abstract: We investigate the problem of quantum search on a noisy quantum computer. Taking a fault-ignorant approach, we design quantum algorithm that solve the task for various different noise strengths, possibly unknown beforehand. One rationale is to avoid costly overheads, such as traditional quantum error correction. Proving lower bounds on algorithm runtimes, which may depend on the actual level of noise, we find that the quadratic speedup is lost. Nevertheless, for low noise levels, our algorithms outperform the best noiseless classical search algorithm.