Modern mobile application processors are required to execute heavier workloads while the battery capacity is rarely increased. This trend leads to the need for a power model that can analyze the power consumed by CPU and GPU at run-time, which are the key components of the application processor in terms of power savings. We propose novel CPU and GPU power models based on the phases using performance monitoring counters for smartphones. Our phase-based power models employ combined per-phase power modeling methods to achieve more accurate power consumption estimations, unlike existing power models. The proposed CPU power model shows estimation errors of 2.51% for ARM Cortex A-53 and 1.97% for Samsung M1 on average, and the proposed GPU power model shows an average error of 8.92% for the Mali-T880. In addition, we integrate proposed CPU and GPU models with the latest display power model into a holistic power model. Our holistic power model can estimate the smartphone′s total power consumption with an error of 6.36% on average while running nine 3D game benchmarks, improving the error rate by about 56% compared with the latest prior model.
Smartphones have limited battery capacity, so efficient power management is required for high-performance applications and to increase usage time. In recent years, efficient power management of smartphones has become very important as the demand for power use of smartphones has grown due to deep learning, games, virtual reality, and augmented reality applications. Existing low-power techniques of smartphones focus only on lowering power consumption without considering actual power consumption based on utilization of the central processing unit (CPU) and graphics processing unit (GPU), which are major components of smartphones. In addition, they do not take into consideration the strict use of resources within the component and what instructions are being processed to operate them. In this paper, we propose a low-power technique that manages power by calculating the actual power consumption of smartphones at execution time and classifying the detailed resource operating states of CPUs and GPUs. The proposed technique was implemented by linking the kernel and native app on a Galaxy S7 smartphone equipped with Android. In experiments with 15 workloads, the proposed technique achieves an energy reduction of 18.11% compared to the low-power technique of the interactive governor built into the Galaxy S7 with a small FPS reduction of 3.12%.
This work presents a shared-column im2col design for energy-efficient TinyML Conv2D. By storing overlapping input regions once and using shifted indexing, it reduces redundant SRAM traffic while preserving identical outputs. Implemented on Arduino Nano 33 BLE, the method achieves consistent execution-time and energy reductions on CIFAR-10 and SqueezeNet models with parity, padding, and stride conditions varied, demonstrating efficiency gains.
As AI research advances, various performance optimization techniques have been studied for TinyML models on microcontrollers with very limited resources. In TinyML frameworks such as TFLM (TensorFlow Lite Micro) and NNOM (Neural Network on Microcontrollers), their execution engines mainly depend on CMSIS-NN for inference acceleration, which is an ARM’s back-end library to execute optimized kernel functions for performance optimization. But we often find that CMSIS-NN is not invincible for inference time optimization on TinyML. In this paper, we examine TinyML frameworks and their CMSIS-NN libraries and consider how to improve CMSIS-NN in terms of runtime. Then, we propose the D2I technique to reduce the overhead of memory operations that occur while performing the Im2col procedure within the convolution function, which takes most of the inference time in CMSIS-NN. The proposed technique creates a necessary index table, finds the location of the input with the corresponding index, and performs direct operations between filters and inputs. Thus, it can quite mitigate data copy operations in Im2col with a small additional amount of memory compared to Im2col. In extensive experiments using an Arduino nano 33 BLE board with Cortex-M4 and an STM32F746G-DISCO board with Cortex-M7, D2I was found to achieve about 16.3% and 14.5% inference time improvements against the Im2col in TFLM’s and NNOM’s CMSIS-NNs, respectively, for the SqueezeNet model. And the additional memory usage was shown to be identically 11.52 kB.