An integrated GPU power and performance model

Abstract

GPU architectures are increasingly important in the multi-core era due to their high number of parallel processors. Performance optimization for multi-core processors has been a challenge for programmers. Furthermore, optimizing for power consumption is even more difficult. Unfortunately, as a result of the high number of processors, the power consumption of many-core processors such as GPUs has increased significantly. Hence, in this paper, we propose an integrated power and performance (IPP) prediction model for a GPU architecture to predict the optimal number of active processors for a given application. The basic intuition is that when an application reaches the peak memory bandwidth, using more cores does not result in performance improvement. We develop an empirical power model for the GPU. Unlike most previous models, which require measured execution times, hardware performance counters, or architectural simulations, IPP predicts execution times to calculate dynamic power events. We then use the outcome of IPP to control the number of running cores. We also model the increases in power consumption that resulted from the increases in temperature. With the predicted optimal number of active cores, we show that we can save up to 22.09% of runtime GPU energy consumption and on average 10.99% of that for the five memory bandwidth-limited benchmarks.

Hybrid CMOS-TFET based register files for energy-efficient GPGPUs

“…The read/write times to CMOS-and TFET-based registers and the total execution time are collected from the modified GPGPU-Sim to evaluate both RF dynamic and leakage energy consumption. Our energy estimation is consistent with previous studies [3,4,8].…”

“…Accessing such sizeable register files leads to massive power consumption [2][3][4][5][6]. It has been reported that the register files consume 15%-20% of the GPU stream multiprocessor's power [8]. Effectively optimizing the register files power consumption is critical and the first step towards the energy-efficient GPUs.…”

Parallel strategy for Optimal Power Efficiency

“…We use average power P to substitute P(t) and estimate the energy consumed. So E is a sum of product of average power , and T obtained by the performance counter, namely EPT =× [9] .…”

Untitled

“…Hong and Kim [30] propose an integrated GPU power and performance analysis model which can be applied without performance measurements. Zhang [43] and Chen [21] use similar strategies to statistically correlate the GPU power consumption and its execution behaviors.…”