| cbuffer cbuffer_u : register(b0, space0) { |
| uint4 u[2]; |
| }; |
| RWByteAddressBuffer s : register(u1, space0); |
| |
| void tint_symbol(RWByteAddressBuffer buffer, uint offset, matrix<float16_t, 3, 4> value) { |
| buffer.Store<vector<float16_t, 4> >((offset + 0u), value[0u]); |
| buffer.Store<vector<float16_t, 4> >((offset + 8u), value[1u]); |
| buffer.Store<vector<float16_t, 4> >((offset + 16u), value[2u]); |
| } |
| |
| matrix<float16_t, 3, 4> tint_symbol_2(uint4 buffer[2], uint offset) { |
| const uint scalar_offset = ((offset + 0u)) / 4; |
| uint4 ubo_load_1 = buffer[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)); |
| vector<float16_t, 2> ubo_load_yw = vector<float16_t, 2>(f16tof32(ubo_load >> 16)); |
| const uint scalar_offset_1 = ((offset + 8u)) / 4; |
| uint4 ubo_load_3 = buffer[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)); |
| vector<float16_t, 2> ubo_load_2_yw = vector<float16_t, 2>(f16tof32(ubo_load_2 >> 16)); |
| const uint scalar_offset_2 = ((offset + 16u)) / 4; |
| uint4 ubo_load_5 = buffer[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)); |
| vector<float16_t, 2> ubo_load_4_yw = vector<float16_t, 2>(f16tof32(ubo_load_4 >> 16)); |
| return matrix<float16_t, 3, 4>(vector<float16_t, 4>(ubo_load_xz[0], ubo_load_yw[0], ubo_load_xz[1], ubo_load_yw[1]), vector<float16_t, 4>(ubo_load_2_xz[0], ubo_load_2_yw[0], ubo_load_2_xz[1], ubo_load_2_yw[1]), vector<float16_t, 4>(ubo_load_4_xz[0], ubo_load_4_yw[0], ubo_load_4_xz[1], ubo_load_4_yw[1])); |
| } |
| |
| [numthreads(1, 1, 1)] |
| void f() { |
| tint_symbol(s, 0u, tint_symbol_2(u, 0u)); |
| uint2 ubo_load_6 = u[0].xy; |
| vector<float16_t, 2> ubo_load_6_xz = vector<float16_t, 2>(f16tof32(ubo_load_6 & 0xFFFF)); |
| vector<float16_t, 2> ubo_load_6_yw = vector<float16_t, 2>(f16tof32(ubo_load_6 >> 16)); |
| s.Store<vector<float16_t, 4> >(8u, vector<float16_t, 4>(ubo_load_6_xz[0], ubo_load_6_yw[0], ubo_load_6_xz[1], ubo_load_6_yw[1])); |
| uint2 ubo_load_7 = u[0].xy; |
| vector<float16_t, 2> ubo_load_7_xz = vector<float16_t, 2>(f16tof32(ubo_load_7 & 0xFFFF)); |
| vector<float16_t, 2> ubo_load_7_yw = vector<float16_t, 2>(f16tof32(ubo_load_7 >> 16)); |
| s.Store<vector<float16_t, 4> >(8u, vector<float16_t, 4>(ubo_load_7_xz[0], ubo_load_7_yw[0], ubo_load_7_xz[1], ubo_load_7_yw[1]).ywxz); |
| s.Store<float16_t>(2u, float16_t(f16tof32(((u[0].z) & 0xFFFF)))); |
| return; |
| } |