| #include <metal_stdlib> |
| using namespace metal; |
| struct strided_arr { |
| float el; |
| }; |
| template<typename T, size_t N> |
| struct tint_array { |
| const constant T& operator[](size_t i) const constant { return elements[i]; } |
| device T& operator[](size_t i) device { return elements[i]; } |
| const device T& operator[](size_t i) const device { return elements[i]; } |
| thread T& operator[](size_t i) thread { return elements[i]; } |
| const thread T& operator[](size_t i) const thread { return elements[i]; } |
| threadgroup T& operator[](size_t i) threadgroup { return elements[i]; } |
| const threadgroup T& operator[](size_t i) const threadgroup { return elements[i]; } |
| T elements[N]; |
| }; |
| |
| struct strided_arr_1 { |
| tint_array<tint_array<strided_arr, 2>, 3> el; |
| }; |
| struct S { |
| tint_array<strided_arr_1, 4> a; |
| }; |
| struct tint_module_vars_struct { |
| device S* s; |
| }; |
| |
| void tint_store_and_preserve_padding_4(device strided_arr* const target, strided_arr value_param) { |
| (*target).el = value_param.el; |
| } |
| void tint_store_and_preserve_padding_3(device tint_array<strided_arr, 2>* const target, tint_array<strided_arr, 2> value_param) { |
| { |
| uint v = 0u; |
| v = 0u; |
| while(true) { |
| uint const v_1 = v; |
| if ((v_1 >= 2u)) { |
| break; |
| } |
| tint_store_and_preserve_padding_4((&(*target)[v_1]), value_param[v_1]); |
| v = (v_1 + 1u); |
| continue; |
| } |
| } |
| } |
| void tint_store_and_preserve_padding_2(device tint_array<tint_array<strided_arr, 2>, 3>* const target, tint_array<tint_array<strided_arr, 2>, 3> value_param) { |
| { |
| uint v_2 = 0u; |
| v_2 = 0u; |
| while(true) { |
| uint const v_3 = v_2; |
| if ((v_3 >= 3u)) { |
| break; |
| } |
| tint_store_and_preserve_padding_3((&(*target)[v_3]), value_param[v_3]); |
| v_2 = (v_3 + 1u); |
| continue; |
| } |
| } |
| } |
| void tint_store_and_preserve_padding_1(device strided_arr_1* const target, strided_arr_1 value_param) { |
| tint_store_and_preserve_padding_2((&(*target).el), value_param.el); |
| } |
| void tint_store_and_preserve_padding(device tint_array<strided_arr_1, 4>* const target, tint_array<strided_arr_1, 4> value_param) { |
| { |
| uint v_4 = 0u; |
| v_4 = 0u; |
| while(true) { |
| uint const v_5 = v_4; |
| if ((v_5 >= 4u)) { |
| break; |
| } |
| tint_store_and_preserve_padding_1((&(*target)[v_5]), value_param[v_5]); |
| v_4 = (v_5 + 1u); |
| continue; |
| } |
| } |
| } |
| void f_1(tint_module_vars_struct tint_module_vars) { |
| tint_array<strided_arr_1, 4> const x_19 = (*tint_module_vars.s).a; |
| tint_array<tint_array<strided_arr, 2>, 3> const x_24 = (*tint_module_vars.s).a[3].el; |
| tint_array<strided_arr, 2> const x_28 = (*tint_module_vars.s).a[3].el[2]; |
| float const x_32 = (*tint_module_vars.s).a[3].el[2][1].el; |
| tint_store_and_preserve_padding((&(*tint_module_vars.s).a), tint_array<strided_arr_1, 4>{}); |
| (*tint_module_vars.s).a[3].el[2][1].el = 5.0f; |
| } |
| kernel void f(device S* s [[buffer(0)]]) { |
| tint_module_vars_struct const tint_module_vars = tint_module_vars_struct{.s=s}; |
| f_1(tint_module_vars); |
| } |