Techniques for orchestrating stages of thread synchronization

Apparatuses, systems, and techniques to execute data-dependent parallel operations in one or more programs utilizing an application programming interface to perform parallel computing, such as CUDA, without relying on a synchronization operation between said one or more programs. For example, at least one embodiment pertains to processors or computing systems used to determine 416 which thread 406, 418 in a group 402 of threads is last to complete modification of shared data, and that thread may be selected to perform additional data-dependent computations from said group of threads.