summaryrefslogtreecommitdiffstats
path: root/src/shader_recompiler/frontend/maxwell/translate
diff options
context:
space:
mode:
Diffstat (limited to 'src/shader_recompiler/frontend/maxwell/translate')
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/attribute_memory_to_physical.cpp1
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/barrier_operations.cpp2
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/common_funcs.h1
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/exit_program.cpp5
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/floating_point_multi_function.cpp1
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_helper.h2
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_set.cpp1
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_set_predicate.cpp1
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/load_constant.cpp14
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/load_store_attribute.cpp1
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/load_store_memory.cpp1
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/logic_operation_three_input.cpp580
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/logic_operation_three_input_lut3.py92
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/move_register.cpp1
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/surface_atomic_operations.cpp3
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/texture_fetch_swizzled.cpp2
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/texture_gather.cpp2
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/texture_gather_swizzled.cpp2
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/texture_gradient.cpp2
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/texture_load.cpp4
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/texture_mipmap_level.cpp4
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/texture_query.cpp1
-rw-r--r--src/shader_recompiler/frontend/maxwell/translate/impl/warp_shuffle.cpp2
23 files changed, 640 insertions, 85 deletions
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/attribute_memory_to_physical.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/attribute_memory_to_physical.cpp
index fb3f00d3f..d26d0982b 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/attribute_memory_to_physical.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/attribute_memory_to_physical.cpp
@@ -4,7 +4,6 @@
#include "common/bit_field.h"
#include "common/common_types.h"
-#include "shader_recompiler/frontend/maxwell/opcodes.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/impl.h"
namespace Shader::Maxwell {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/barrier_operations.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/barrier_operations.cpp
index 86e433e41..d92d4e929 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/barrier_operations.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/barrier_operations.cpp
@@ -4,8 +4,6 @@
#include "common/bit_field.h"
#include "common/common_types.h"
-#include "shader_recompiler/frontend/ir/modifiers.h"
-#include "shader_recompiler/frontend/maxwell/opcodes.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/impl.h"
namespace Shader::Maxwell {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/common_funcs.h b/src/shader_recompiler/frontend/maxwell/translate/impl/common_funcs.h
index 214d0af3c..24f041a4f 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/common_funcs.h
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/common_funcs.h
@@ -4,7 +4,6 @@
#pragma once
-#include "common/common_types.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/impl.h"
namespace Shader::Maxwell {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/exit_program.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/exit_program.cpp
index c2443c886..62d20ebe4 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/exit_program.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/exit_program.cpp
@@ -3,7 +3,6 @@
// Refer to the license.txt file included.
#include "common/common_types.h"
-#include "shader_recompiler/exception.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/impl.h"
namespace Shader::Maxwell {
@@ -12,9 +11,13 @@ void ExitFragment(TranslatorVisitor& v) {
const ProgramHeader sph{v.env.SPH()};
IR::Reg src_reg{IR::Reg::R0};
for (u32 render_target = 0; render_target < 8; ++render_target) {
+ if (!sph.ps.HasOutputComponents(render_target)) {
+ continue;
+ }
const std::array<bool, 4> mask{sph.ps.EnabledOutputComponents(render_target)};
for (u32 component = 0; component < 4; ++component) {
if (!mask[component]) {
+ ++src_reg;
continue;
}
v.ir.SetFragColor(render_target, component, v.F(src_reg));
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/floating_point_multi_function.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/floating_point_multi_function.cpp
index 2f8605619..bb8512400 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/floating_point_multi_function.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/floating_point_multi_function.cpp
@@ -5,7 +5,6 @@
#include "common/bit_field.h"
#include "common/common_types.h"
#include "shader_recompiler/exception.h"
-#include "shader_recompiler/frontend/maxwell/opcodes.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/impl.h"
namespace Shader::Maxwell {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_helper.h b/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_helper.h
index 59da56a7e..c5e80a559 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_helper.h
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_helper.h
@@ -6,8 +6,6 @@
#include "common/common_types.h"
#include "shader_recompiler/exception.h"
-#include "shader_recompiler/frontend/maxwell/translate/impl/common_encoding.h"
-#include "shader_recompiler/frontend/maxwell/translate/impl/common_funcs.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/impl.h"
namespace Shader::Maxwell {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_set.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_set.cpp
index cca5b831f..01bc9c49f 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_set.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_set.cpp
@@ -2,6 +2,7 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
+#include "shader_recompiler/frontend/maxwell/translate/impl/common_funcs.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_helper.h"
namespace Shader::Maxwell {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_set_predicate.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_set_predicate.cpp
index b3931dae3..2b9c4fc0f 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_set_predicate.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_set_predicate.cpp
@@ -2,6 +2,7 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
+#include "shader_recompiler/frontend/maxwell/translate/impl/common_funcs.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/half_floating_point_helper.h"
namespace Shader::Maxwell {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/load_constant.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/load_constant.cpp
index 2300088e3..8007a4d46 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/load_constant.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/load_constant.cpp
@@ -11,10 +11,20 @@ namespace Shader::Maxwell {
using namespace LDC;
namespace {
std::pair<IR::U32, IR::U32> Slot(IR::IREmitter& ir, Mode mode, const IR::U32& imm_index,
- const IR::U32& reg, const IR::U32& imm) {
+ const IR::U32& reg, const IR::U32& imm_offset) {
switch (mode) {
case Mode::Default:
- return {imm_index, ir.IAdd(reg, imm)};
+ return {imm_index, ir.IAdd(reg, imm_offset)};
+ case Mode::IS: {
+ // Segmented addressing mode
+ // Ra+imm_offset points into a flat mapping of const buffer
+ // address space
+ const IR::U32 address{ir.IAdd(reg, imm_offset)};
+ const IR::U32 index{ir.BitFieldExtract(address, ir.Imm32(16), ir.Imm32(16))};
+ const IR::U32 offset{ir.BitFieldExtract(address, ir.Imm32(0), ir.Imm32(16))};
+
+ return {ir.IAdd(index, imm_index), offset};
+ }
default:
break;
}
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/load_store_attribute.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/load_store_attribute.cpp
index 924fb7a40..00d1d8438 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/load_store_attribute.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/load_store_attribute.cpp
@@ -6,7 +6,6 @@
#include "common/common_types.h"
#include "shader_recompiler/exception.h"
#include "shader_recompiler/frontend/ir/ir_emitter.h"
-#include "shader_recompiler/frontend/maxwell/opcodes.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/impl.h"
namespace Shader::Maxwell {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/load_store_memory.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/load_store_memory.cpp
index 36c5cff2f..4792cd4a5 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/load_store_memory.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/load_store_memory.cpp
@@ -5,7 +5,6 @@
#include "common/bit_field.h"
#include "common/common_types.h"
#include "shader_recompiler/exception.h"
-#include "shader_recompiler/frontend/maxwell/opcodes.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/impl.h"
namespace Shader::Maxwell {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/logic_operation_three_input.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/logic_operation_three_input.cpp
index e0fe47912..f3c7ceb57 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/logic_operation_three_input.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/logic_operation_three_input.cpp
@@ -13,59 +13,535 @@ namespace {
// Emulate GPU's LOP3.LUT (three-input logic op with 8-bit truth table)
IR::U32 ApplyLUT(IR::IREmitter& ir, const IR::U32& a, const IR::U32& b, const IR::U32& c,
u64 ttbl) {
- IR::U32 r{ir.Imm32(0)};
- const IR::U32 not_a{ir.BitwiseNot(a)};
- const IR::U32 not_b{ir.BitwiseNot(b)};
- const IR::U32 not_c{ir.BitwiseNot(c)};
- if (ttbl & 0x01) {
- // r |= ~a & ~b & ~c;
- const auto lhs{ir.BitwiseAnd(not_a, not_b)};
- const auto rhs{ir.BitwiseAnd(lhs, not_c)};
- r = ir.BitwiseOr(r, rhs);
+ switch (ttbl) {
+ // generated code, do not edit manually
+ case 0:
+ return ir.Imm32(0);
+ case 1:
+ return ir.BitwiseNot(ir.BitwiseOr(a, ir.BitwiseOr(b, c)));
+ case 2:
+ return ir.BitwiseAnd(c, ir.BitwiseNot(ir.BitwiseOr(a, b)));
+ case 3:
+ return ir.BitwiseNot(ir.BitwiseOr(a, b));
+ case 4:
+ return ir.BitwiseAnd(b, ir.BitwiseNot(ir.BitwiseOr(a, c)));
+ case 5:
+ return ir.BitwiseNot(ir.BitwiseOr(a, c));
+ case 6:
+ return ir.BitwiseAnd(ir.BitwiseNot(a), ir.BitwiseXor(b, c));
+ case 7:
+ return ir.BitwiseNot(ir.BitwiseOr(a, ir.BitwiseAnd(b, c)));
+ case 8:
+ return ir.BitwiseAnd(ir.BitwiseAnd(b, c), ir.BitwiseNot(a));
+ case 9:
+ return ir.BitwiseNot(ir.BitwiseOr(a, ir.BitwiseXor(b, c)));
+ case 10:
+ return ir.BitwiseAnd(c, ir.BitwiseNot(a));
+ case 11:
+ return ir.BitwiseAnd(ir.BitwiseNot(a), ir.BitwiseOr(c, ir.BitwiseNot(b)));
+ case 12:
+ return ir.BitwiseAnd(b, ir.BitwiseNot(a));
+ case 13:
+ return ir.BitwiseAnd(ir.BitwiseNot(a), ir.BitwiseOr(b, ir.BitwiseNot(c)));
+ case 14:
+ return ir.BitwiseAnd(ir.BitwiseNot(a), ir.BitwiseOr(b, c));
+ case 15:
+ return ir.BitwiseNot(a);
+ case 16:
+ return ir.BitwiseAnd(a, ir.BitwiseNot(ir.BitwiseOr(b, c)));
+ case 17:
+ return ir.BitwiseNot(ir.BitwiseOr(b, c));
+ case 18:
+ return ir.BitwiseAnd(ir.BitwiseNot(b), ir.BitwiseXor(a, c));
+ case 19:
+ return ir.BitwiseNot(ir.BitwiseOr(b, ir.BitwiseAnd(a, c)));
+ case 20:
+ return ir.BitwiseAnd(ir.BitwiseNot(c), ir.BitwiseXor(a, b));
+ case 21:
+ return ir.BitwiseNot(ir.BitwiseOr(c, ir.BitwiseAnd(a, b)));
+ case 22:
+ return ir.BitwiseXor(ir.BitwiseOr(a, b), ir.BitwiseOr(c, ir.BitwiseAnd(a, b)));
+ case 23:
+ return ir.BitwiseXor(ir.BitwiseAnd(ir.BitwiseXor(a, b), ir.BitwiseXor(a, c)),
+ ir.BitwiseNot(a));
+ case 24:
+ return ir.BitwiseAnd(ir.BitwiseXor(a, b), ir.BitwiseXor(a, c));
+ case 25:
+ return ir.BitwiseNot(ir.BitwiseOr(ir.BitwiseAnd(a, b), ir.BitwiseXor(b, c)));
+ case 26:
+ return ir.BitwiseAnd(ir.BitwiseOr(c, ir.BitwiseNot(b)), ir.BitwiseXor(a, c));
+ case 27:
+ return ir.BitwiseXor(ir.BitwiseOr(a, ir.BitwiseNot(c)), ir.BitwiseOr(b, c));
+ case 28:
+ return ir.BitwiseAnd(ir.BitwiseOr(b, ir.BitwiseNot(c)), ir.BitwiseXor(a, b));
+ case 29:
+ return ir.BitwiseXor(ir.BitwiseOr(a, ir.BitwiseNot(b)), ir.BitwiseOr(b, c));
+ case 30:
+ return ir.BitwiseXor(a, ir.BitwiseOr(b, c));
+ case 31:
+ return ir.BitwiseNot(ir.BitwiseAnd(a, ir.BitwiseOr(b, c)));
+ case 32:
+ return ir.BitwiseAnd(ir.BitwiseAnd(a, c), ir.BitwiseNot(b));
+ case 33:
+ return ir.BitwiseNot(ir.BitwiseOr(b, ir.BitwiseXor(a, c)));
+ case 34:
+ return ir.BitwiseAnd(c, ir.BitwiseNot(b));
+ case 35:
+ return ir.BitwiseAnd(ir.BitwiseNot(b), ir.BitwiseOr(c, ir.BitwiseNot(a)));
+ case 36:
+ return ir.BitwiseAnd(ir.BitwiseXor(a, b), ir.BitwiseXor(b, c));
+ case 37:
+ return ir.BitwiseNot(ir.BitwiseOr(ir.BitwiseAnd(a, b), ir.BitwiseXor(a, c)));
+ case 38:
+ return ir.BitwiseAnd(ir.BitwiseOr(c, ir.BitwiseNot(a)), ir.BitwiseXor(b, c));
+ case 39:
+ return ir.BitwiseXor(ir.BitwiseOr(a, c), ir.BitwiseOr(b, ir.BitwiseNot(c)));
+ case 40:
+ return ir.BitwiseAnd(c, ir.BitwiseXor(a, b));
+ case 41:
+ return ir.BitwiseXor(ir.BitwiseOr(a, b),
+ ir.BitwiseOr(ir.BitwiseAnd(a, b), ir.BitwiseNot(c)));
+ case 42:
+ return ir.BitwiseAnd(c, ir.BitwiseNot(ir.BitwiseAnd(a, b)));
+ case 43:
+ return ir.BitwiseXor(ir.BitwiseOr(a, ir.BitwiseNot(c)),
+ ir.BitwiseOr(b, ir.BitwiseXor(a, c)));
+ case 44:
+ return ir.BitwiseAnd(ir.BitwiseOr(b, c), ir.BitwiseXor(a, b));
+ case 45:
+ return ir.BitwiseXor(a, ir.BitwiseOr(b, ir.BitwiseNot(c)));
+ case 46:
+ return ir.BitwiseXor(ir.BitwiseAnd(a, b), ir.BitwiseOr(b, c));
+ case 47:
+ return ir.BitwiseOr(ir.BitwiseAnd(c, ir.BitwiseNot(b)), ir.BitwiseNot(a));
+ case 48:
+ return ir.BitwiseAnd(a, ir.BitwiseNot(b));
+ case 49:
+ return ir.BitwiseAnd(ir.BitwiseNot(b), ir.BitwiseOr(a, ir.BitwiseNot(c)));
+ case 50:
+ return ir.BitwiseAnd(ir.BitwiseNot(b), ir.BitwiseOr(a, c));
+ case 51:
+ return ir.BitwiseNot(b);
+ case 52:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, ir.BitwiseNot(c)), ir.BitwiseXor(a, b));
+ case 53:
+ return ir.BitwiseXor(ir.BitwiseOr(a, c), ir.BitwiseOr(b, ir.BitwiseNot(a)));
+ case 54:
+ return ir.BitwiseXor(b, ir.BitwiseOr(a, c));
+ case 55:
+ return ir.BitwiseNot(ir.BitwiseAnd(b, ir.BitwiseOr(a, c)));
+ case 56:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, c), ir.BitwiseXor(a, b));
+ case 57:
+ return ir.BitwiseXor(b, ir.BitwiseOr(a, ir.BitwiseNot(c)));
+ case 58:
+ return ir.BitwiseXor(ir.BitwiseAnd(a, b), ir.BitwiseOr(a, c));
+ case 59:
+ return ir.BitwiseOr(ir.BitwiseAnd(c, ir.BitwiseNot(a)), ir.BitwiseNot(b));
+ case 60:
+ return ir.BitwiseXor(a, b);
+ case 61:
+ return ir.BitwiseOr(ir.BitwiseNot(ir.BitwiseOr(a, c)), ir.BitwiseXor(a, b));
+ case 62:
+ return ir.BitwiseOr(ir.BitwiseAnd(c, ir.BitwiseNot(a)), ir.BitwiseXor(a, b));
+ case 63:
+ return ir.BitwiseNot(ir.BitwiseAnd(a, b));
+ case 64:
+ return ir.BitwiseAnd(ir.BitwiseAnd(a, b), ir.BitwiseNot(c));
+ case 65:
+ return ir.BitwiseNot(ir.BitwiseOr(c, ir.BitwiseXor(a, b)));
+ case 66:
+ return ir.BitwiseAnd(ir.BitwiseXor(a, c), ir.BitwiseXor(b, c));
+ case 67:
+ return ir.BitwiseNot(ir.BitwiseOr(ir.BitwiseAnd(a, c), ir.BitwiseXor(a, b)));
+ case 68:
+ return ir.BitwiseAnd(b, ir.BitwiseNot(c));
+ case 69:
+ return ir.BitwiseAnd(ir.BitwiseNot(c), ir.BitwiseOr(b, ir.BitwiseNot(a)));
+ case 70:
+ return ir.BitwiseAnd(ir.BitwiseOr(b, ir.BitwiseNot(a)), ir.BitwiseXor(b, c));
+ case 71:
+ return ir.BitwiseXor(ir.BitwiseOr(a, b), ir.BitwiseOr(c, ir.BitwiseNot(b)));
+ case 72:
+ return ir.BitwiseAnd(b, ir.BitwiseXor(a, c));
+ case 73:
+ return ir.BitwiseXor(ir.BitwiseOr(a, c),
+ ir.BitwiseOr(ir.BitwiseAnd(a, c), ir.BitwiseNot(b)));
+ case 74:
+ return ir.BitwiseAnd(ir.BitwiseOr(b, c), ir.BitwiseXor(a, c));
+ case 75:
+ return ir.BitwiseXor(a, ir.BitwiseOr(c, ir.BitwiseNot(b)));
+ case 76:
+ return ir.BitwiseAnd(b, ir.BitwiseNot(ir.BitwiseAnd(a, c)));
+ case 77:
+ return ir.BitwiseXor(ir.BitwiseOr(a, ir.BitwiseNot(b)),
+ ir.BitwiseOr(c, ir.BitwiseXor(a, b)));
+ case 78:
+ return ir.BitwiseXor(ir.BitwiseAnd(a, c), ir.BitwiseOr(b, c));
+ case 79:
+ return ir.BitwiseOr(ir.BitwiseAnd(b, ir.BitwiseNot(c)), ir.BitwiseNot(a));
+ case 80:
+ return ir.BitwiseAnd(a, ir.BitwiseNot(c));
+ case 81:
+ return ir.BitwiseAnd(ir.BitwiseNot(c), ir.BitwiseOr(a, ir.BitwiseNot(b)));
+ case 82:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, ir.BitwiseNot(b)), ir.BitwiseXor(a, c));
+ case 83:
+ return ir.BitwiseXor(ir.BitwiseOr(a, b), ir.BitwiseOr(c, ir.BitwiseNot(a)));
+ case 84:
+ return ir.BitwiseAnd(ir.BitwiseNot(c), ir.BitwiseOr(a, b));
+ case 85:
+ return ir.BitwiseNot(c);
+ case 86:
+ return ir.BitwiseXor(c, ir.BitwiseOr(a, b));
+ case 87:
+ return ir.BitwiseNot(ir.BitwiseAnd(c, ir.BitwiseOr(a, b)));
+ case 88:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, b), ir.BitwiseXor(a, c));
+ case 89:
+ return ir.BitwiseXor(c, ir.BitwiseOr(a, ir.BitwiseNot(b)));
+ case 90:
+ return ir.BitwiseXor(a, c);
+ case 91:
+ return ir.BitwiseOr(ir.BitwiseNot(ir.BitwiseOr(a, b)), ir.BitwiseXor(a, c));
+ case 92:
+ return ir.BitwiseXor(ir.BitwiseAnd(a, c), ir.BitwiseOr(a, b));
+ case 93:
+ return ir.BitwiseOr(ir.BitwiseAnd(b, ir.BitwiseNot(a)), ir.BitwiseNot(c));
+ case 94:
+ return ir.BitwiseOr(ir.BitwiseAnd(b, ir.BitwiseNot(a)), ir.BitwiseXor(a, c));
+ case 95:
+ return ir.BitwiseNot(ir.BitwiseAnd(a, c));
+ case 96:
+ return ir.BitwiseAnd(a, ir.BitwiseXor(b, c));
+ case 97:
+ return ir.BitwiseXor(ir.BitwiseOr(b, c),
+ ir.BitwiseOr(ir.BitwiseAnd(b, c), ir.BitwiseNot(a)));
+ case 98:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, c), ir.BitwiseXor(b, c));
+ case 99:
+ return ir.BitwiseXor(b, ir.BitwiseOr(c, ir.BitwiseNot(a)));
+ case 100:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, b), ir.BitwiseXor(b, c));
+ case 101:
+ return ir.BitwiseXor(c, ir.BitwiseOr(b, ir.BitwiseNot(a)));
+ case 102:
+ return ir.BitwiseXor(b, c);
+ case 103:
+ return ir.BitwiseOr(ir.BitwiseNot(ir.BitwiseOr(a, b)), ir.BitwiseXor(b, c));
+ case 104:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, b), ir.BitwiseXor(c, ir.BitwiseAnd(a, b)));
+ case 105:
+ return ir.BitwiseXor(ir.BitwiseNot(a), ir.BitwiseXor(b, c));
+ case 106:
+ return ir.BitwiseXor(c, ir.BitwiseAnd(a, b));
+ case 107:
+ return ir.BitwiseXor(ir.BitwiseAnd(c, ir.BitwiseOr(a, b)),
+ ir.BitwiseXor(a, ir.BitwiseNot(b)));
+ case 108:
+ return ir.BitwiseXor(b, ir.BitwiseAnd(a, c));
+ case 109:
+ return ir.BitwiseXor(ir.BitwiseAnd(b, ir.BitwiseOr(a, c)),
+ ir.BitwiseXor(a, ir.BitwiseNot(c)));
+ case 110:
+ return ir.BitwiseOr(ir.BitwiseAnd(b, ir.BitwiseNot(a)), ir.BitwiseXor(b, c));
+ case 111:
+ return ir.BitwiseOr(ir.BitwiseNot(a), ir.BitwiseXor(b, c));
+ case 112:
+ return ir.BitwiseAnd(a, ir.BitwiseNot(ir.BitwiseAnd(b, c)));
+ case 113:
+ return ir.BitwiseXor(ir.BitwiseOr(b, ir.BitwiseNot(a)),
+ ir.BitwiseOr(c, ir.BitwiseXor(a, b)));
+ case 114:
+ return ir.BitwiseXor(ir.BitwiseAnd(b, c), ir.BitwiseOr(a, c));
+ case 115:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, ir.BitwiseNot(c)), ir.BitwiseNot(b));
+ case 116:
+ return ir.BitwiseXor(ir.BitwiseAnd(b, c), ir.BitwiseOr(a, b));
+ case 117:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, ir.BitwiseNot(b)), ir.BitwiseNot(c));
+ case 118:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, ir.BitwiseNot(b)), ir.BitwiseXor(b, c));
+ case 119:
+ return ir.BitwiseNot(ir.BitwiseAnd(b, c));
+ case 120:
+ return ir.BitwiseXor(a, ir.BitwiseAnd(b, c));
+ case 121:
+ return ir.BitwiseXor(ir.BitwiseAnd(a, ir.BitwiseOr(b, c)),
+ ir.BitwiseXor(b, ir.BitwiseNot(c)));
+ case 122:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, ir.BitwiseNot(b)), ir.BitwiseXor(a, c));
+ case 123:
+ return ir.BitwiseOr(ir.BitwiseNot(b), ir.BitwiseXor(a, c));
+ case 124:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, ir.BitwiseNot(c)), ir.BitwiseXor(a, b));
+ case 125:
+ return ir.BitwiseOr(ir.BitwiseNot(c), ir.BitwiseXor(a, b));
+ case 126:
+ return ir.BitwiseOr(ir.BitwiseXor(a, b), ir.BitwiseXor(a, c));
+ case 127:
+ return ir.BitwiseNot(ir.BitwiseAnd(a, ir.BitwiseAnd(b, c)));
+ case 128:
+ return ir.BitwiseAnd(a, ir.BitwiseAnd(b, c));
+ case 129:
+ return ir.BitwiseNot(ir.BitwiseOr(ir.BitwiseXor(a, b), ir.BitwiseXor(a, c)));
+ case 130:
+ return ir.BitwiseAnd(c, ir.BitwiseXor(a, ir.BitwiseNot(b)));
+ case 131:
+ return ir.BitwiseAnd(ir.BitwiseOr(c, ir.BitwiseNot(a)), ir.BitwiseXor(a, ir.BitwiseNot(b)));
+ case 132:
+ return ir.BitwiseAnd(b, ir.BitwiseXor(a, ir.BitwiseNot(c)));
+ case 133:
+ return ir.BitwiseAnd(ir.BitwiseOr(b, ir.BitwiseNot(a)), ir.BitwiseXor(a, ir.BitwiseNot(c)));
+ case 134:
+ return ir.BitwiseAnd(ir.BitwiseOr(b, c), ir.BitwiseXor(a, ir.BitwiseXor(b, c)));
+ case 135:
+ return ir.BitwiseXor(ir.BitwiseAnd(b, c), ir.BitwiseNot(a));
+ case 136:
+ return ir.BitwiseAnd(b, c);
+ case 137:
+ return ir.BitwiseAnd(ir.BitwiseOr(b, ir.BitwiseNot(a)), ir.BitwiseXor(b, ir.BitwiseNot(c)));
+ case 138:
+ return ir.BitwiseAnd(c, ir.BitwiseOr(b, ir.BitwiseNot(a)));
+ case 139:
+ return ir.BitwiseOr(ir.BitwiseAnd(b, c), ir.BitwiseNot(ir.BitwiseOr(a, b)));
+ case 140:
+ return ir.BitwiseAnd(b, ir.BitwiseOr(c, ir.BitwiseNot(a)));
+ case 141:
+ return ir.BitwiseOr(ir.BitwiseAnd(b, c), ir.BitwiseNot(ir.BitwiseOr(a, c)));
+ case 142:
+ return ir.BitwiseXor(a, ir.BitwiseOr(ir.BitwiseXor(a, b), ir.BitwiseXor(a, c)));
+ case 143:
+ return ir.BitwiseOr(ir.BitwiseAnd(b, c), ir.BitwiseNot(a));
+ case 144:
+ return ir.BitwiseAnd(a, ir.BitwiseXor(b, ir.BitwiseNot(c)));
+ case 145:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, ir.BitwiseNot(b)), ir.BitwiseXor(b, ir.BitwiseNot(c)));
+ case 146:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, c), ir.BitwiseXor(a, ir.BitwiseXor(b, c)));
+ case 147:
+ return ir.BitwiseXor(ir.BitwiseAnd(a, c), ir.BitwiseNot(b));
+ case 148:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, b), ir.BitwiseXor(a, ir.BitwiseXor(b, c)));
+ case 149:
+ return ir.BitwiseXor(ir.BitwiseAnd(a, b), ir.BitwiseNot(c));
+ case 150:
+ return ir.BitwiseXor(a, ir.BitwiseXor(b, c));
+ case 151:
+ return ir.BitwiseOr(ir.BitwiseNot(ir.BitwiseOr(a, b)),
+ ir.BitwiseXor(a, ir.BitwiseXor(b, c)));
+ case 152:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, b), ir.BitwiseXor(b, ir.BitwiseNot(c)));
+ case 153:
+ return ir.BitwiseXor(b, ir.BitwiseNot(c));
+ case 154:
+ return ir.BitwiseXor(c, ir.BitwiseAnd(a, ir.BitwiseNot(b)));
+ case 155:
+ return ir.BitwiseNot(ir.BitwiseAnd(ir.BitwiseOr(a, b), ir.BitwiseXor(b, c)));
+ case 156:
+ return ir.BitwiseXor(b, ir.BitwiseAnd(a, ir.BitwiseNot(c)));
+ case 157:
+ return ir.BitwiseNot(ir.BitwiseAnd(ir.BitwiseOr(a, c), ir.BitwiseXor(b, c)));
+ case 158:
+ return ir.BitwiseOr(ir.BitwiseAnd(b, c), ir.BitwiseXor(a, ir.BitwiseOr(b, c)));
+ case 159:
+ return ir.BitwiseNot(ir.BitwiseAnd(a, ir.BitwiseXor(b, c)));
+ case 160:
+ return ir.BitwiseAnd(a, c);
+ case 161:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, ir.BitwiseNot(b)), ir.BitwiseXor(a, ir.BitwiseNot(c)));
+ case 162:
+ return ir.BitwiseAnd(c, ir.BitwiseOr(a, ir.BitwiseNot(b)));
+ case 163:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, c), ir.BitwiseNot(ir.BitwiseOr(a, b)));
+ case 164:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, b), ir.BitwiseXor(a, ir.BitwiseNot(c)));
+ case 165:
+ return ir.BitwiseXor(a, ir.BitwiseNot(c));
+ case 166:
+ return ir.BitwiseXor(c, ir.BitwiseAnd(b, ir.BitwiseNot(a)));
+ case 167:
+ return ir.BitwiseNot(ir.BitwiseAnd(ir.BitwiseOr(a, b), ir.BitwiseXor(a, c)));
+ case 168:
+ return ir.BitwiseAnd(c, ir.BitwiseOr(a, b));
+ case 169:
+ return ir.BitwiseXor(ir.BitwiseNot(c), ir.BitwiseOr(a, b));
+ case 170:
+ return c;
+ case 171:
+ return ir.BitwiseOr(c, ir.BitwiseNot(ir.BitwiseOr(a, b)));
+ case 172:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, b), ir.BitwiseOr(c, ir.BitwiseNot(a)));
+ case 173:
+ return ir.BitwiseOr(ir.BitwiseAnd(b, c), ir.BitwiseXor(a, ir.BitwiseNot(c)));
+ case 174:
+ return ir.BitwiseOr(c, ir.BitwiseAnd(b, ir.BitwiseNot(a)));
+ case 175:
+ return ir.BitwiseOr(c, ir.BitwiseNot(a));
+ case 176:
+ return ir.BitwiseAnd(a, ir.BitwiseOr(c, ir.BitwiseNot(b)));
+ case 177:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, c), ir.BitwiseNot(ir.BitwiseOr(b, c)));
+ case 178:
+ return ir.BitwiseXor(b, ir.BitwiseOr(ir.BitwiseXor(a, b), ir.BitwiseXor(a, c)));
+ case 179:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, c), ir.BitwiseNot(b));
+ case 180:
+ return ir.BitwiseXor(a, ir.BitwiseAnd(b, ir.BitwiseNot(c)));
+ case 181:
+ return ir.BitwiseNot(ir.BitwiseAnd(ir.BitwiseOr(b, c), ir.BitwiseXor(a, c)));
+ case 182:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, c), ir.BitwiseXor(b, ir.BitwiseOr(a, c)));
+ case 183:
+ return ir.BitwiseNot(ir.BitwiseAnd(b, ir.BitwiseXor(a, c)));
+ case 184:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, b), ir.BitwiseOr(c, ir.BitwiseNot(b)));
+ case 185:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, c), ir.BitwiseXor(b, ir.BitwiseNot(c)));
+ case 186:
+ return ir.BitwiseOr(c, ir.BitwiseAnd(a, ir.BitwiseNot(b)));
+ case 187:
+ return ir.BitwiseOr(c, ir.BitwiseNot(b));
+ case 188:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, c), ir.BitwiseXor(a, b));
+ case 189:
+ return ir.BitwiseOr(ir.BitwiseXor(a, b), ir.BitwiseXor(a, ir.BitwiseNot(c)));
+ case 190:
+ return ir.BitwiseOr(c, ir.BitwiseXor(a, b));
+ case 191:
+ return ir.BitwiseOr(c, ir.BitwiseNot(ir.BitwiseAnd(a, b)));
+ case 192:
+ return ir.BitwiseAnd(a, b);
+ case 193:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, ir.BitwiseNot(c)), ir.BitwiseXor(a, ir.BitwiseNot(b)));
+ case 194:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, c), ir.BitwiseXor(a, ir.BitwiseNot(b)));
+ case 195:
+ return ir.BitwiseXor(a, ir.BitwiseNot(b));
+ case 196:
+ return ir.BitwiseAnd(b, ir.BitwiseOr(a, ir.BitwiseNot(c)));
+ case 197:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, b), ir.BitwiseNot(ir.BitwiseOr(a, c)));
+ case 198:
+ return ir.BitwiseXor(b, ir.BitwiseAnd(c, ir.BitwiseNot(a)));
+ case 199:
+ return ir.BitwiseNot(ir.BitwiseAnd(ir.BitwiseOr(a, c), ir.BitwiseXor(a, b)));
+ case 200:
+ return ir.BitwiseAnd(b, ir.BitwiseOr(a, c));
+ case 201:
+ return ir.BitwiseXor(ir.BitwiseNot(b), ir.BitwiseOr(a, c));
+ case 202:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, c), ir.BitwiseOr(b, ir.BitwiseNot(a)));
+ case 203:
+ return ir.BitwiseOr(ir.BitwiseAnd(b, c), ir.BitwiseXor(a, ir.BitwiseNot(b)));
+ case 204:
+ return b;
+ case 205:
+ return ir.BitwiseOr(b, ir.BitwiseNot(ir.BitwiseOr(a, c)));
+ case 206:
+ return ir.BitwiseOr(b, ir.BitwiseAnd(c, ir.BitwiseNot(a)));
+ case 207:
+ return ir.BitwiseOr(b, ir.BitwiseNot(a));
+ case 208:
+ return ir.BitwiseAnd(a, ir.BitwiseOr(b, ir.BitwiseNot(c)));
+ case 209:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, b), ir.BitwiseNot(ir.BitwiseOr(b, c)));
+ case 210:
+ return ir.BitwiseXor(a, ir.BitwiseAnd(c, ir.BitwiseNot(b)));
+ case 211:
+ return ir.BitwiseNot(ir.BitwiseAnd(ir.BitwiseOr(b, c), ir.BitwiseXor(a, b)));
+ case 212:
+ return ir.BitwiseXor(c, ir.BitwiseOr(ir.BitwiseXor(a, b), ir.BitwiseXor(a, c)));
+ case 213:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, b), ir.BitwiseNot(c));
+ case 214:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, b), ir.BitwiseXor(c, ir.BitwiseOr(a, b)));
+ case 215:
+ return ir.BitwiseNot(ir.BitwiseAnd(c, ir.BitwiseXor(a, b)));
+ case 216:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, c), ir.BitwiseOr(b, ir.BitwiseNot(c)));
+ case 217:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, b), ir.BitwiseXor(b, ir.BitwiseNot(c)));
+ case 218:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, b), ir.BitwiseXor(a, c));
+ case 219:
+ return ir.BitwiseOr(ir.BitwiseXor(a, c), ir.BitwiseXor(a, ir.BitwiseNot(b)));
+ case 220:
+ return ir.BitwiseOr(b, ir.BitwiseAnd(a, ir.BitwiseNot(c)));
+ case 221:
+ return ir.BitwiseOr(b, ir.BitwiseNot(c));
+ case 222:
+ return ir.BitwiseOr(b, ir.BitwiseXor(a, c));
+ case 223:
+ return ir.BitwiseOr(b, ir.BitwiseNot(ir.BitwiseAnd(a, c)));
+ case 224:
+ return ir.BitwiseAnd(a, ir.BitwiseOr(b, c));
+ case 225:
+ return ir.BitwiseXor(ir.BitwiseNot(a), ir.BitwiseOr(b, c));
+ case 226:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, ir.BitwiseNot(b)), ir.BitwiseOr(b, c));
+ case 227:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, c), ir.BitwiseXor(a, ir.BitwiseNot(b)));
+ case 228:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, ir.BitwiseNot(c)), ir.BitwiseOr(b, c));
+ case 229:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, b), ir.BitwiseXor(a, ir.BitwiseNot(c)));
+ case 230:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, b), ir.BitwiseXor(b, c));
+ case 231:
+ return ir.BitwiseOr(ir.BitwiseXor(a, ir.BitwiseNot(b)), ir.BitwiseXor(b, c));
+ case 232:
+ return ir.BitwiseAnd(ir.BitwiseOr(a, b), ir.BitwiseOr(c, ir.BitwiseAnd(a, b)));
+ case 233:
+ return ir.BitwiseOr(ir.BitwiseAnd(a, b),
+ ir.BitwiseXor(ir.BitwiseNot(c), ir.BitwiseOr(a, b)));
+ case 234:
+ return ir.BitwiseOr(c, ir.BitwiseAnd(a, b));
+ case 235:
+ return ir.BitwiseOr(c, ir.BitwiseXor(a, ir.BitwiseNot(b)));
+ case 236:
+ return ir.BitwiseOr(b, ir.BitwiseAnd(a, c));
+ case 237:
+ return ir.BitwiseOr(b, ir.BitwiseXor(a, ir.BitwiseNot(c)));
+ case 238:
+ return ir.BitwiseOr(b, c);
+ case 239:
+ return ir.BitwiseOr(ir.BitwiseNot(a), ir.BitwiseOr(b, c));
+ case 240:
+ return a;
+ case 241:
+ return ir.BitwiseOr(a, ir.BitwiseNot(ir.BitwiseOr(b, c)));
+ case 242:
+ return ir.BitwiseOr(a, ir.BitwiseAnd(c, ir.BitwiseNot(b)));
+ case 243:
+ return ir.BitwiseOr(a, ir.BitwiseNot(b));
+ case 244:
+ return ir.BitwiseOr(a, ir.BitwiseAnd(b, ir.BitwiseNot(c)));
+ case 245:
+ return ir.BitwiseOr(a, ir.BitwiseNot(c));
+ case 246:
+ return ir.BitwiseOr(a, ir.BitwiseXor(b, c));
+ case 247:
+ return ir.BitwiseOr(a, ir.BitwiseNot(ir.BitwiseAnd(b, c)));
+ case 248:
+ return ir.BitwiseOr(a, ir.BitwiseAnd(b, c));
+ case 249:
+ return ir.BitwiseOr(a, ir.BitwiseXor(b, ir.BitwiseNot(c)));
+ case 250:
+ return ir.BitwiseOr(a, c);
+ case 251:
+ return ir.BitwiseOr(ir.BitwiseNot(b), ir.BitwiseOr(a, c));
+ case 252:
+ return ir.BitwiseOr(a, b);
+ case 253:
+ return ir.BitwiseOr(ir.BitwiseNot(c), ir.BitwiseOr(a, b));
+ case 254:
+ return ir.BitwiseOr(a, ir.BitwiseOr(b, c));
+ case 255:
+ return ir.Imm32(0xFFFFFFFF);
+ // end of generated code
}
- if (ttbl & 0x02) {
- // r |= ~a & ~b & c;
- const auto lhs{ir.BitwiseAnd(not_a, not_b)};
- const auto rhs{ir.BitwiseAnd(lhs, c)};
- r = ir.BitwiseOr(r, rhs);
- }
- if (ttbl & 0x04) {
- // r |= ~a & b & ~c;
- const auto lhs{ir.BitwiseAnd(not_a, b)};
- const auto rhs{ir.BitwiseAnd(lhs, not_c)};
- r = ir.BitwiseOr(r, rhs);
- }
- if (ttbl & 0x08) {
- // r |= ~a & b & c;
- const auto lhs{ir.BitwiseAnd(not_a, b)};
- const auto rhs{ir.BitwiseAnd(lhs, c)};
- r = ir.BitwiseOr(r, rhs);
- }
- if (ttbl & 0x10) {
- // r |= a & ~b & ~c;
- const auto lhs{ir.BitwiseAnd(a, not_b)};
- const auto rhs{ir.BitwiseAnd(lhs, not_c)};
- r = ir.BitwiseOr(r, rhs);
- }
- if (ttbl & 0x20) {
- // r |= a & ~b & c;
- const auto lhs{ir.BitwiseAnd(a, not_b)};
- const auto rhs{ir.BitwiseAnd(lhs, c)};
- r = ir.BitwiseOr(r, rhs);
- }
- if (ttbl & 0x40) {
- // r |= a & b & ~c;
- const auto lhs{ir.BitwiseAnd(a, b)};
- const auto rhs{ir.BitwiseAnd(lhs, not_c)};
- r = ir.BitwiseOr(r, rhs);
- }
- if (ttbl & 0x80) {
- // r |= a & b & c;
- const auto lhs{ir.BitwiseAnd(a, b)};
- const auto rhs{ir.BitwiseAnd(lhs, c)};
- r = ir.BitwiseOr(r, rhs);
- }
- return r;
+ throw NotImplementedException("LOP3 with out of range ttbl");
}
IR::U32 LOP3(TranslatorVisitor& v, u64 insn, const IR::U32& op_b, const IR::U32& op_c, u64 lut) {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/logic_operation_three_input_lut3.py b/src/shader_recompiler/frontend/maxwell/translate/impl/logic_operation_three_input_lut3.py
new file mode 100644
index 000000000..8f547c266
--- /dev/null
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/logic_operation_three_input_lut3.py
@@ -0,0 +1,92 @@
+# Copyright © 2022 degasus <markus@selfnet.de>
+# This work is free. You can redistribute it and/or modify it under the
+# terms of the Do What The Fuck You Want To Public License, Version 2,
+# as published by Sam Hocevar. See http://www.wtfpl.net/ for more details.
+
+from itertools import product
+
+# The primitive instructions
+OPS = {
+ 'ir.BitwiseAnd({}, {})' : (2, 1, lambda a,b: a&b),
+ 'ir.BitwiseOr({}, {})' : (2, 1, lambda a,b: a|b),
+ 'ir.BitwiseXor({}, {})' : (2, 1, lambda a,b: a^b),
+ 'ir.BitwiseNot({})' : (1, 0.1, lambda a: (~a) & 255), # Only tiny cost, as this can often inlined in other instructions
+}
+
+# Our database of combination of instructions
+optimized_calls = {}
+def cmp(lhs, rhs):
+ if lhs is None: # new entry
+ return True
+ if lhs[3] > rhs[3]: # costs
+ return True
+ if lhs[3] < rhs[3]: # costs
+ return False
+ if len(lhs[0]) > len(rhs[0]): # string len
+ return True
+ if len(lhs[0]) < len(rhs[0]): # string len
+ return False
+ if lhs[0] > rhs[0]: # string sorting
+ return True
+ if lhs[0] < rhs[0]: # string sorting
+ return False
+ assert lhs == rhs, "redundant instruction, bug in brute force"
+ return False
+def register(imm, instruction, count, latency):
+ # Use the sum of instruction count and latency as costs to evaluate which combination is best
+ costs = count + latency
+
+ old = optimized_calls.get(imm, None)
+ new = (instruction, count, latency, costs)
+
+ # Update if new or better
+ if cmp(old, new):
+ optimized_calls[imm] = new
+ return True
+
+ return False
+
+# Constants: 0, 1 (for free)
+register(0, 'ir.Imm32(0)', 0, 0)
+register(255, 'ir.Imm32(0xFFFFFFFF)', 0, 0)
+
+# Inputs: a, b, c (for free)
+ta = 0xF0
+tb = 0xCC
+tc = 0xAA
+inputs = {
+ ta : 'a',
+ tb : 'b',
+ tc : 'c',
+}
+for imm, instruction in inputs.items():
+ register(imm, instruction, 0, 0)
+ register((~imm) & 255, 'ir.BitwiseNot({})'.format(instruction), 0.099, 0.099) # slightly cheaper NEG on inputs
+
+# Try to combine two values from the db with an instruction.
+# If it is better than the old method, update it.
+while True:
+ registered = 0
+ calls_copy = optimized_calls.copy()
+ for OP, (argc, cost, f) in OPS.items():
+ for args in product(calls_copy.items(), repeat=argc):
+ # unpack(transponse) the arrays
+ imm = [arg[0] for arg in args]
+ value = [arg[1][0] for arg in args]
+ count = [arg[1][1] for arg in args]
+ latency = [arg[1][2] for arg in args]
+
+ registered += register(
+ f(*imm),
+ OP.format(*value),
+ sum(count) + cost,
+ max(latency) + cost)
+ if registered == 0:
+ # No update at all? So terminate
+ break
+
+# Hacky output. Please improve me to output valid C++ instead.
+s = """ case {imm}:
+ return {op};"""
+for imm in range(256):
+ print(s.format(imm=imm, op=optimized_calls[imm][0]))
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/move_register.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/move_register.cpp
index 6bb08db8a..c317e14c9 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/move_register.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/move_register.cpp
@@ -5,7 +5,6 @@
#include "common/bit_field.h"
#include "common/common_types.h"
#include "shader_recompiler/exception.h"
-#include "shader_recompiler/frontend/maxwell/opcodes.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/impl.h"
namespace Shader::Maxwell {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/surface_atomic_operations.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/surface_atomic_operations.cpp
index 63b588ad4..100c94c19 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/surface_atomic_operations.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/surface_atomic_operations.cpp
@@ -2,9 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include <array>
-#include <bit>
-
#include "common/bit_field.h"
#include "common/common_types.h"
#include "shader_recompiler/frontend/ir/modifiers.h"
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_fetch_swizzled.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_fetch_swizzled.cpp
index 154e7f1a1..00dbcd86f 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_fetch_swizzled.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_fetch_swizzled.cpp
@@ -2,8 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include <utility>
-
#include "common/bit_field.h"
#include "common/common_types.h"
#include "shader_recompiler/frontend/ir/modifiers.h"
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gather.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gather.cpp
index 218cbc1a8..959c63ba9 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gather.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gather.cpp
@@ -2,8 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include <optional>
-
#include "common/bit_field.h"
#include "common/common_types.h"
#include "shader_recompiler/frontend/ir/modifiers.h"
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gather_swizzled.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gather_swizzled.cpp
index 34efa2d50..86e68a830 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gather_swizzled.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gather_swizzled.cpp
@@ -2,8 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include <utility>
-
#include "common/bit_field.h"
#include "common/common_types.h"
#include "shader_recompiler/frontend/ir/modifiers.h"
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gradient.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gradient.cpp
index c3fe3ffda..6f4feff11 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gradient.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_gradient.cpp
@@ -2,8 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include <optional>
-
#include "common/bit_field.h"
#include "common/common_types.h"
#include "shader_recompiler/frontend/ir/modifiers.h"
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_load.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_load.cpp
index 983058303..60732215b 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_load.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_load.cpp
@@ -2,8 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include <optional>
-
#include "common/bit_field.h"
#include "common/common_types.h"
#include "shader_recompiler/frontend/ir/modifiers.h"
@@ -134,7 +132,7 @@ void Impl(TranslatorVisitor& v, u64 insn, bool is_bindless) {
multisample = v.X(meta_reg++);
}
if (tld.clamp != 0) {
- throw NotImplementedException("TLD.CL - CLAMP is not implmented");
+ throw NotImplementedException("TLD.CL - CLAMP is not implemented");
}
IR::TextureInstInfo info{};
info.type.Assign(GetType(tld.type));
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_mipmap_level.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_mipmap_level.cpp
index aea3c0e62..f89ce1b68 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_mipmap_level.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_mipmap_level.cpp
@@ -2,8 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include <optional>
-
#include "common/bit_field.h"
#include "common/common_types.h"
#include "shader_recompiler/frontend/ir/modifiers.h"
@@ -83,7 +81,7 @@ void Impl(TranslatorVisitor& v, u64 insn, bool is_bindless) {
} const tmml{insn};
if ((tmml.mask & 0b1100) != 0) {
- throw NotImplementedException("TMML BA results are not implmented");
+ throw NotImplementedException("TMML BA results are not implemented");
}
const IR::Value coords{MakeCoords(v, tmml.coord_reg, tmml.type)};
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_query.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_query.cpp
index 0459e5473..72131301c 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/texture_query.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/texture_query.cpp
@@ -6,7 +6,6 @@
#include "common/bit_field.h"
#include "common/common_types.h"
-#include "shader_recompiler/frontend/ir/modifiers.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/impl.h"
namespace Shader::Maxwell {
diff --git a/src/shader_recompiler/frontend/maxwell/translate/impl/warp_shuffle.cpp b/src/shader_recompiler/frontend/maxwell/translate/impl/warp_shuffle.cpp
index 550fed55c..f98f66940 100644
--- a/src/shader_recompiler/frontend/maxwell/translate/impl/warp_shuffle.cpp
+++ b/src/shader_recompiler/frontend/maxwell/translate/impl/warp_shuffle.cpp
@@ -2,8 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include <optional>
-
#include "common/bit_field.h"
#include "common/common_types.h"
#include "shader_recompiler/frontend/maxwell/translate/impl/impl.h"