struct S_atomic { int x; uint a; uint y; }; static uint local_invocation_index_1 = 0u; groupshared S_atomic wg[10]; void compute_main_inner(uint local_invocation_index) { uint idx = 0u; idx = local_invocation_index; { [loop] for(; !(!((idx < 10u))); idx = (idx + 1u)) { const uint x_28 = idx; wg[x_28].x = 0; uint atomic_result = 0u; InterlockedExchange(wg[x_28].a, 0u, atomic_result); wg[x_28].y = 0u; } } GroupMemoryBarrierWithGroupSync(); uint atomic_result_1 = 0u; InterlockedExchange(wg[4].a, 1u, atomic_result_1); return; } void compute_main_1() { compute_main_inner(local_invocation_index_1); return; } struct tint_symbol_1 { uint local_invocation_index_1_param : SV_GroupIndex; }; void compute_main_inner_1(uint local_invocation_index_1_param) { { [loop] for(uint idx_1 = local_invocation_index_1_param; (idx_1 < 10u); idx_1 = (idx_1 + 1u)) { const uint i = idx_1; wg[i].x = 0; uint atomic_result_2 = 0u; InterlockedExchange(wg[i].a, 0u, atomic_result_2); wg[i].y = 0u; } } GroupMemoryBarrierWithGroupSync(); local_invocation_index_1 = local_invocation_index_1_param; compute_main_1(); } [numthreads(1, 1, 1)] void compute_main(tint_symbol_1 tint_symbol) { compute_main_inner_1(tint_symbol.local_invocation_index_1_param); return; }