mirror of
https://github.com/shadps4-emu/shadPS4.git
synced 2025-08-05 17:02:40 +00:00
1f9ac53c28
Some checks are pending
Build and Release / reuse (push) Waiting to run
Build and Release / clang-format (push) Waiting to run
Build and Release / get-info (push) Waiting to run
Build and Release / windows-sdl (push) Blocked by required conditions
Build and Release / windows-qt (push) Blocked by required conditions
Build and Release / macos-sdl (push) Blocked by required conditions
Build and Release / macos-qt (push) Blocked by required conditions
Build and Release / linux-sdl (push) Blocked by required conditions
Build and Release / linux-qt (push) Blocked by required conditions
Build and Release / linux-sdl-gcc (push) Blocked by required conditions
Build and Release / linux-qt-gcc (push) Blocked by required conditions
Build and Release / pre-release (push) Blocked by required conditions
* control_flow_graph: Improve divergence handling * recompiler: Simplify optimization passes Removes a redudant constant propagation and cleans up the passes a little * ir_passes: Add new readlane elimination pass The algorithm has grown complex enough where it deserves its own pass. The old implementation could only handle a single phi level properly, however this one should be able to eliminate vast majority of lane cases remaining. It first performs a traversal of the phi tree to ensure that all phi sources can be rewritten into an expected value and then performs elimintation by recursively duplicating the phi nodes at each step, in order to preserve control flow. * clang format * control_flow_graph: Remove debug code
89 lines
3.7 KiB
C++
89 lines
3.7 KiB
C++
// SPDX-FileCopyrightText: Copyright 2024 shadPS4 Emulator Project
|
|
// SPDX-License-Identifier: GPL-2.0-or-later
|
|
|
|
#include "shader_recompiler/frontend/control_flow_graph.h"
|
|
#include "shader_recompiler/frontend/decode.h"
|
|
#include "shader_recompiler/frontend/structured_control_flow.h"
|
|
#include "shader_recompiler/ir/passes/ir_passes.h"
|
|
#include "shader_recompiler/ir/post_order.h"
|
|
#include "shader_recompiler/recompiler.h"
|
|
|
|
namespace Shader {
|
|
|
|
IR::BlockList GenerateBlocks(const IR::AbstractSyntaxList& syntax_list) {
|
|
size_t num_syntax_blocks{};
|
|
for (const auto& node : syntax_list) {
|
|
if (node.type == IR::AbstractSyntaxNode::Type::Block) {
|
|
++num_syntax_blocks;
|
|
}
|
|
}
|
|
IR::BlockList blocks;
|
|
blocks.reserve(num_syntax_blocks);
|
|
u32 order_index{};
|
|
for (const auto& node : syntax_list) {
|
|
if (node.type == IR::AbstractSyntaxNode::Type::Block) {
|
|
blocks.push_back(node.data.block);
|
|
}
|
|
}
|
|
return blocks;
|
|
}
|
|
|
|
IR::Program TranslateProgram(std::span<const u32> code, Pools& pools, Info& info,
|
|
RuntimeInfo& runtime_info, const Profile& profile) {
|
|
// Ensure first instruction is expected.
|
|
constexpr u32 token_mov_vcchi = 0xBEEB03FF;
|
|
if (code[0] != token_mov_vcchi) {
|
|
LOG_WARNING(Render_Recompiler, "First instruction is not s_mov_b32 vcc_hi, #imm");
|
|
}
|
|
|
|
Gcn::GcnCodeSlice slice(code.data(), code.data() + code.size());
|
|
Gcn::GcnDecodeContext decoder;
|
|
|
|
// Decode and save instructions
|
|
IR::Program program{info};
|
|
program.ins_list.reserve(code.size());
|
|
while (!slice.atEnd()) {
|
|
program.ins_list.emplace_back(decoder.decodeInstruction(slice));
|
|
}
|
|
|
|
// Clear any previous pooled data.
|
|
pools.ReleaseContents();
|
|
|
|
// Create control flow graph
|
|
Common::ObjectPool<Gcn::Block> gcn_block_pool{64};
|
|
Gcn::CFG cfg{gcn_block_pool, program.ins_list};
|
|
|
|
// Structurize control flow graph and create program.
|
|
program.syntax_list = Shader::Gcn::BuildASL(pools.inst_pool, pools.block_pool, cfg,
|
|
program.info, runtime_info, profile);
|
|
program.blocks = GenerateBlocks(program.syntax_list);
|
|
program.post_order_blocks = Shader::IR::PostOrder(program.syntax_list.front());
|
|
|
|
// Run optimization passes
|
|
Shader::Optimization::SsaRewritePass(program.post_order_blocks);
|
|
Shader::Optimization::ConstantPropagationPass(program.post_order_blocks);
|
|
Shader::Optimization::IdentityRemovalPass(program.blocks);
|
|
if (info.l_stage == LogicalStage::TessellationControl) {
|
|
Shader::Optimization::TessellationPreprocess(program, runtime_info);
|
|
Shader::Optimization::HullShaderTransform(program, runtime_info);
|
|
} else if (info.l_stage == LogicalStage::TessellationEval) {
|
|
Shader::Optimization::TessellationPreprocess(program, runtime_info);
|
|
Shader::Optimization::DomainShaderTransform(program, runtime_info);
|
|
}
|
|
Shader::Optimization::RingAccessElimination(program, runtime_info);
|
|
Shader::Optimization::ReadLaneEliminationPass(program);
|
|
Shader::Optimization::FlattenExtendedUserdataPass(program);
|
|
Shader::Optimization::ResourceTrackingPass(program);
|
|
Shader::Optimization::LowerBufferFormatToRaw(program);
|
|
Shader::Optimization::SharedMemoryToStoragePass(program, runtime_info, profile);
|
|
Shader::Optimization::SharedMemoryBarrierPass(program, runtime_info, profile);
|
|
Shader::Optimization::IdentityRemovalPass(program.blocks);
|
|
Shader::Optimization::DeadCodeEliminationPass(program);
|
|
Shader::Optimization::ConstantPropagationPass(program.post_order_blocks);
|
|
Shader::Optimization::CollectShaderInfoPass(program);
|
|
|
|
return program;
|
|
}
|
|
|
|
} // namespace Shader
|