[Mesa-dev] [PATCH 2/2] nv50/ir: optimize sub(a, 0) to a

[Karol Herbst]

helps a 0ad shader in default shader-db:
total instructions in shared programs : 102863 -> 102861 (-0.00%)
total gprs used in shared programs    : 14725 -> 14725 (0.00%)
total local used in shared programs   : 129 -> 129 (0.00%)
total bytes used in shared programs   : 943944 -> 943928 (-0.00%)

                local        gpr       inst      bytes
    helped           0           0           1           1
      hurt           0           0           0           0

helps some borderlands shaders:
total instructions in shared programs : 1794261 -> 1794255 (-0.00%)
total gprs used in shared programs    : 233059 -> 233059 (0.00%)
total local used in shared programs   : 5569 -> 5569 (0.00%)
total bytes used in shared programs   : 16451664 -> 16451616 (-0.00%)

                local        gpr       inst      bytes
    helped           0           0           3           3
      hurt           0           0           0           0

Signed-off-by: Karol Herbst <nouveau at karolherbst.de>
---
 src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp b/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp
index 5c3ee37..fa07dff 100644
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp
@@ -1003,12 +1003,15 @@ ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s)
       }
       break;
    case OP_ADD:
+   case OP_SUB:
       if (i->usesFlags())
          break;
       if (imm0.isInteger(0)) {
          if (s == 0) {
             i->setSrc(0, i->getSrc(1));
             i->src(0).mod = i->src(1).mod;
+            if (i->op == OP_SUB)
+               i->src(0).mod = i->src(0).mod ^ Modifier(NV50_IR_MOD_NEG);
          }
          i->setSrc(1, NULL);
          i->op = i->src(0).mod.getOp();
-- 
2.7.1