# Performance Evaluation and Modeling of HPC I/O on Non-Volatile Memory

## TL;DR

This paper evaluates and models the performance of HPC I/O systems with non-volatile memory, revealing key impacts on caching, I/O methods, and collective operations, guiding future HPC storage optimization.

## Contribution

It provides a comprehensive performance analysis and modeling of HPC I/O with NVM, highlighting the effects on page cache, I/O methods, and collective I/O performance.

## Key findings

- Page cache effects diminish with NVM.
- Minor performance difference between MPI and POSIX I/O.
- Collective I/O performance is worse due to data shuffling.

## Abstract

HPC applications pose high demands on I/O performance and storage capability. The emerging non-volatile memory (NVM) techniques offer low-latency, high bandwidth, and persistence for HPC applications. However, the existing I/O stack are designed and optimized based on an assumption of disk-based storage. To effectively use NVM, we must re-examine the existing high performance computing (HPC) I/O sub-system to properly integrate NVM into it. Using NVM as a fast storage, the previous assumption on the inferior performance of storage (e.g., hard drive) is not valid any more. The performance problem caused by slow storage may be mitigated; the existing mechanisms to narrow the performance gap between storage and CPU may be unnecessary and result in large overhead. Thus fully understanding the impact of introducing NVM into the HPC software stack demands a thorough performance study. In this paper, we analyze and model the performance of I/O intensive HPC applications with NVM as a block device. We study the performance from three perspectives: (1) the impact of NVM on the performance of traditional page cache; (2) a performance comparison between MPI individual I/O and POSIX I/O; and (3) the impact of NVM on the performance of collective I/O. We reveal the diminishing effects of page cache, minor performance difference between MPI individual I/O and POSIX I/O, and performance disadvantage of collective I/O on NVM due to unnecessary data shuffling. We also model the performance of MPI collective I/O and study the complex interaction between data shuffling, storage performance, and I/O access patterns.