| #extension GL_AMD_gpu_shader_half_float : require |
| layout(binding = 0, std140) uniform a_block_std140_ubo { |
| f16mat4x3 conv_mat4x3_f16(mat4x3_f16 val) { |
| return f16mat4x3(val.col0, val.col1, val.col2, val.col3); |
| f16mat4x3[4] conv_arr4_mat4x3_f16(mat4x3_f16 val[4]) { |
| f16mat4x3 arr[4] = f16mat4x3[4](f16mat4x3(0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf), f16mat4x3(0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf), f16mat4x3(0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf), f16mat4x3(0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf, 0.0hf)); |
| for(uint i = 0u; (i < 4u); i = (i + 1u)) { |
| arr[i] = conv_mat4x3_f16(val[i]); |
| f16vec3 load_a_inner_p0_p1(uint p0, uint p1) { |
| f16mat4x3 p_a[4] = conv_arr4_mat4x3_f16(a.inner); |
| f16mat4x3 p_a_i = conv_mat4x3_f16(a.inner[tint_symbol]); |
| f16vec3 p_a_i_i = load_a_inner_p0_p1(uint(tint_symbol), uint(tint_symbol_1)); |
| f16mat4x3 l_a[4] = conv_arr4_mat4x3_f16(a.inner); |
| f16mat4x3 l_a_i = conv_mat4x3_f16(a.inner[tint_symbol]); |
| f16vec3 l_a_i_i = load_a_inner_p0_p1(uint(tint_symbol), uint(tint_symbol_1)); |
| layout(local_size_x = 1, local_size_y = 1, local_size_z = 1) in; |