nv50/ir: optimmize shl(a, 0) to a
PATCH 1/1 nv50/ir: optimmize shl(a, 0) to a
helps two alien isolation shaders
shader-db:
total instructions in shared programs : 4251497 -> 4251494 (-0.00%)
total gprs used in shared programs : 513962 -> 513962 (0.00%) total local used in shared programs : 29797 -> 29797 (0.00%) total bytes used in shared programs : 38960264 -> 38960232 (-0.00%)
local gpr inst bytes
helped 0 0 2 2
hurt 0 0 0 0
--- src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp | 5 +++++ 1 file changed, 5 insertions(+)
src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp +5 -0
@@ -1284,6 +1284,11 @@ ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s) case OP_SHL: { + if (s == 1 && imm0.isInteger(0)) { + i->op = OP_MOV; + i->setSrc(1, NULL); + break; + } if (s != 1 || i->src(0).mod != Modifier(0)) break; // try to concatenate shifts
Signature · 6 lines
2.12.2 _______________________________________________ mesa-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/mesa-dev
A message that is not in the archive (sent off-list or not yet received): 20170429164604.3263-1-karolherbst-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org
Samuel Pitoiset
Reviewed-by"optimmize" ? No need to resend just for that though.
Quoted 24 lines · Karol Herbst
helps two alien isolation shaders
shader-db:
total instructions in shared programs : 4251497 -> 4251494 (-0.00%)total gprs used in shared programs : 513962 -> 513962 (0.00%) total local used in shared programs : 29797 -> 29797 (0.00%) total bytes used in shared programs : 38960264 -> 38960232 (-0.00%)local gpr inst bytes
helped 0 0 2 2
hurt 0 0 0 0Signed-off-by Karol Herbst <[email protected]>--- src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp | 5 +++++ 1 file changed, 5 insertions(+)src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp +0
-0@@ -1284,6 +1284,11 @@ ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s)
src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s) · line 1294
Quoted 10 lines
src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp +5
-0case OP_SHL: { + if (s == 1 && imm0.isInteger(0)) { + i->op = OP_MOV; + i->setSrc(1, NULL); + break; + } if (s != 1 || i->src(0).mod != Modifier(0)) break; // try to concatenate shifts
_______________________________________________
Nouveau mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/nouveau
Ilia Mirkin
Reviewed-byQuoted 22 lines · Karol Herbst
helps two alien isolation shaders
shader-db:
total instructions in shared programs : 4251497 -> 4251494 (-0.00%)total gprs used in shared programs : 513962 -> 513962 (0.00%) total local used in shared programs : 29797 -> 29797 (0.00%) total bytes used in shared programs : 38960264 -> 38960232 (-0.00%)local gpr inst bytes
helped 0 0 2 2
hurt 0 0 0 0Signed-off-by Karol Herbst <[email protected]>--- src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp | 5 +++++ 1 file changed, 5 insertions(+)src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp +0
-0
src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s) · line 1293
Quoted 10 lines
src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp +5
-0@@ -1284,6 +1284,11 @@ ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s) case OP_SHL: { + if (s == 1 && imm0.isInteger(0)) { + i->op = OP_MOV; + i->setSrc(1, NULL); + break; + } if (s != 1 || i->src(0).mod != Modifier(0)) break;
Interesting. This suggests that src(0) may have modifiers, although I can't imagine what that'd be. Perhaps it can take a neg? In that case, you need to fix the OP_MOV above -- should probably do it anyways, i.e. i->op = i->src(0).mod.getOp() or something. With that, this is
Separately, I just noticed that we don't appear to have any 0 << x or 0 >> y or EXTBF or anything like that. I doubt it comes up too often though.
Quoted 4 lines
src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp +0
-1// try to concatenate shifts --2.12.2
_______________________________________________
Nouveau mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/nouveau
2017-04-29 21:03 GMT+02:00 Ilia Mirkin <[email protected]>:
Quoted 24 lines · Karol Herbst
Quoted 23 lines
helps two alien isolation shaders
shader-db:
total instructions in shared programs : 4251497 -> 4251494 (-0.00%)total gprs used in shared programs : 513962 -> 513962 (0.00%) total local used in shared programs : 29797 -> 29797 (0.00%) total bytes used in shared programs : 38960264 -> 38960232 (-0.00%)local gpr inst bytes
helped 0 0 2 2
hurt 0 0 0 0Signed-off-by Karol Herbst <[email protected]>--- src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp | 5 +++++ 1 file changed, 5 insertions(+)src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp +0
-0@@ -1284,6 +1284,11 @@ ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s)
Quoted 15 lines
src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s) · line 1293
Quoted 9 lines
src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp +5
-0case OP_SHL: { + if (s == 1 && imm0.isInteger(0)) { + i->op = OP_MOV; + i->setSrc(1, NULL); + break; + } if (s != 1 || i->src(0).mod != Modifier(0)) break;Interesting. This suggests that src(0) may have modifiers, although I can't imagine what that'd be. Perhaps it can take a neg? In that case, you need to fix the
OP_MOVabove -- should probably do it anyways, i.e.i->op= i->src(0).mod.getOp() or something. With that, this is
I looked through the emitter and there seem to be no flags. Envydis and nvdisasm seem to think the same.
Will fix it up nethertheless.
Quoted 10 lines
Reviewed-by Ilia Mirkin <[email protected]>Separately, I just noticed that we don't appear to have any 0 << x or 0 >> y or EXTBF or anything like that. I doubt it comes up too often though.
Quoted 4 lines
src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp +0
-1// try to concatenate shifts --2.12.2
_______________________________________________
mesa-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/mesa-dev