The past few years have witnessed significant growth in the computational capabilities of GPUs. The race for computing performance makes the uses of many-core accelerators more necessary. However, GPUs consume a significant amount of energy as compared with CPUs. One way to reduce the energy consumption is to scale the speed and/or voltage of the processor. Typically, the faster the processor runs, the faster we finish jobs, but the more power is required by the processor. It is hence important to balance between performance and power consumption. In this paper, we consider the following scheduling problem. We have a set of jobs to be assigned to different processors. Each job may have different characteristics depending on the type of processor that it is assigned to. The goal is to minimize the total energy consumption. After proving the NP-hardness of this problem, we propose a constant approximation algorithm for the case when processors can scale to any continuous speed. When processors have a set of discrete speeds, we propose a heuristic algorithm and compare with some classical scheduling algorithms experimentally. Then, we extend this heuristic to the online case where jobs arrive over time. Our simulation results show that the proposed heuristic algorithms are effective and can achieve near-optimal performance.

In many IoT applications, information needs to be gathered from multiple heterogeneous sources to the base station for real-time processing and follow-up actions. Undoubtedly, information freshness, measured by age of information (AoI), is critical in taking responsive actions. Recent studies have taken AoI into the consideration of transmission scheduling over wireless channels. However, existing studies on guaranteeing AoI either assume error-free wireless channels or priorly known link reliability, which is unrealistic. In this paper, we tackle the AoI-guaranteed transmission scheduling problem over an unreliable channel with the aim of throughput maximization, which is modelled as an AoI-Guaranteed Multi-Armed Bandit (AG-MAB) problem. Since the problem has not been studied in the literature even for the oracle case with given link reliability, we first propose an optimal stationary randomized sampling (SRS) policy for the oracle case. For the AG-MAB problem with unknown link reliability, we propose learning algorithms that meet the AoI requirements with probability 1 and incur sublinear regret compared to Oracle SRS, which can also detect the unsatisfiability of the AoI constraint and switch to the fallback policy promptly with guaranteed accuracy. Numerical results show that our algorithm outperforms the AoI-constraint-aware baselines on throughput with per-source AoI requirement guaranteed.