artemist/yuzu
1
0
Fork 0
Code Issues Pull requests Packages Projects Releases Wiki Activity

shader: Constant propagation and global memory to storage buffer

Browse source
This commit is contained in:
ReinUsesLisp 2021-02-05 05:58:02 -03:00 committed by ameerj
parent d24a16045f
commit e81739493a
17 changed files with 652 additions and 63 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
src/shader_recompiler/CMakeLists.txt
View file

@ -59,7 +59,9 @@ add_executable(shader_recompiler
frontend/maxwell/translate/impl/move_special_register.cpp
frontend/maxwell/translate/translate.cpp
frontend/maxwell/translate/translate.h
ir_opt/constant_propagation_pass.cpp
ir_opt/dead_code_elimination_pass.cpp
ir_opt/global_memory_to_storage_buffer_pass.cpp
ir_opt/identity_removal_pass.cpp
ir_opt/passes.h
ir_opt/ssa_rewrite_pass.cpp

20
src/shader_recompiler/frontend/ir/ir_emitter.cpp
View file

@ -504,6 +504,20 @@ U32U64 IREmitter::IAdd(const U32U64& a, const U32U64& b) {
}
}
U32U64 IREmitter::ISub(const U32U64& a, const U32U64& b) {
if (a.Type() != b.Type()) {
throw InvalidArgument("Mismatching types {} and {}", a.Type(), b.Type());
}
switch (a.Type()) {
case Type::U32:
return Inst<U32>(Opcode::ISub32, a, b);
case Type::U64:
return Inst<U64>(Opcode::ISub64, a, b);
default:
ThrowInvalidType(a.Type());
}
}
U32 IREmitter::IMul(const U32& a, const U32& b) {
return Inst<U32>(Opcode::IMul32, a, b);
}
@ -679,8 +693,8 @@ U32U64 IREmitter::ConvertFToI(size_t bitsize, bool is_signed, const U16U32U64& v
}
}
U32U64 IREmitter::ConvertU(size_t bitsize, const U32U64& value) {
switch (bitsize) {
U32U64 IREmitter::ConvertU(size_t result_bitsize, const U32U64& value) {
switch (result_bitsize) {
case 32:
switch (value.Type()) {
case Type::U32:
@ -703,7 +717,7 @@ U32U64 IREmitter::ConvertU(size_t bitsize, const U32U64& value) {
break;
}
}
throw NotImplementedException("Conversion from {} to {} bits", value.Type(), bitsize);
throw NotImplementedException("Conversion from {} to {} bits", value.Type(), result_bitsize);
}
} // namespace Shader::IR

5
src/shader_recompiler/frontend/ir/ir_emitter.h
View file

@ -17,6 +17,8 @@ namespace Shader::IR {
class IREmitter {
public:
explicit IREmitter(Block& block_) : block{block_}, insertion_point{block.end()} {}
explicit IREmitter(Block& block_, Block::iterator insertion_point_)
: block{block_}, insertion_point{insertion_point_} {}
Block& block;
@ -125,6 +127,7 @@ public:
[[nodiscard]] U16U32U64 FPTrunc(const U16U32U64& value);
[[nodiscard]] U32U64 IAdd(const U32U64& a, const U32U64& b);
[[nodiscard]] U32U64 ISub(const U32U64& a, const U32U64& b);
[[nodiscard]] U32 IMul(const U32& a, const U32& b);
[[nodiscard]] U32 INeg(const U32& value);
[[nodiscard]] U32 IAbs(const U32& value);
@ -155,7 +158,7 @@ public:
[[nodiscard]] U32U64 ConvertFToU(size_t bitsize, const U16U32U64& value);
[[nodiscard]] U32U64 ConvertFToI(size_t bitsize, bool is_signed, const U16U32U64& value);
[[nodiscard]] U32U64 ConvertU(size_t bitsize, const U32U64& value);
[[nodiscard]] U32U64 ConvertU(size_t result_bitsize, const U32U64& value);
private:
IR::Block::iterator insertion_point;

26
src/shader_recompiler/frontend/ir/microinstruction.cpp
View file

@ -2,6 +2,8 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <algorithm>
#include "shader_recompiler/exception.h"
#include "shader_recompiler/frontend/ir/microinstruction.h"
#include "shader_recompiler/frontend/ir/type.h"
@ -44,6 +46,13 @@ bool Inst::MayHaveSideEffects() const noexcept {
case Opcode::WriteGlobal32:
case Opcode::WriteGlobal64:
case Opcode::WriteGlobal128:
case Opcode::WriteStorageU8:
case Opcode::WriteStorageS8:
case Opcode::WriteStorageU16:
case Opcode::WriteStorageS16:
case Opcode::WriteStorage32:
case Opcode::WriteStorage64:
case Opcode::WriteStorage128:
return true;
default:
return false;
@ -56,15 +65,19 @@ bool Inst::IsPseudoInstruction() const noexcept {
case Opcode::GetSignFromOp:
case Opcode::GetCarryFromOp:
case Opcode::GetOverflowFromOp:
case Opcode::GetZSCOFromOp:
return true;
default:
return false;
}
}
bool Inst::AreAllArgsImmediates() const noexcept {
return std::all_of(args.begin(), args.begin() + NumArgs(),
[](const IR::Value& value) { return value.IsImmediate(); });
}
bool Inst::HasAssociatedPseudoOperation() const noexcept {
return zero_inst || sign_inst || carry_inst || overflow_inst || zsco_inst;
return zero_inst || sign_inst || carry_inst || overflow_inst;
}
Inst* Inst::GetAssociatedPseudoOperation(IR::Opcode opcode) {
@ -82,9 +95,6 @@ Inst* Inst::GetAssociatedPseudoOperation(IR::Opcode opcode) {
case Opcode::GetOverflowFromOp:
CheckPseudoInstruction(overflow_inst, Opcode::GetOverflowFromOp);
return overflow_inst;
case Opcode::GetZSCOFromOp:
CheckPseudoInstruction(zsco_inst, Opcode::GetZSCOFromOp);
return zsco_inst;
default:
throw InvalidArgument("{} is not a pseudo-instruction", opcode);
}
@ -176,9 +186,6 @@ void Inst::Use(const Value& value) {
case Opcode::GetOverflowFromOp:
SetPseudoInstruction(value.Inst()->overflow_inst, this);
break;
case Opcode::GetZSCOFromOp:
SetPseudoInstruction(value.Inst()->zsco_inst, this);
break;
default:
break;
}
@ -200,9 +207,6 @@ void Inst::UndoUse(const Value& value) {
case Opcode::GetOverflowFromOp:
RemovePseudoInstruction(value.Inst()->overflow_inst, Opcode::GetOverflowFromOp);
break;
case Opcode::GetZSCOFromOp:
RemovePseudoInstruction(value.Inst()->zsco_inst, Opcode::GetZSCOFromOp);
break;
default:
break;
}

4
src/shader_recompiler/frontend/ir/microinstruction.h
View file

@ -49,6 +49,9 @@ public:
/// Pseudo-instructions depend on their parent instructions for their semantics.
[[nodiscard]] bool IsPseudoInstruction() const noexcept;
/// Determines if all arguments of this instruction are immediates.
[[nodiscard]] bool AreAllArgsImmediates() const noexcept;
/// Determines if there is a pseudo-operation associated with this instruction.
[[nodiscard]] bool HasAssociatedPseudoOperation() const noexcept;
/// Gets a pseudo-operation associated with this instruction
@ -94,7 +97,6 @@ private:
Inst* sign_inst{};
Inst* carry_inst{};
Inst* overflow_inst{};
Inst* zsco_inst{};
std::vector<std::pair<Block*, Value>> phi_operands;
u64 flags{};
};

22
src/shader_recompiler/frontend/ir/opcode.inc
View file

@ -24,9 +24,6 @@ OPCODE(GetAttribute, U32, Attr
OPCODE(SetAttribute, U32, Attribute, )
OPCODE(GetAttributeIndexed, U32, U32, )
OPCODE(SetAttributeIndexed, U32, U32, )
OPCODE(GetZSCORaw, U32, )
OPCODE(SetZSCORaw, Void, U32, )
OPCODE(SetZSCO, Void, ZSCO, )
OPCODE(GetZFlag, U1, Void, )
OPCODE(GetSFlag, U1, Void, )
OPCODE(GetCFlag, U1, Void, )
@ -65,6 +62,22 @@ OPCODE(WriteGlobal32, Void, U64,
OPCODE(WriteGlobal64, Void, U64, Opaque, )
OPCODE(WriteGlobal128, Void, U64, Opaque, )
// Storage buffer operations
OPCODE(LoadStorageU8, U32, U32, U32, )
OPCODE(LoadStorageS8, U32, U32, U32, )
OPCODE(LoadStorageU16, U32, U32, U32, )
OPCODE(LoadStorageS16, U32, U32, U32, )
OPCODE(LoadStorage32, U32, U32, U32, )
OPCODE(LoadStorage64, Opaque, U32, U32, )
OPCODE(LoadStorage128, Opaque, U32, U32, )
OPCODE(WriteStorageU8, Void, U32, U32, U32, )
OPCODE(WriteStorageS8, Void, U32, U32, U32, )
OPCODE(WriteStorageU16, Void, U32, U32, U32, )
OPCODE(WriteStorageS16, Void, U32, U32, U32, )
OPCODE(WriteStorage32, Void, U32, U32, U32, )
OPCODE(WriteStorage64, Void, U32, U32, Opaque, )
OPCODE(WriteStorage128, Void, U32, U32, Opaque, )
// Vector utility
OPCODE(CompositeConstruct2, Opaque, Opaque, Opaque, )
OPCODE(CompositeConstruct3, Opaque, Opaque, Opaque, Opaque, )
@ -90,7 +103,6 @@ OPCODE(GetZeroFromOp, U1, Opaq
OPCODE(GetSignFromOp, U1, Opaque, )
OPCODE(GetCarryFromOp, U1, Opaque, )
OPCODE(GetOverflowFromOp, U1, Opaque, )
OPCODE(GetZSCOFromOp, ZSCO, Opaque, )
// Floating-point operations
OPCODE(FPAbs16, U16, U16, )
@ -143,6 +155,8 @@ OPCODE(FPTrunc64, U64, U64,
// Integer operations
OPCODE(IAdd32, U32, U32, U32, )
OPCODE(IAdd64, U64, U64, U64, )
OPCODE(ISub32, U32, U32, U32, )
OPCODE(ISub64, U64, U64, U64, )
OPCODE(IMul32, U32, U32, U32, )
OPCODE(INeg32, U32, U32, )
OPCODE(IAbs32, U32, U32, )

2
src/shader_recompiler/frontend/ir/type.cpp
View file

@ -11,7 +11,7 @@ namespace Shader::IR {
std::string NameOf(Type type) {
static constexpr std::array names{
"Opaque", "Label", "Reg", "Pred", "Attribute", "U1", "U8", "U16", "U32", "U64", "ZSCO",
"Opaque", "Label", "Reg", "Pred", "Attribute", "U1", "U8", "U16", "U32", "U64",
};
const size_t bits{static_cast<size_t>(type)};
if (bits == 0) {

1
src/shader_recompiler/frontend/ir/type.h
View file

@ -25,7 +25,6 @@ enum class Type {
U16 = 1 << 7,
U32 = 1 << 8,
U64 = 1 << 9,
ZSCO = 1 << 10,
};
DECLARE_ENUM_FLAG_OPERATORS(Type)

17
src/shader_recompiler/frontend/ir/value.cpp
View file

@ -91,26 +91,41 @@ IR::Attribute Value::Attribute() const {
}
bool Value::U1() const {
if (IsIdentity()) {
return inst->Arg(0).U1();
}
ValidateAccess(Type::U1);
return imm_u1;
}
u8 Value::U8() const {
if (IsIdentity()) {
return inst->Arg(0).U8();
}
ValidateAccess(Type::U8);
return imm_u8;
}
u16 Value::U16() const {
if (IsIdentity()) {
return inst->Arg(0).U16();
}
ValidateAccess(Type::U16);
return imm_u16;
}
u32 Value::U32() const {
if (IsIdentity()) {
return inst->Arg(0).U32();
}
ValidateAccess(Type::U32);
return imm_u32;
}
u64 Value::U64() const {
if (IsIdentity()) {
return inst->Arg(0).U64();
}
ValidateAccess(Type::U64);
return imm_u64;
}
@ -142,8 +157,6 @@ bool Value::operator==(const Value& other) const {
return imm_u32 == other.imm_u32;
case Type::U64:
return imm_u64 == other.imm_u64;
case Type::ZSCO:
throw NotImplementedException("ZSCO comparison");
}
throw LogicError("Invalid type {}", type);
}

1
src/shader_recompiler/frontend/ir/value.h
View file

@ -96,6 +96,5 @@ using U64 = TypedValue<Type::U64>;
using U32U64 = TypedValue<Type::U32 | Type::U64>;
using U16U32U64 = TypedValue<Type::U16 | Type::U32 | Type::U64>;
using UAny = TypedValue<Type::U8 | Type::U16 | Type::U32 | Type::U64>;
using ZSCO = TypedValue<Type::ZSCO>;
} // namespace Shader::IR

6
src/shader_recompiler/frontend/maxwell/program.cpp
View file

@ -52,9 +52,11 @@ Program::Program(Environment& env, const Flow::CFG& cfg) {
}
std::ranges::for_each(functions, Optimization::SsaRewritePass);
for (IR::Function& function : functions) {
Optimization::Invoke(Optimization::GlobalMemoryToStorageBufferPass, function);
Optimization::Invoke(Optimization::ConstantPropagationPass, function);
Optimization::Invoke(Optimization::DeadCodeEliminationPass, function);
Optimization::Invoke(Optimization::IdentityRemovalPass, function);
// Optimization::Invoke(Optimization::VerificationPass, function);
Optimization::IdentityRemovalPass(function);
Optimization::VerificationPass(function);
}
//*/
}

146
src/shader_recompiler/ir_opt/constant_propagation_pass.cpp Normal file
View file

@ -0,0 +1,146 @@
// Copyright 2021 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <algorithm>
#include <type_traits>
#include "common/bit_util.h"
#include "shader_recompiler/exception.h"
#include "shader_recompiler/frontend/ir/microinstruction.h"
#include "shader_recompiler/ir_opt/passes.h"
namespace Shader::Optimization {
namespace {
[[nodiscard]] u32 BitFieldUExtract(u32 base, u32 shift, u32 count) {
if (static_cast<size_t>(shift) + static_cast<size_t>(count) > Common::BitSize<u32>()) {
throw LogicError("Undefined result in BitFieldUExtract({}, {}, {})", base, shift, count);
}
return (base >> shift) & ((1U << count) - 1);
}
template <typename T>
[[nodiscard]] T Arg(const IR::Value& value) {
if constexpr (std::is_same_v<T, bool>) {
return value.U1();
} else if constexpr (std::is_same_v<T, u32>) {
return value.U32();
} else if constexpr (std::is_same_v<T, u64>) {
return value.U64();
}
}
template <typename ImmFn>
bool FoldCommutative(IR::Inst& inst, ImmFn&& imm_fn) {
const auto arg = [](const IR::Value& value) {
if constexpr (std::is_invocable_r_v<bool, ImmFn, bool, bool>) {
return value.U1();
} else if constexpr (std::is_invocable_r_v<u32, ImmFn, u32, u32>) {
return value.U32();
} else if constexpr (std::is_invocable_r_v<u64, ImmFn, u64, u64>) {
return value.U64();
}
};
const IR::Value lhs{inst.Arg(0)};
const IR::Value rhs{inst.Arg(1)};
const bool is_lhs_immediate{lhs.IsImmediate()};
const bool is_rhs_immediate{rhs.IsImmediate()};
if (is_lhs_immediate && is_rhs_immediate) {
const auto result{imm_fn(arg(lhs), arg(rhs))};
inst.ReplaceUsesWith(IR::Value{result});
return false;
}
if (is_lhs_immediate && !is_rhs_immediate) {
IR::Inst* const rhs_inst{rhs.InstRecursive()};
if (rhs_inst->Opcode() == inst.Opcode() && rhs_inst->Arg(1).IsImmediate()) {
const auto combined{imm_fn(arg(lhs), arg(rhs_inst->Arg(1)))};
inst.SetArg(0, rhs_inst->Arg(0));
inst.SetArg(1, IR::Value{combined});
} else {
// Normalize
inst.SetArg(0, rhs);
inst.SetArg(1, lhs);
}
}
if (!is_lhs_immediate && is_rhs_immediate) {
const IR::Inst* const lhs_inst{lhs.InstRecursive()};
if (lhs_inst->Opcode() == inst.Opcode() && lhs_inst->Arg(1).IsImmediate()) {
const auto combined{imm_fn(arg(rhs), arg(lhs_inst->Arg(1)))};
inst.SetArg(0, lhs_inst->Arg(0));
inst.SetArg(1, IR::Value{combined});
}
}
return true;
}
void FoldGetRegister(IR::Inst& inst) {
if (inst.Arg(0).Reg() == IR::Reg::RZ) {
inst.ReplaceUsesWith(IR::Value{u32{0}});
}
}
void FoldGetPred(IR::Inst& inst) {
if (inst.Arg(0).Pred() == IR::Pred::PT) {
inst.ReplaceUsesWith(IR::Value{true});
}
}
template <typename T>
void FoldAdd(IR::Inst& inst) {
if (inst.HasAssociatedPseudoOperation()) {
return;
}
if (!FoldCommutative(inst, [](T a, T b) { return a + b; })) {
return;
}
const IR::Value rhs{inst.Arg(1)};
if (rhs.IsImmediate() && Arg<T>(rhs) == 0) {
inst.ReplaceUsesWith(inst.Arg(0));
}
}
void FoldLogicalAnd(IR::Inst& inst) {
if (!FoldCommutative(inst, [](bool a, bool b) { return a && b; })) {
return;
}
const IR::Value rhs{inst.Arg(1)};
if (rhs.IsImmediate()) {
if (rhs.U1()) {
inst.ReplaceUsesWith(inst.Arg(0));
} else {
inst.ReplaceUsesWith(IR::Value{false});
}
}
}
void ConstantPropagation(IR::Inst& inst) {
switch (inst.Opcode()) {
case IR::Opcode::GetRegister:
return FoldGetRegister(inst);
case IR::Opcode::GetPred:
return FoldGetPred(inst);
case IR::Opcode::IAdd32:
return FoldAdd<u32>(inst);
case IR::Opcode::IAdd64:
return FoldAdd<u64>(inst);
case IR::Opcode::BitFieldUExtract:
if (inst.AreAllArgsImmediates() && !inst.HasAssociatedPseudoOperation()) {
inst.ReplaceUsesWith(IR::Value{
BitFieldUExtract(inst.Arg(0).U32(), inst.Arg(1).U32(), inst.Arg(2).U32())});
}
break;
case IR::Opcode::LogicalAnd:
return FoldLogicalAnd(inst);
default:
break;
}
}
} // Anonymous namespace
void ConstantPropagationPass(IR::Block& block) {
std::ranges::for_each(block, ConstantPropagation);
}
} // namespace Shader::Optimization

331
src/shader_recompiler/ir_opt/global_memory_to_storage_buffer_pass.cpp Normal file
View file

@ -0,0 +1,331 @@
// Copyright 2021 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <algorithm>
#include <compare>
#include <optional>
#include <ranges>
#include <boost/container/flat_set.hpp>
#include <boost/container/small_vector.hpp>
#include "shader_recompiler/frontend/ir/basic_block.h"
#include "shader_recompiler/frontend/ir/ir_emitter.h"
#include "shader_recompiler/frontend/ir/microinstruction.h"
#include "shader_recompiler/ir_opt/passes.h"
namespace Shader::Optimization {
namespace {
/// Address in constant buffers to the storage buffer descriptor
struct StorageBufferAddr {
auto operator<=>(const StorageBufferAddr&) const noexcept = default;
u32 index;
u32 offset;
};
/// Block iterator to a global memory instruction and the storage buffer it uses
struct StorageInst {
StorageBufferAddr storage_buffer;
IR::Block::iterator inst;
};
/// Bias towards a certain range of constant buffers when looking for storage buffers
struct Bias {
u32 index;
u32 offset_begin;
u32 offset_end;
};
using StorageBufferSet =
boost::container::flat_set<StorageBufferAddr, std::less<StorageBufferAddr>,
boost::container::small_vector<StorageBufferAddr, 16>>;
using StorageInstVector = boost::container::small_vector<StorageInst, 32>;
/// Returns true when the instruction is a global memory instruction
bool IsGlobalMemory(const IR::Inst& inst) {
switch (inst.Opcode()) {
case IR::Opcode::LoadGlobalS8:
case IR::Opcode::LoadGlobalU8:
case IR::Opcode::LoadGlobalS16:
case IR::Opcode::LoadGlobalU16:
case IR::Opcode::LoadGlobal32:
case IR::Opcode::LoadGlobal64:
case IR::Opcode::LoadGlobal128:
case IR::Opcode::WriteGlobalS8:
case IR::Opcode::WriteGlobalU8:
case IR::Opcode::WriteGlobalS16:
case IR::Opcode::WriteGlobalU16:
case IR::Opcode::WriteGlobal32:
case IR::Opcode::WriteGlobal64:
case IR::Opcode::WriteGlobal128:
return true;
default:
return false;
}
}
/// Converts a global memory opcode to its storage buffer equivalent
IR::Opcode GlobalToStorage(IR::Opcode opcode) {
switch (opcode) {
case IR::Opcode::LoadGlobalS8:
return IR::Opcode::LoadStorageS8;
case IR::Opcode::LoadGlobalU8:
return IR::Opcode::LoadStorageU8;
case IR::Opcode::LoadGlobalS16:
return IR::Opcode::LoadStorageS16;
case IR::Opcode::LoadGlobalU16:
return IR::Opcode::LoadStorageU16;
case IR::Opcode::LoadGlobal32:
return IR::Opcode::LoadStorage32;
case IR::Opcode::LoadGlobal64:
return IR::Opcode::LoadStorage64;
case IR::Opcode::LoadGlobal128:
return IR::Opcode::LoadStorage128;
case IR::Opcode::WriteGlobalS8:
return IR::Opcode::WriteStorageS8;
case IR::Opcode::WriteGlobalU8:
return IR::Opcode::WriteStorageU8;
case IR::Opcode::WriteGlobalS16:
return IR::Opcode::WriteStorageS16;
case IR::Opcode::WriteGlobalU16:
return IR::Opcode::WriteStorageU16;
case IR::Opcode::WriteGlobal32:
return IR::Opcode::WriteStorage32;
case IR::Opcode::WriteGlobal64:
return IR::Opcode::WriteStorage64;
case IR::Opcode::WriteGlobal128:
return IR::Opcode::WriteStorage128;
default:
throw InvalidArgument("Invalid global memory opcode {}", opcode);
}
}
/// Returns true when a storage buffer address satisfies a bias
bool MeetsBias(const StorageBufferAddr& storage_buffer, const Bias& bias) noexcept {
return storage_buffer.index == bias.index && storage_buffer.offset >= bias.offset_begin &&
storage_buffer.offset < bias.offset_end;
}
/// Ignores a global memory operation, reads return zero and writes are ignored
void IgnoreGlobalMemory(IR::Block& block, IR::Block::iterator inst) {
const IR::Value zero{u32{0}};
switch (inst->Opcode()) {
case IR::Opcode::LoadGlobalS8:
case IR::Opcode::LoadGlobalU8:
case IR::Opcode::LoadGlobalS16:
case IR::Opcode::LoadGlobalU16:
case IR::Opcode::LoadGlobal32:
inst->ReplaceUsesWith(zero);
break;
case IR::Opcode::LoadGlobal64:
inst->ReplaceUsesWith(
IR::Value{&*block.PrependNewInst(inst, IR::Opcode::CompositeConstruct2, {zero, zero})});
break;
case IR::Opcode::LoadGlobal128:
inst->ReplaceUsesWith(IR::Value{&*block.PrependNewInst(
inst, IR::Opcode::CompositeConstruct4, {zero, zero, zero, zero})});
break;
case IR::Opcode::WriteGlobalS8:
case IR::Opcode::WriteGlobalU8:
case IR::Opcode::WriteGlobalS16:
case IR::Opcode::WriteGlobalU16:
case IR::Opcode::WriteGlobal32:
case IR::Opcode::WriteGlobal64:
case IR::Opcode::WriteGlobal128:
inst->Invalidate();
break;
default:
throw LogicError("Invalid opcode to ignore its global memory operation {}", inst->Opcode());
}
}
/// Recursively tries to track the storage buffer address used by a global memory instruction
std::optional<StorageBufferAddr> Track(const IR::Value& value, const Bias* bias) {
if (value.IsImmediate()) {
// Immediates can't be a storage buffer
return std::nullopt;
}
const IR::Inst* const inst{value.InstRecursive()};
if (inst->Opcode() == IR::Opcode::GetCbuf) {
const IR::Value index{inst->Arg(0)};
const IR::Value offset{inst->Arg(1)};
if (!index.IsImmediate()) {
// Definitely not a storage buffer if it's read from a non-immediate index
return std::nullopt;
}
if (!offset.IsImmediate()) {
// TODO: Support SSBO arrays
return std::nullopt;
}
const StorageBufferAddr storage_buffer{
.index = index.U32(),
.offset = offset.U32(),
};
if (bias && !MeetsBias(storage_buffer, *bias)) {
// We have to blacklist some addresses in case we wrongly point to them
return std::nullopt;
}
return storage_buffer;
}
// Reversed loops are more likely to find the right result
for (size_t arg = inst->NumArgs(); arg--;) {
if (const std::optional storage_buffer{Track(inst->Arg(arg), bias)}) {
return *storage_buffer;
}
}
return std::nullopt;
}
/// Collects the storage buffer used by a global memory instruction and the instruction itself
void CollectStorageBuffers(IR::Block& block, IR::Block::iterator inst,
StorageBufferSet& storage_buffer_set, StorageInstVector& to_replace) {
// NVN puts storage buffers in a specific range, we have to bias towards these addresses to
// avoid getting false positives
static constexpr Bias nvn_bias{
.index{0},
.offset_begin{0x110},
.offset_end{0x610},
};
// First try to find storage buffers in the NVN address
const IR::U64 addr{inst->Arg(0)};
std::optional<StorageBufferAddr> storage_buffer{Track(addr, &nvn_bias)};
if (!storage_buffer) {
// If it fails, track without a bias
storage_buffer = Track(addr, nullptr);
if (!storage_buffer) {
// If that also failed, drop the global memory usage
IgnoreGlobalMemory(block, inst);
}
}
// Collect storage buffer and the instruction
storage_buffer_set.insert(*storage_buffer);
to_replace.push_back(StorageInst{
.storage_buffer{*storage_buffer},
.inst{inst},
});
}
/// Tries to track the first 32-bits of a global memory instruction
std::optional<IR::U32> TrackLowAddress(IR::IREmitter& ir, IR::Inst* inst) {
// The first argument is the low level GPU pointer to the global memory instruction
const IR::U64 addr{inst->Arg(0)};
if (addr.IsImmediate()) {
// Not much we can do if it's an immediate
return std::nullopt;
}
// This address is expected to either be a PackUint2x32 or a IAdd64
IR::Inst* addr_inst{addr.InstRecursive()};
s32 imm_offset{0};
if (addr_inst->Opcode() == IR::Opcode::IAdd64) {
// If it's an IAdd64, get the immediate offset it is applying and grab the address
// instruction. This expects for the instruction to be canonicalized having the address on
// the first argument and the immediate offset on the second one.
const IR::U64 imm_offset_value{addr_inst->Arg(1)};
if (!imm_offset_value.IsImmediate()) {
return std::nullopt;
}
imm_offset = static_cast<s32>(static_cast<s64>(imm_offset_value.U64()));
const IR::U64 iadd_addr{addr_inst->Arg(0)};
if (iadd_addr.IsImmediate()) {
return std::nullopt;
}
addr_inst = iadd_addr.Inst();
}
// With IAdd64 handled, now PackUint2x32 is expected without exceptions
if (addr_inst->Opcode() != IR::Opcode::PackUint2x32) {
return std::nullopt;
}
// PackUint2x32 is expected to be generated from a vector
const IR::Value vector{addr_inst->Arg(0)};
if (vector.IsImmediate()) {
return std::nullopt;
}
// This vector is expected to be a CompositeConstruct2
IR::Inst* const vector_inst{vector.InstRecursive()};
if (vector_inst->Opcode() != IR::Opcode::CompositeConstruct2) {
return std::nullopt;
}
// Grab the first argument from the CompositeConstruct2, this is the low address.
// Re-apply the offset in case we found one.
const IR::U32 low_addr{vector_inst->Arg(0)};
return imm_offset != 0 ? IR::U32{ir.IAdd(low_addr, ir.Imm32(imm_offset))} : low_addr;
}
/// Returns the offset in indices (not bytes) for an equivalent storage instruction
IR::U32 StorageOffset(IR::Block& block, IR::Block::iterator inst, StorageBufferAddr buffer) {
IR::IREmitter ir{block, inst};
IR::U32 offset;
if (const std::optional<IR::U32> low_addr{TrackLowAddress(ir, &*inst)}) {
offset = *low_addr;
} else {
offset = ir.ConvertU(32, IR::U64{inst->Arg(0)});
}
// Subtract the least significant 32 bits from the guest offset. The result is the storage
// buffer offset in bytes.
const IR::U32 low_cbuf{ir.GetCbuf(ir.Imm32(buffer.index), ir.Imm32(buffer.offset))};
return ir.ISub(offset, low_cbuf);
}
/// Replace a global memory load instruction with its storage buffer equivalent
void ReplaceLoad(IR::Block& block, IR::Block::iterator inst, const IR::U32& storage_index,
const IR::U32& offset) {
const IR::Opcode new_opcode{GlobalToStorage(inst->Opcode())};
const IR::Value value{&*block.PrependNewInst(inst, new_opcode, {storage_index, offset})};
inst->ReplaceUsesWith(value);
}
/// Replace a global memory write instruction with its storage buffer equivalent
void ReplaceWrite(IR::Block& block, IR::Block::iterator inst, const IR::U32& storage_index,
const IR::U32& offset) {
const IR::Opcode new_opcode{GlobalToStorage(inst->Opcode())};
block.PrependNewInst(inst, new_opcode, {storage_index, offset, inst->Arg(1)});
inst->Invalidate();
}
/// Replace a global memory instruction with its storage buffer equivalent
void Replace(IR::Block& block, IR::Block::iterator inst, const IR::U32& storage_index,
const IR::U32& offset) {
switch (inst->Opcode()) {
case IR::Opcode::LoadGlobalS8:
case IR::Opcode::LoadGlobalU8:
case IR::Opcode::LoadGlobalS16:
case IR::Opcode::LoadGlobalU16:
case IR::Opcode::LoadGlobal32:
case IR::Opcode::LoadGlobal64:
case IR::Opcode::LoadGlobal128:
return ReplaceLoad(block, inst, storage_index, offset);
case IR::Opcode::WriteGlobalS8:
case IR::Opcode::WriteGlobalU8:
case IR::Opcode::WriteGlobalS16:
case IR::Opcode::WriteGlobalU16:
case IR::Opcode::WriteGlobal32:
case IR::Opcode::WriteGlobal64:
case IR::Opcode::WriteGlobal128:
return ReplaceWrite(block, inst, storage_index, offset);
default:
throw InvalidArgument("Invalid global memory opcode {}", inst->Opcode());
}
}
} // Anonymous namespace
void GlobalMemoryToStorageBufferPass(IR::Block& block) {
StorageBufferSet storage_buffers;
StorageInstVector to_replace;
for (IR::Block::iterator inst{block.begin()}; inst != block.end(); ++inst) {
if (!IsGlobalMemory(*inst)) {
continue;
}
CollectStorageBuffers(block, inst, storage_buffers, to_replace);
}
for (const auto [storage_buffer, inst] : to_replace) {
const auto it{storage_buffers.find(storage_buffer)};
const IR::U32 storage_index{IR::Value{static_cast<u32>(storage_buffers.index_of(it))}};
const IR::U32 offset{StorageOffset(block, inst, storage_buffer)};
Replace(block, inst, storage_index, offset);
}
}
} // namespace Shader::Optimization

28
src/shader_recompiler/ir_opt/identity_removal_pass.cpp
View file

@ -10,22 +10,24 @@
namespace Shader::Optimization {
void IdentityRemovalPass(IR::Block& block) {
void IdentityRemovalPass(IR::Function& function) {
std::vector<IR::Inst*> to_invalidate;
for (auto inst = block.begin(); inst != block.end();) {
const size_t num_args{inst->NumArgs()};
for (size_t i = 0; i < num_args; ++i) {
IR::Value arg;
while ((arg = inst->Arg(i)).IsIdentity()) {
inst->SetArg(i, arg.Inst()->Arg(0));
for (auto& block : function.blocks) {
for (auto inst = block->begin(); inst != block->end();) {
const size_t num_args{inst->NumArgs()};
for (size_t i = 0; i < num_args; ++i) {
IR::Value arg;
while ((arg = inst->Arg(i)).IsIdentity()) {
inst->SetArg(i, arg.Inst()->Arg(0));
}
}
if (inst->Opcode() == IR::Opcode::Identity || inst->Opcode() == IR::Opcode::Void) {
to_invalidate.push_back(&*inst);
inst = block->Instructions().erase(inst);
} else {
++inst;
}
}
if (inst->Opcode() == IR::Opcode::Identity || inst->Opcode() == IR::Opcode::Void) {
to_invalidate.push_back(&*inst);
inst = block.Instructions().erase(inst);
} else {
++inst;
}
}
for (IR::Inst* const inst : to_invalidate) {

6
src/shader_recompiler/ir_opt/passes.h
View file

@ -16,9 +16,11 @@ void Invoke(Func&& func, IR::Function& function) {
}
}
void ConstantPropagationPass(IR::Block& block);
void DeadCodeEliminationPass(IR::Block& block);
void IdentityRemovalPass(IR::Block& block);
void GlobalMemoryToStorageBufferPass(IR::Block& block);
void IdentityRemovalPass(IR::Function& function);
void SsaRewritePass(IR::Function& function);
void VerificationPass(const IR::Block& block);
void VerificationPass(const IR::Function& function);
} // namespace Shader::Optimization

56
src/shader_recompiler/ir_opt/ssa_rewrite_pass.cpp
View file

@ -14,8 +14,6 @@
// https://link.springer.com/chapter/10.1007/978-3-642-37051-9_6
//
#include <map>
#include <boost/container/flat_map.hpp>
#include "shader_recompiler/frontend/ir/basic_block.h"
@ -30,6 +28,12 @@ namespace Shader::Optimization {
namespace {
using ValueMap = boost::container::flat_map<IR::Block*, IR::Value, std::less<IR::Block*>>;
struct FlagTag {};
struct ZeroFlagTag : FlagTag {};
struct SignFlagTag : FlagTag {};
struct CarryFlagTag : FlagTag {};
struct OverflowFlagTag : FlagTag {};
struct DefTable {
[[nodiscard]] ValueMap& operator[](IR::Reg variable) noexcept {
return regs[IR::RegIndex(variable)];
@ -39,8 +43,28 @@ struct DefTable {
return preds[IR::PredIndex(variable)];
}
[[nodiscard]] ValueMap& operator[](ZeroFlagTag) noexcept {
return zero_flag;
}
[[nodiscard]] ValueMap& operator[](SignFlagTag) noexcept {
return sign_flag;
}
[[nodiscard]] ValueMap& operator[](CarryFlagTag) noexcept {
return carry_flag;
}
[[nodiscard]] ValueMap& operator[](OverflowFlagTag) noexcept {
return overflow_flag;
}
std::array<ValueMap, IR::NUM_USER_REGS> regs;
std::array<ValueMap, IR::NUM_USER_PREDS> preds;
ValueMap zero_flag;
ValueMap sign_flag;
ValueMap carry_flag;
ValueMap overflow_flag;
};
IR::Opcode UndefOpcode(IR::Reg) noexcept {
@ -51,6 +75,10 @@ IR::Opcode UndefOpcode(IR::Pred) noexcept {
return IR::Opcode::Undef1;
}
IR::Opcode UndefOpcode(const FlagTag&) noexcept {
return IR::Opcode::Undef1;
}
[[nodiscard]] bool IsPhi(const IR::Inst& inst) noexcept {
return inst.Opcode() == IR::Opcode::Phi;
}
@ -135,6 +163,18 @@ void SsaRewritePass(IR::Function& function) {
pass.WriteVariable(pred, block.get(), inst.Arg(1));
}
break;
case IR::Opcode::SetZFlag:
pass.WriteVariable(ZeroFlagTag{}, block.get(), inst.Arg(0));
break;
case IR::Opcode::SetSFlag:
pass.WriteVariable(SignFlagTag{}, block.get(), inst.Arg(0));
break;
case IR::Opcode::SetCFlag:
pass.WriteVariable(CarryFlagTag{}, block.get(), inst.Arg(0));
break;
case IR::Opcode::SetOFlag:
pass.WriteVariable(OverflowFlagTag{}, block.get(), inst.Arg(0));
break;
case IR::Opcode::GetRegister:
if (const IR::Reg reg{inst.Arg(0).Reg()}; reg != IR::Reg::RZ) {
inst.ReplaceUsesWith(pass.ReadVariable(reg, block.get()));
@ -145,6 +185,18 @@ void SsaRewritePass(IR::Function& function) {
inst.ReplaceUsesWith(pass.ReadVariable(pred, block.get()));
}
break;
case IR::Opcode::GetZFlag:
inst.ReplaceUsesWith(pass.ReadVariable(ZeroFlagTag{}, block.get()));
break;
case IR::Opcode::GetSFlag:
inst.ReplaceUsesWith(pass.ReadVariable(SignFlagTag{}, block.get()));
break;
case IR::Opcode::GetCFlag:
inst.ReplaceUsesWith(pass.ReadVariable(CarryFlagTag{}, block.get()));
break;
case IR::Opcode::GetOFlag:
inst.ReplaceUsesWith(pass.ReadVariable(OverflowFlagTag{}, block.get()));
break;
default:
break;
}

42
src/shader_recompiler/ir_opt/verification_pass.cpp
View file

@ -11,40 +11,44 @@
namespace Shader::Optimization {
static void ValidateTypes(const IR::Block& block) {
for (const IR::Inst& inst : block) {
const size_t num_args{inst.NumArgs()};
for (size_t i = 0; i < num_args; ++i) {
const IR::Type t1{inst.Arg(i).Type()};
const IR::Type t2{IR::ArgTypeOf(inst.Opcode(), i)};
if (!IR::AreTypesCompatible(t1, t2)) {
throw LogicError("Invalid types in block:\n{}", IR::DumpBlock(block));
static void ValidateTypes(const IR::Function& function) {
for (const auto& block : function.blocks) {
for (const IR::Inst& inst : *block) {
const size_t num_args{inst.NumArgs()};
for (size_t i = 0; i < num_args; ++i) {
const IR::Type t1{inst.Arg(i).Type()};
const IR::Type t2{IR::ArgTypeOf(inst.Opcode(), i)};
if (!IR::AreTypesCompatible(t1, t2)) {
throw LogicError("Invalid types in block:\n{}", IR::DumpBlock(*block));
}
}
}
}
}
static void ValidateUses(const IR::Block& block) {
static void ValidateUses(const IR::Function& function) {
std::map<IR::Inst*, int> actual_uses;
for (const IR::Inst& inst : block) {
const size_t num_args{inst.NumArgs()};
for (size_t i = 0; i < num_args; ++i) {
const IR::Value arg{inst.Arg(i)};
if (!arg.IsImmediate()) {
++actual_uses[arg.Inst()];
for (const auto& block : function.blocks) {
for (const IR::Inst& inst : *block) {
const size_t num_args{inst.NumArgs()};
for (size_t i = 0; i < num_args; ++i) {
const IR::Value arg{inst.Arg(i)};
if (!arg.IsImmediate()) {
++actual_uses[arg.Inst()];
}
}
}
}
for (const auto [inst, uses] : actual_uses) {
if (inst->UseCount() != uses) {
throw LogicError("Invalid uses in block:\n{}", IR::DumpBlock(block));
throw LogicError("Invalid uses in block:" /*, IR::DumpFunction(function)*/);
}
}
}
void VerificationPass(const IR::Block& block) {
ValidateTypes(block);
ValidateUses(block);
void VerificationPass(const IR::Function& function) {
ValidateTypes(function);
ValidateUses(function);
}
} // namespace Shader::Optimization