-bit aware at all. For now, let's make sure that we
don't attempt this model whith 64-bit inputs and we always fall back to pull
model for them.
Also, don't enable the VUE handles in the thread payload on the fly when we
find an input for which we need the pull model, this is not safe: if we need
to resort to the pull model we need to account for that when we setup the
thread payload so we compute the first non-payload register properly. If we
didn't do that correctly and we enable it on-the-fly here then we will end up
VUE handles on the first non-payload register which will probably lead to
GPU hangs. Instead, always enable the VUE handles for the pull model so we
can safely use them when needed. The GS is going to resort to pull model
almost in every situation anyway, so this shouldn't make a significant
difference and it makes things easier and safer.
v2: Always enable the VUE handles for pull model, this is easier and safer
and the GS is going to fallback to pull model almost always anyway (Ken)
v3: Only clamp the URB read length if we are over the maximum reserved for
push inputs as we were doing in the original code (Ken).
v4: No need to clamp the urb read length if invocations > 1
Reviewed-by: Kenneth Graunke <[email protected]>
Diffstat (limited to 'src/intel/compiler/brw_fs.cpp')
-rw-r--r-- | src/intel/compiler/brw_fs.cpp | 18 |
1 files changed, 12 insertions, 6 deletions
diff --git a/src/intel/compiler/brw_fs.cpp b/src/intel/compiler/brw_fs.cpp index eb9b4c38909..e33cb0e1186 100644 --- a/src/intel/compiler/brw_fs.cpp +++ b/src/intel/compiler/brw_fs.cpp @@ -5602,6 +5602,17 @@ fs_visitor::setup_gs_payload() payload.num_regs++; } + /* Always enable VUE handles so we can safely use pull model if needed. + * + * The push model for a GS uses a ton of register space even for trivial + * scenarios with just a few inputs, so just make things easier and a bit + * safer by always having pull model available. + */ + gs_prog_data->base.include_vue_handles = true; + + /* R3..RN: ICP Handles for each incoming vertex (when using pull model) */ + payload.num_regs += nir->info.gs.vertices_in; + /* Use a maximum of 24 registers for push-model inputs. */ const unsigned max_push_components = 24; @@ -5612,12 +5623,7 @@ fs_visitor::setup_gs_payload() * have to multiply by VerticesIn to obtain the total storage requirement. */ if (8 * vue_prog_data->urb_read_length * nir->info.gs.vertices_in > - max_push_components || gs_prog_data->invocations > 1) { - gs_prog_data->base.include_vue_handles = true; - - /* R3..RN: ICP Handles for each incoming vertex (when using pull model) */ - payload.num_regs += nir->info.gs.vertices_in; - + max_push_components) { vue_prog_data->urb_read_length = ROUND_DOWN_TO(max_push_components / nir->info.gs.vertices_in, 8) / 8; } |