Adaptively Periodic I/O Scheduling for Concurrent HPC Applications

Abstract

With the convergence of big data and HPC (high-performance computing), various machine learning applications and traditional large-scale simulations with a stochastically iterative I/O periodicity are running concurrently on HPC platforms, which poses more challenges on the scarcely shared I/O resources due to the ever-growing data transfer demand. Currently the existing heuristic online and periodic offline I/O scheduling methods for traditional HPC applications with a fixed I/O periodicity are not suitable for the applications with stochastically iterative I/O periodicities, which are required to schedule the concurrent I/Os from different applications under I/O congestion. In this work, we propose an adaptively periodic I/O scheduling (APIO) method that optimizes the system efficiency and application dilation by taking the stochastically iterative I/O periodicity of the applications into account. We first build a periodic offline scheduling method within a specified duration to capture the iterative nature. After that, APIO adjusts the bandwidth allocation to resist stochasticity based on the actual length of the computing phrase. In the case where the specified duration does not satisfy the actual running requirements, the period length will be extended to adapt to the actual duration. Theoretical analysis and extensive simulations demonstrate the efficiency of our proposed I/O scheduling method over the existing online approach.

Revisiting I/O bandwidth-sharing strategies for HPC applications

“…The main limitation is of course the assumed periodicity of each application. An extension is provided by other authors in [28], where applications still consist of phases with work followed by I/O transfers, but now CPU phases have stochastic lengths taken from some probability distribution, while I/O phases have constant length. As a motivation, for CPU phases, we can think of a constant amount of flops to perform, with some system-dependent or data-dependent noise, while for I/O transfers, we can think of a fixed-size checkpoint operation.…”

“…Several researchers have already identified and addressed this problem (see [7,12,25,2,28,27] among others). Performance degradation due to I/O is already significant for current state-of-the-art platforms and is expected to worsen due to the faster increase in processing speed than in I/O bandwidth [22].…”

Revisiting I/O Bandwidth-Sharing Strategies for Hpc Applications

“…Several researchers have already identified and addressed this problem (see [7,12,25,2,28,27] among others). Performance degradation due to I/O is already significant for current state-of-the-art platforms and is expected to worsen due to the faster increase in processing speed than in I/O bandwidth [22].…”