0ba40d4ccf
Ports the existing DAG combines, minus the simplify demanded bits which seems to have no equivalent now. Without these, this isn't particularly helpful in most of the IR sample cases.
65 lines
2.6 KiB
TableGen
65 lines
2.6 KiB
TableGen
//=- AMDGPUCombine.td - Define AMDGPU Combine Rules ----------*- tablegen -*-=//
|
|
//
|
|
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
|
|
// See https://llvm.org/LICENSE.txt for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
include "llvm/Target/GlobalISel/Combine.td"
|
|
|
|
// TODO: This really belongs after legalization after scalarization.
|
|
// TODO: GICombineRules should accept subtarget predicates
|
|
|
|
def fmin_fmax_legacy_matchdata : GIDefMatchData<"FMinFMaxLegacyInfo">;
|
|
|
|
def fcmp_select_to_fmin_fmax_legacy : GICombineRule<
|
|
(defs root:$select, fmin_fmax_legacy_matchdata:$matchinfo),
|
|
(match (wip_match_opcode G_SELECT):$select,
|
|
[{ return matchFMinFMaxLegacy(*${select}, MRI, *MF, ${matchinfo}); }]),
|
|
(apply [{ applySelectFCmpToFMinToFMaxLegacy(*${select}, ${matchinfo}); }])>;
|
|
|
|
|
|
def uchar_to_float : GICombineRule<
|
|
(defs root:$itofp),
|
|
(match (wip_match_opcode G_UITOFP, G_SITOFP):$itofp,
|
|
[{ return matchUCharToFloat(*${itofp}, MRI, *MF, Helper); }]),
|
|
(apply [{ applyUCharToFloat(*${itofp}); }])>;
|
|
|
|
def cvt_f32_ubyteN_matchdata : GIDefMatchData<"CvtF32UByteMatchInfo">;
|
|
|
|
def cvt_f32_ubyteN : GICombineRule<
|
|
(defs root:$cvt_f32_ubyteN, cvt_f32_ubyteN_matchdata:$matchinfo),
|
|
(match (wip_match_opcode G_AMDGPU_CVT_F32_UBYTE0,
|
|
G_AMDGPU_CVT_F32_UBYTE1,
|
|
G_AMDGPU_CVT_F32_UBYTE2,
|
|
G_AMDGPU_CVT_F32_UBYTE3):$cvt_f32_ubyteN,
|
|
[{ return matchCvtF32UByteN(*${cvt_f32_ubyteN}, MRI, *MF, ${matchinfo}); }]),
|
|
(apply [{ applyCvtF32UByteN(*${cvt_f32_ubyteN}, ${matchinfo}); }])>;
|
|
|
|
// Combines which should only apply on SI/VI
|
|
def gfx6gfx7_combines : GICombineGroup<[fcmp_select_to_fmin_fmax_legacy]>;
|
|
|
|
|
|
def AMDGPUPreLegalizerCombinerHelper: GICombinerHelper<
|
|
"AMDGPUGenPreLegalizerCombinerHelper", [all_combines,
|
|
elide_br_by_inverting_cond]> {
|
|
let DisableRuleOption = "amdgpuprelegalizercombiner-disable-rule";
|
|
}
|
|
|
|
|
|
// FIXME: combines_for_extload can introduce illegal extloads which
|
|
// aren't re-legalized.
|
|
// FIXME: Is there a way to remove a single item from all_combines?
|
|
def all_combines_minus_extload : GICombineGroup<[trivial_combines,
|
|
ptr_add_immed_chain, combine_indexed_load_store, undef_combines,
|
|
identity_combines]
|
|
>;
|
|
|
|
def AMDGPUPostLegalizerCombinerHelper: GICombinerHelper<
|
|
"AMDGPUGenPostLegalizerCombinerHelper",
|
|
[all_combines_minus_extload, gfx6gfx7_combines,
|
|
uchar_to_float, cvt_f32_ubyteN]> {
|
|
let DisableRuleOption = "amdgpupostlegalizercombiner-disable-rule";
|
|
}
|