张量处理单元的学习性能模型
A Learned Performance Model for Tensor Processing Units
摘要：
准确的硬件性能模型对于高效的代码生成至关重要。它们可以被编译器用来做启发式决策，被超级优化器用来作为最小化目标，或者被自动调优器用来为特定程序找到最优配置。然而，它们很难开发，因为当代处理器很复杂，最近深度学习加速器的激增增加了开发负担。我们展示了一种从张量处理单元（TPU）实例的张量计算图程序语料库中学习性能模型的方法。我们表明，我们的学习模型在两个任务上的性能优于经过高度优化的分析性能模型-瓷砖大小选择和运营商融合-并且它可以帮助自动调谐器在TPU访问受限或昂贵的设置中发现更快的程序。
英文摘要：
Accurate hardware performance models are critical to efficient code generation. They can be used by compilers to make heuristic decisions, by superoptimizers as a minimization objective, or by autotuners to find an optimal configuration for a specific program. However, they are difficult to develop because contemporary processors are complex, and the recent proliferation of deep learning accelerators has increased the development burden. We demonstrate a method of learning performance models from a corpus of tensor computation graph programs for Tensor Processing Unit (TPU) instances. We show that our learned model outperforms a heavily-optimized analytical performance model on two tasks -- tile-size selection and operator fusion -- and that it helps an autotuner discover faster programs in a setting where access to TPUs is limited or expensive.
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。
