Performance Per Watt Benefits of Dynamic Core Morphing in Asymmetric Multicores

Abstract

Abstract-The trend toward multicore processors is moving the emphasis in computation from sequential to parallel processing. However, not all applications can be parallelized and benefit from multiple cores. Such applications lead to under-utilization of parallel resources, hence sub-optimal performance/watt. They may however, benefit from powerful uniprocessors. On the other hand, not all applications can take advantage of more powerful uniprocessors. To address competing requirements of diverse applications, we propose a heterogeneous multicore architecture with a Dynamic Core Morphing (DCM) capability. Depending on the computational demands of the currently executing applications, the resources of a few tightly coupled cores are morphed at runtime. We present a simple hardware-based algorithm to monitor the time-varying computational needs of the application and when deemed beneficial, trigger reconfiguration of the cores at fine-grain time scales to maximize the performance/watt of the application. The proposed dynamic scheme is then compared against a baseline static heterogeneous multicore configuration and an equivalent homogeneous configuration. Our results show that dynamic morphing of cores can provide performance/watt gains of 43% and 16% on an average, when compared to the homogeneous and baseline heterogeneous configurations, respectively.

Reducing Energy per Instruction via Dynamic Resource Allocation and Voltage and Frequency Adaptation in Asymmetric Multicores

“…To illustrate the efficiency of our proposed scheme, we compare the EPI metric of our scheme against three baseline heterogeneous configurations -static, with DVFA capability only and with DRA feature only [14]. Besides the performance/power metric (EPI), we also evaluate the stand-alone performance in terms of weighted improvement in throughput using our scheme.…”

“…The size of the floating-point and integer core parameters were taken from [14], as the proposed scheme is compared against their heterogeneous baseline configuration with DRA-only capability. The execution latencies for the cores were taken from [13] (see Table I).…”

“…Static morphing of the cores can not suit all the different phases in an application. In [14], Rodrigues et al presented dynamic morphing of cores at runtime to maximize performance/Watt. In comparison, our proposed scheme could both dynamically allocate resources and scale/boost the voltage/frequency of the individual cores.…”

“…The DVFA capability of our scheme not only helps in reducing the energy consumption but also in efficiently redistributing the saved power to the other cores in order to increase the overall system's throughput. Moreover, improving performance and performance/Watt are targeted in [13] and [14], respectively, while the objective of the proposed scheme is to minimize EPI. Finally, we present an in-depth analysis of the impact of reconfiguration overhead in this work.…”

Does the Sharing of Execution Units Improve Performance/Power of Multicores?

“…Dynamic exchange of execution units between pairs of cores was investigated in [Rodrigues et al 2011;2013]. Here, depending on the current workload characteristics, the cores may exchange execution units to maximize performance/Watt.…”