Detecting I/O Access Patterns of HPC Workloads at Runtime

Abstract

In this paper, we seek to guide optimization and tuning strategies by identifying the application's I/O access pattern. We evaluate three machine learning techniques to automatically detect the I/O access pattern of HPC applications at runtime: decision trees, random forests, and neural networks. We focus on the detection using metrics from file-level accesses as seen by the clients, I/O nodes, and parallel file system servers. We evaluated these detection strategies in a case study in which the accurate detection of the current access pattern is fundamental to adjust a parameter of an I/O scheduling algorithm. We demonstrate that such approaches correctly classify the access pattern, regarding file layout and spatiality of accesses-into the most common ones used by the community and by I/O benchmarking tools to test new I/O optimization-with up to 99% precision. Furthermore, when applied to our study case, it guides a tuning mechanism to achieve 99% of the performance of an Oracle solution.

Adaptive request scheduling for the I/O forwarding layer using reinforcement learning

“…Specifically to our case study, we classified the access pattern regarding the operation (read or write), spatiality (contiguous or 1D-strided), number of accessed files, and request size aspects, since our extensive performance evaluation of TWINS showed these to uniquely identify the different behaviors. The classification of the spatiality uses a neural network described in our previous work [17].…”

Capturing Periodic I/O Using Frequency Techniques

“…The main advantages of FTIO in this comparison basically stream from the unique properties of DFT. Compared to popular machine learning (ML) approaches from the time domain, like neuronal networks (NN) [43] and LSTMs [44], [45], decision trees and other supervised methods [46], or a combination of supervised and unsupervised techniques [20], FTIO, and in particular DFT, does not require a learning phase. Additionally, FTIO does not require past system logs, different from recent regressionbased approaches [47] and other strategies [20], [44]- [48], Moreover, compared to approaches that predict future I/O activity, such as ARIMA [41], DFT does not require defining several thresholds and parameter estimations.…”

“…More general characterization efforts usually focus on aspects such as spatiality and request size [49], [50], using information from MPI-IO [51]- [53], ML-based methods [43], etc. In contrast, FTIO focuses on the temporal behavior (specifically on the periodicity), and hence is also complementary to those.…”

Executando na Nuvem um Detector de Padrões de Acesso de E/S

“…Desta forma, evitamos sobrecarregar ainda mais o sistema, que já é muito concorrido pelas aplicac ¸ões que estão executando. Adaptamos o modelo de rede neural implementado por Bez et al [Bez et al 2019], a fim de possibilitar o treinamento em dispositivos Tensor Processing Unit (TPU) em Nuvem no Google Cloud Platform.…”