A modern general-purpose graphics processing unit (GPGPU) usually consists of multiple streaming multiprocessors (SMs), each having a pipeline that incorporates a group of threads executing a common instruction flow. Although SMs are designed to work independently, we observe that they tend to exhibit very similar behavior for many workloads. If multiple SMs can be grouped and work in the lock-step manner, it is possible to save energy by sharing the front-end units among multiple SMs, including the instruction fetch, decode, and schedule components. However, such sharing brings architectural challenges and sometime causes performance degradation. In this article, we show our design, implementation, and evaluation for such an architecture, which we call Buddy SM. Specifically, multiple SMs can be opportunistically grouped into a buddy cluster. One SM becomes the master, and the rest become the slaves. The front-end unit of the master works actively for itself as well as for the slaves, whereas the front-end logics of the slaves are power gated. For efficient flow control and program correctness, the proposed architecture can identify unfavorable conditions and ungroup the buddy cluster when necessary. We analyze various techniques to improve the performance and energy efficiency of Buddy SM. Detailed experiments manifest that 37.2% front-end and 7.5% total GPU energy reduction can be achieved. This article is extended from a six-page conference paper entitled "Dynamic Front-End Sharing in Graphics Processing Units," presented at the 32nd IEEE International Conference on Computer Design (ICCD). We do the further research and extend the conference paper in several directions: -We propose an adaptive buddy cluster formation technique that can improve performance and save an average of 10% more energy than that in the conference paper (Sections 4.8 and 6.3). -We construct simple large warp architectures and an eight-buddy architecture to compare with two-buddy and four-buddy architectures (Section 6.5). We evaluate additional regrouping strategies (Sections 4.6 and 6.2). -We run additional experiments to evaluate the characteristics of the benchmarks and categorize them so that we can thoroughly analyze and explain the final performance results (Section 5.2). -We perform new experiments to investigate the direct impact of the Buddy SM architecture on applications:issuing stalls and the memory access latency (Section 6.4). -We add a background section to introduce the components in the SM front-end and the mechanism of instruction issue to help readers understanding of this work better (Section 3).