arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1712.04892cs.AR
收藏

作为非线性优化的加速器协同设计

Accelerator Codesign as Non-Linear Optimization

Nirmal Prajapati, Sanjay Rajopadhye, Hristo Djidjev, Nandkishore Santhi, Tobias Grosser, Rumen Andonov

摘要:

我们提出了一种优化方法,用于确定硬件和软件参数的高效率的实施(家庭)的应用程序称为密集模板计算可编程GPGPU。我们首先介绍了一个简单的,分析模型的硅面积使用的加速器架构和模板计算的工作负载表征。我们结合这个特性与参数化的执行时间模型,并制定一个数学优化问题。该问题寻求最大化“所有硬件和软件参数”的共同目标函数。因此,解决这个问题的方案“解决”了协同设计问题:同时选择软件-硬件参数以优化总体性能。我们通过提出NVIDIA Maxwell GTX-980(分别为Titan X)的架构变体来验证这种方法,这些架构变体专门针对四种常见的2D stentleman(Heat,Jacobi,Laplacian和Gradient)和两种3D stentleman(Heat和Laplacian)的预定工作负载进行调整。我们的模型预测,通过对硬件参数进行简单的调整,例如通过改变流式多处理器的数量和每个处理器包含的计算核心的数量来调整粗粒度和细粒度并行性,性能可能会提高28%(分别为33%)。我们提出了一组帕累托最优设计点,以利用性能和硅面积之间的权衡,并表明通过额外消除GPU缓存,我们可以得到进一步的2倍改进。

英文摘要:

We propose an optimization approach for determining both hardware and software parameters for the efficient implementation of a (family of) applications called dense stencil computations on programmable GPGPUs. We first introduce a simple, analytical model for the silicon area usage of accelerator architectures and a workload characterization of stencil computations. We combine this characterization with a parametric execution time model and formulate a mathematical optimization problem. That problem seeks to maximize a common objective function of 'all the hardware and software parameters'. The solution to this problem, therefore "solves" the codesign problem: simultaneously choosing software-hardware parameters to optimize total performance. We validate this approach by proposing architectural variants of the NVIDIA Maxwell GTX-980 (respectively, Titan X) specifically tuned to a predetermined workload of four common 2D stencils (Heat, Jacobi, Laplacian, and Gradient) and two 3D ones (Heat and Laplacian). Our model predicts that performance would potentially improve by 28% (respectively, 33%) with simple tweaks to the hardware parameters such as adapting coarse and fine-grained parallelism by changing the number of streaming multiprocessors and the number of compute cores each contains. We propose a set of Pareto-optimal design points to exploit the trade-off between performance and silicon area and show that by additionally eliminating GPU caches, we can get a further 2-fold improvement.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑