A common approach to performance estimation of queueing system models is based on point estimators. Confidence intervals provide more insight on the system performance, and basically a regenerative structure of the studied process is required. However, classical regeneration epochs (arrivals into empty system) are too rare (or even might not exist) in complex queueing systems (such as high performance computing clusters), thus are inappropriate for practical usage. Instead, we propose the so-called artificial regeneration based on exponential splitting, which provide the desired estimators with reasonable accuracy. We illustrate the approach considering a multiserver model with energy efficiency management by activating the so-called sleep mode of servers which we hope is relevant to energy/performance tradeoff in cloud computing.