The problem of SPARQL query containment has recently attracted a lot of attention due to its fundamental purpose in query optimization and information integration. New approaches to this problem, have been put forth, that can be implemented in practice. However, these approaches suffer from various limitations: coverage (size and type of queries), response time (how long it takes to determine containment), and the technique applied to encode the problem. In order to experimentally assess implementation limitations, we designed a benchmark suite offering different experimental settings depending on the type of queries, projection and reasoning (RDFS). We have applied this benchmark to three available systems using different techniques highlighting the strengths and weaknesses of such systems.