發(fā)布時(shí)間：2018-04-21 03:03

  本文選題：負(fù)載預(yù)測(cè)調(diào)度算法 + CPU�。� 參考：《上海大學(xué)》2016年博士論文

【摘要】：由于性價(jià)比和能效比很高,多核CPU-GPU計(jì)算平臺(tái)得到了廣泛應(yīng)用,這也使系統(tǒng)內(nèi)同時(shí)存在兩種異構(gòu)的計(jì)算資源。但是多核CPU和GPU的性能必須通過(guò)高效的調(diào)度算法才能得到充分發(fā)揮。因此如何充分利用異構(gòu)資源的計(jì)算能力,如何實(shí)現(xiàn)負(fù)載均衡成為研究的熱點(diǎn)。傳統(tǒng)的調(diào)度方法有靜態(tài)調(diào)度和動(dòng)態(tài)調(diào)度。靜態(tài)調(diào)度開銷非常小,但容易導(dǎo)致負(fù)載不均衡,降低計(jì)算資源的利用率;動(dòng)態(tài)調(diào)度能更好地實(shí)現(xiàn)負(fù)載均衡,但調(diào)度開銷比較大。如果將上述兩種調(diào)度方法結(jié)合起來(lái),將大大減少調(diào)度開銷,并有效地實(shí)現(xiàn)負(fù)載均衡。CPU-GPU異構(gòu)計(jì)算平臺(tái)中,基于SIMD結(jié)構(gòu)的GPU適合并行度和計(jì)算量大的計(jì)算任務(wù),GPU的計(jì)算性能遠(yuǎn)遠(yuǎn)大于CPU的計(jì)算性能,但是現(xiàn)有的調(diào)度算法無(wú)法根據(jù)硬件特點(diǎn)進(jìn)行任務(wù)分配。本文針對(duì)上述問題,提出了一種新的調(diào)度方法--負(fù)載預(yù)測(cè)調(diào)度算法(Load-prediction scheduling--LPS),該算法可以充分發(fā)揮異構(gòu)的多核CPU和GPU的計(jì)算能力,并實(shí)現(xiàn)靜態(tài)和動(dòng)態(tài)調(diào)度的有效結(jié)合。本文完成的主要工作包括:1、本文提出了負(fù)載預(yù)測(cè)調(diào)度算法,該算法具有以下特點(diǎn):(1)根據(jù)GPU硬件特點(diǎn)分配任務(wù),充分發(fā)揮GPU計(jì)算性能。(2)有效結(jié)合了動(dòng)態(tài)調(diào)度和靜態(tài)調(diào)度,實(shí)現(xiàn)了負(fù)載均衡和減少調(diào)度開銷,適合應(yīng)用在異構(gòu)環(huán)境中。(3)充分發(fā)揮多核CPU的計(jì)算性能。2、將負(fù)載預(yù)測(cè)調(diào)度算法應(yīng)用在心電仿真計(jì)算中,實(shí)現(xiàn)了上述特點(diǎn)。此外在計(jì)算中通過(guò)負(fù)載預(yù)測(cè)消除了分支,同時(shí)提高了GPU的計(jì)算粒度,因此拓寬了GPU的計(jì)算范圍,進(jìn)一步提高了計(jì)算效率。3、將負(fù)載預(yù)測(cè)調(diào)度算法應(yīng)用在多體問題計(jì)算中。
[Abstract]:Multi-core CPU-GPU computing platform has been widely used because of its high performance-to-price ratio and energy-efficiency ratio, which also makes two heterogeneous computing resources exist in the system at the same time. But the performance of multi-core CPU and GPU must be achieved by efficient scheduling algorithm. Therefore, how to make full use of the computing power of heterogeneous resources and how to achieve load balancing has become a hot topic. Traditional scheduling methods include static scheduling and dynamic scheduling. Static scheduling overhead is very small, but it is easy to lead to load imbalance, reduce the utilization of computing resources, dynamic scheduling can better achieve load balancing, but scheduling overhead is relatively large. If the above two scheduling methods are combined, the scheduling overhead will be greatly reduced, and the load balancing. CPU-GPU heterogeneous computing platform will be implemented effectively. GPU based on SIMD structure is suitable for computing tasks with high degree of parallelism and large amount of computation. The performance of GPU is much higher than that of CPU, but the existing scheduling algorithms can not assign tasks according to the hardware characteristics. In this paper, a new scheduling method, Load-Prediction scheduling algorithm, is proposed in this paper. The algorithm can give full play to the computing power of heterogeneous multi-core CPU and GPU, and realize the effective combination of static and dynamic scheduling. The main work accomplished in this paper includes: 1. This paper proposes a load prediction scheduling algorithm. The algorithm has the following characteristics: 1) assign tasks according to the hardware characteristics of GPU, give full play to the performance of GPU computing. 2) effectively combine dynamic scheduling with static scheduling. It realizes load balancing and reducing scheduling overhead. It is suitable for application in heterogeneous environment. It can give full play to the computing performance of multi-core CPU. The load predictive scheduling algorithm is applied to ECG simulation, and the above characteristics are realized. In addition, the branch is eliminated by load prediction, and the computational granularity of GPU is improved. Therefore, the computational range of GPU is widened, and the computational efficiency is further improved. The load prediction scheduling algorithm is applied to the computation of multi-body problem.
【學(xué)位授予單位】：上海大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2016
【分類號(hào)】：TP301.6

【參考文獻(xiàn)】

相關(guān)期刊論文 前2條

1 肖漢;張祖勛;;基于GPGPU的并行影像匹配算法[J];測(cè)繪學(xué)報(bào);2010年01期

2 王惠春;朱定局;曹學(xué)年;樊建平;;基于SMP集群的混合并行編程模型研究[J];計(jì)算機(jī)工程;2009年03期

相關(guān)碩士學(xué)位論文 前2條

1 陳信;基于經(jīng)濟(jì)模型的網(wǎng)格資源調(diào)度算法研究[D];山東師范大學(xué);2010年

2 曹婷婷;基于多核處理器串行程序并行化改造和性能優(yōu)化[D];西南交通大學(xué);2009年

，

本文編號(hào)：1780629