| cbuffer cbuffer_data : register(b0) { |
| uint4 data[2]; |
| }; |
| |
| matrix<float16_t, 3, 3> data_load_1(uint offset) { |
| const uint scalar_offset = ((offset + 0u)) / 4; |
| uint4 ubo_load_1 = data[scalar_offset / 4]; |
| uint2 ubo_load = ((scalar_offset & 2) ? ubo_load_1.zw : ubo_load_1.xy); |
| vector<float16_t, 2> ubo_load_xz = vector<float16_t, 2>(f16tof32(ubo_load & 0xFFFF)); |
| float16_t ubo_load_y = f16tof32(ubo_load[0] >> 16); |
| const uint scalar_offset_1 = ((offset + 8u)) / 4; |
| uint4 ubo_load_3 = data[scalar_offset_1 / 4]; |
| uint2 ubo_load_2 = ((scalar_offset_1 & 2) ? ubo_load_3.zw : ubo_load_3.xy); |
| vector<float16_t, 2> ubo_load_2_xz = vector<float16_t, 2>(f16tof32(ubo_load_2 & 0xFFFF)); |
| float16_t ubo_load_2_y = f16tof32(ubo_load_2[0] >> 16); |
| const uint scalar_offset_2 = ((offset + 16u)) / 4; |
| uint4 ubo_load_5 = data[scalar_offset_2 / 4]; |
| uint2 ubo_load_4 = ((scalar_offset_2 & 2) ? ubo_load_5.zw : ubo_load_5.xy); |
| vector<float16_t, 2> ubo_load_4_xz = vector<float16_t, 2>(f16tof32(ubo_load_4 & 0xFFFF)); |
| float16_t ubo_load_4_y = f16tof32(ubo_load_4[0] >> 16); |
| return matrix<float16_t, 3, 3>(vector<float16_t, 3>(ubo_load_xz[0], ubo_load_y, ubo_load_xz[1]), vector<float16_t, 3>(ubo_load_2_xz[0], ubo_load_2_y, ubo_load_2_xz[1]), vector<float16_t, 3>(ubo_load_4_xz[0], ubo_load_4_y, ubo_load_4_xz[1])); |
| } |
| |
| void main() { |
| uint2 ubo_load_6 = data[1].zw; |
| vector<float16_t, 2> ubo_load_6_xz = vector<float16_t, 2>(f16tof32(ubo_load_6 & 0xFFFF)); |
| float16_t ubo_load_6_y = f16tof32(ubo_load_6[0] >> 16); |
| vector<float16_t, 3> x = mul(data_load_1(0u), vector<float16_t, 3>(ubo_load_6_xz[0], ubo_load_6_y, ubo_load_6_xz[1])); |
| return; |
| } |