Add an option to SPIRV-Cross to enforce invariant floating point math to prevent different depth calculation between prepass & basepass when running on Metal 2.0 and earlier.

spirv_msl.cpp

@@ -977,7 +977,30 @@ void CompilerMSL::preprocess_op_codes()
 	/* UE Change Begin: Allow Metal to use the array<T> template to make arrays a value type */
 	// UnsafeArray
 	{
-		add_header_line("	");
+		if (msl_options.invariant_float_math)
+		{
+			add_header_line("template <typename T> T fmul(T l, T r) { return metal::fma(l, r, T(0)); }");
+			add_header_line("template <typename T> T fadd(T l, T r) { return metal::fma(T(1), l, r); }");
+			
+			add_header_line("template <typename T, int Cols, int Rows> metal::vec<T, Rows> fmul_mv(metal::matrix<T, Cols, Rows> m, metal::vec<T, Cols> v) { metal::vec<T, Rows> res = metal::vec<T, Rows>(0); for(uint i = Cols; i > 0; --i) { res = metal::fma(m[i-1], metal::vec<T, Rows>(v[i-1]), res); } return res; }");
+			
+			add_header_line("template <typename T, int LCols, int LRows, int RCols, int RRows>");
+			add_header_line("metal::matrix<T, RCols, LRows> fmul_mat(metal::matrix<T, LCols, LRows> l, metal::matrix<T, RCols, RRows> r)");
+			add_header_line("{");
+			add_header_line("	metal::matrix<T, RCols, LRows> res;");
+			add_header_line("	for(uint i = 0; i < RCols; i++)");
+			add_header_line("	{");
+			add_header_line("		metal::vec<T, RCols> tmp(0);");
+			add_header_line("		for(uint j = 0; j < LCols; j++)");
+			add_header_line("		{");
+			add_header_line("			tmp = metal::fma(metal::vec<T, RCols>(r[i][j]),l[j],tmp);");
+			add_header_line("		}");
+			add_header_line("		res[i] = tmp;");
+			add_header_line("	}");
+			add_header_line("	return res;");
+			add_header_line("}");
+		}
+		
 		add_header_line("template <typename T, size_t Num>");
 		add_header_line("struct unsafe_array");
 		add_header_line("{");
@@ -5360,6 +5383,20 @@ void CompilerMSL::emit_instruction(const Instruction &instruction)
 		MSL_BFOP(fmod);
 		break;
 			
+	case OpFMul:
+		if (msl_options.invariant_float_math)
+			MSL_BFOP(fmul);
+		else
+			MSL_BOP(*);
+		break;
+			
+	case OpFAdd:
+		if (msl_options.invariant_float_math)
+			MSL_BFOP(fadd);
+		else
+			MSL_BOP(+);
+		break;
+
 	// Atomics
 	case OpAtomicExchange:
 	{
@@ -5782,14 +5819,31 @@ void CompilerMSL::emit_instruction(const Instruction &instruction)
 		uint32_t a = ops[2];
 		uint32_t b = ops[3];
 		
+		auto *type_a = maybe_get<SPIRType>(a);
+		auto *type_b = maybe_get<SPIRType>(b);
+
 		auto &type = get<SPIRType>(result_type);
 		string expr = type_to_glsl_constructor(type);
 		expr += "(";
 		for (uint32_t col = 0; col < type.columns; col++)
+		{
+			if (msl_options.invariant_float_math && type_a && type_b &&
+				( (is_matrix(*type_a) && is_matrix(*type_b)) ||
+				  (is_matrix(*type_a) && is_vector(*type_b)) ||
+				  (is_vector(*type_a) && is_matrix(*type_b)) ) )
+			{
+				expr += "fmul_mat(";
+				expr += to_enclosed_expression(a);
+				expr += ", ";
+				expr += to_extract_component_expression(b, col);
+				expr += ")";
+			}
+			else
 			{
 				expr += to_enclosed_expression(a);
 				expr += " * ";
 				expr += to_extract_component_expression(b, col);
+			}
 			if (col + 1 < type.columns)
 				expr += ", ";
 		}

spirv_msl.hpp

@@ -299,6 +299,8 @@ public:
 		bool enforce_storge_buffer_bounds = false;
 		/* UE Change End: Storage buffer robustness - clamps access to SSBOs to the size of the buffer */
 		
+		bool invariant_float_math = false;
+
 		// Requires MSL 2.1, use the native support for texel buffers.
 		bool texture_buffer_native = false;