intel/fs: Use nir_lower_interpolation on gen11+

On gen11, the removed the PLN instruction so we have to emit a pile of
MAD to emulate it.  We may as well do that in NIR so we can optimize and
later schedule it.

Shader-db results on Ice Lake:

    total instructions in shared programs: 17145644 -> 16556440 (-3.44%)
    instructions in affected programs: 11507454 -> 10918250 (-5.12%)
    helped: 35763
    HURT: 42085
    helped stats (abs) min: 1 max: 140 x̄: 19.09 x̃: 18
    helped stats (rel) min: 0.04% max: 37.93% x̄: 15.40% x̃: 14.49%
    HURT stats (abs)   min: 1 max: 248 x̄: 2.22 x̃: 2
    HURT stats (rel)   min: 0.05% max: 50.00% x̄: 5.00% x̃: 2.47%
    95% mean confidence interval for instructions value: -7.67 -7.47
    95% mean confidence interval for instructions %-change: -4.46% -4.29%
    Instructions are helped.

    total loops in shared programs: 4370 -> 4370 (0.00%)
    loops in affected programs: 0 -> 0
    helped: 0
    HURT: 0

    total cycles in shared programs: 360624645 -> 368220857 (2.11%)
    cycles in affected programs: 269631244 -> 277227456 (2.82%)
    helped: 15583
    HURT: 65874
    helped stats (abs) min: 1 max: 28561 x̄: 78.45 x̃: 32
    helped stats (rel) min: <.01% max: 67.81% x̄: 5.38% x̃: 2.44%
    HURT stats (abs)   min: 1 max: 238638 x̄: 133.87 x̃: 20
    HURT stats (rel)   min: <.01% max: 306.25% x̄: 5.81% x̃: 3.97%
    95% mean confidence interval for cycles value: 67.42 119.09
    95% mean confidence interval for cycles %-change: 3.61% 3.73%
    Cycles are HURT.

    total spills in shared programs: 8943 -> 8981 (0.42%)
    spills in affected programs: 1925 -> 1963 (1.97%)
    helped: 44
    HURT: 14

    total fills in shared programs: 21815 -> 21925 (0.50%)
    fills in affected programs: 3511 -> 3621 (3.13%)
    helped: 41
    HURT: 18

    LOST:   70
    GAINED: 14

Reviewed-by: Matt Turner <mattst88@gmail.com>

diff --git a/src/intel/compiler/brw_fs.cpp b/src/intel/compiler/brw_fs.cpp
index 36d8191..bd0ac1a 100644 (file)
--- a/src/intel/compiler/brw_fs.cpp
+++ b/src/intel/compiler/brw_fs.cpp
@@ -3858,47 +3858,6 @@ fs_visitor::lower_load_payload()
 }
 
 bool
-fs_visitor::lower_linterp()
-{
-   bool progress = false;
-
-   if (devinfo->gen < 11)
-      return false;
-
-   foreach_block_and_inst_safe(block, fs_inst, inst, cfg) {
-      const fs_builder ibld(this, block, inst);
-
-      if (inst->opcode != FS_OPCODE_LINTERP)
-         continue;
-
-      fs_reg dwP = component(inst->src[1], 0);
-      fs_reg dwQ = component(inst->src[1], 1);
-      fs_reg dwR = component(inst->src[1], 3);
-      for (unsigned i = 0; i < DIV_ROUND_UP(dispatch_width, 8); i++) {
-         const fs_builder hbld(ibld.half(i));
-         fs_reg dst = half(inst->dst, i);
-         fs_reg delta_xy = offset(inst->src[0], ibld, i);
-         hbld.MAD(dst, dwR, half(delta_xy, 0), dwP);
-         fs_inst *mad = hbld.MAD(dst, dst, half(delta_xy, 1), dwQ);
-
-         /* Propagate conditional mod and saturate from the original
-          * instruction to the second MAD instruction.
-          */
-         set_saturate(inst->saturate, mad);
-         set_condmod(inst->conditional_mod, mad);
-      }
-
-      inst->remove(block);
-      progress = true;
-   }
-
-   if (progress)
-      invalidate_live_intervals();
-
-   return progress;
-}
-
-bool
 fs_visitor::lower_integer_multiplication()
 {
    bool progress = false;
@@ -7095,11 +7054,6 @@ fs_visitor::optimize()
       OPT(compact_virtual_grfs);
    } while (progress);
 
-   if (OPT(lower_linterp)) {
-      OPT(opt_copy_propagation);
-      OPT(dead_code_eliminate);
-   }
-
    /* Do this after cmod propagation has had every possible opportunity to
     * propagate results into SEL instructions.
     */

diff --git a/src/intel/compiler/brw_fs.h b/src/intel/compiler/brw_fs.h
index 230f9ab..d93fb29 100644 (file)
--- a/src/intel/compiler/brw_fs.h
+++ b/src/intel/compiler/brw_fs.h
@@ -163,7 +163,6 @@ public:
    bool lower_pack();
    bool lower_regioning();
    bool lower_logical_sends();
-   bool lower_linterp();
    bool lower_integer_multiplication();
    bool lower_minmax();
    bool lower_simd_width();

diff --git a/src/intel/compiler/brw_fs_generator.cpp b/src/intel/compiler/brw_fs_generator.cpp
index d068d1a..be6a00e 100644 (file)
--- a/src/intel/compiler/brw_fs_generator.cpp
+++ b/src/intel/compiler/brw_fs_generator.cpp
@@ -822,7 +822,7 @@ fs_generator::generate_linterp(fs_inst *inst,
    struct brw_reg interp = stride(src[1], 0, 1, 0);
    brw_inst *i[2];
 
-   /* fs_visitor::lower_linterp() will do the lowering to MAD instructions for
+   /* nir_lower_interpolation() will do the lowering to MAD instructions for
     * us on gen11+
     */
    assert(devinfo->gen < 11);

diff --git a/src/intel/compiler/brw_nir.c b/src/intel/compiler/brw_nir.c
index c96e030..4dea234 100644 (file)
--- a/src/intel/compiler/brw_nir.c
+++ b/src/intel/compiler/brw_nir.c
@@ -452,6 +452,8 @@ brw_nir_lower_fs_inputs(nir_shader *nir,
       lower_io_options |= nir_lower_io_force_sample_interpolation;
 
    nir_lower_io(nir, nir_var_shader_in, type_size_vec4, lower_io_options);
+   if (devinfo->gen >= 11)
+      nir_lower_interpolation(nir, ~0);
 
    /* This pass needs actual constants */
    nir_opt_constant_folding(nir);