Search NASA⌕ Search

Engineering topics

Khan, Awais

Publications and source records attributed to Khan, Awais.

18 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

An Evaluation of the Effect of Network Cost Optimization for Leadership Class Supercomputers

Dragonfly-based networks are an extensively deployed network topology in large-scale high-performance computing due to their cost-effectiveness and efficiency. The US will soon have three Exascale supercomputers for leadership class workloads deployed using dragonfly networks. Compared to indirect networks of similar scale, the dragonfly network has considerably reduced cable lengths, cable counts, and switch counts, resulting in significant network cost savings for a given system size, however, these cost reductions result in reduced global minimal paths and more challenging routing. Additionally, large scale dragonfly networks often require a taper at the global link level, resulting in less bisection bandwidth than is achievable in other traditional non-blocking topologies of equivalent scale. While dragonfly networks have been extensively studied, they have yet to be fully evaluated in an extreme scale (i.e., exascale) system that targets capability workloads. In this paper, we present the results of the first large scale evaluation of a dragonfly network on an exascale system (Frontier) and compare its behavior to a similar scale fat-tree network on a previous generation TOP500 system (Summit). This evaluation aims to determine the effect of network cost optimizations by measuring a tapered topology’s impact on capability workloads. Our evaluation is based on a collection of synthetic microbenchmarks, mini-apps, and full scale applications. It compares the scaling efficiencies of each benchmark between the dragonfly-based Frontier and the fat-tree-based Summit systems. Our results show that a dragonfly network is $\sim \mathbf{3 0 \%}$ more cost efficient than a fat-tree topology, which amortizes to $\sim 3 \%$ of an exascale system cost. Furthermore, while tapered dragonfly networks impose significant tradeoffs, the impacts are not as broad as initially thought and are mostly seen in applications with global communication patterns, particularly all-to-all (e.g., FFT-based algorithms), but also local communication patterns (e.g., nearest-neighbor algorithms) that are sensitive to network performance variability.

Khan, Awais↗
DOE OSTI2024

Fault-Tolerant Deep Learning Cache with Hash Ring for Load Balancing in HPC Systems

Large-scale DL on HPC systems like Frontier and Summit uses distributed node-local caching to address scalability and performance challenges. However, as these systems grow more complex, the risk of node failures increases, and current caching approaches lack fault tolerance, jeopardizing large-scale training jobs. We analyzed six months of SLURM job logs from Frontier and found that over 30% of jobs failed after an average of 75 minutes. To address this, we propose fault-tolerance strategies that recache data lost from failed nodes using a hash ring technique for balanced data recaching in the distributed node-local caching, reducing reliance on the PFS. Our extensive evaluations on Frontier showed that the hash ring-based recaching approach reduced training time by approximately 25% compared to the approach that redirects I/O to the PFS after node failures and demonstrated effective load balancing of training data across nodes.

Lee, Seoyeong↗
DOE OSTI2024

Accelerating Application Bulk Synchronous Writes in HPC Environments

High-bandwidth storage tiers are becoming more common for their capability to absorb high-rate, bursty I/Os. Notably, the designs of these fast storage tiers differ from system to system. The variation of these layers and non-uniform methods of access can pose chal- lenges for applications seeking to run at multiple HPC facilities. Therefore, in this work, we present Spectral, a rapid-output ab- straction library to accelerate application, bulk-synchronous writes on HPC systems. We design Spectral to enable applications to use high-bandwidth storage, such as node-local storage and dis- tributed, write-caches (e.g., burst buffers) transparently without requiring modifications to the application or file system source code. The key idea is to allow applications to spend most of the time performing productive work and to not require any source code changes for maximum portability on different HPC archi- tectures. Spectral internally re-routes write-only files through available, high-performance I/O resources before ultimately mi- grating them to the shared global parallel file system. For instance, on Summit, Spectral transparently places application outputs on node-local storage and then utilizes asynchronous migration to the center-wide GPFS file system. We evaluate Spectral on the Summit HPC system (1024 nodes) using the IOR benchmark and real scientific applications. Spectral shows linear performance scaling, improving application write performance by over an order of magnitude when compared to GPFS.

Khan, Awais↗
DOE OSTI2024

Summit Darshan Archival Dataset

Summit Darshan Archival Dataset contains 2021 Summit Darshan log data for 25 applications and is grouped into science domains. The dataset is processed, and all the propriety fields are anonymized. The resultant data is converted into a tabular structure and saved in parquet file format. In this notebook, we demonstrate how to access the data. Data Organization: The data is organized into two directories: Darshan total (`darshan_total`): List all the high levels generated by the `darshan-parser --total` command on `.darshan` files. There is one parquet file for each application. Note: `uid` and `exe` field are masked Darshan detail (`darshan_detail`): This data contains detailed job level log information extracted by command `darshan-parser` on the raw `.darshan` files. The data is sorted by directory hierarchy in the order of `year/month/day (2021/12/07)`. For instance, to get the data for a `job_id` 3819766 of application `App11`, which was executed on `2021-12-07`can be accessed as follows. Note:`uid` and `filename` fields are masked

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

OctoFAS: A Two-Level Fair Scheduler That Increases Fairness in Network-Based Key-Value Storage

We identified a fairness problem in a network-based key-value storage system using Intel Storage Performance Development Kit (SPDK) in a multitenant environment. In such an environment, each tenant’s I/O service rate is not fairly guaranteed compared to that of other tenants. To address the fairness problem, we propose OctoFAS, a two-level fair scheduler designed to improve overall throughput and fairness among tenants. The two-level scheduler of OctoFAS consists of (i) inter-core scheduling and (ii) intra-core scheduling. Through inter-core scheduling, OctoFAS addresses the load imbalance problem that is inherent in SPDK on the storage server by dynamically migrating I/O requests from overloaded cores to underloaded cores, thereby increasing overall throughput. Intra-core scheduling prioritizes handling requests from starving tenants over well-fed tenants within core-specific event queues to ensure fair I/O services among multiple tenants. OctoFAS is deployed on a Linux cluster with SPDK. Through extensive evaluations, we found that OctoFAS ensures that the total system throughput remains high and balanced, while enhancing fairness by approximately 10% compared to the baseline, when both scheduling levels operate in a hybrid fashion.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

OCTOKV: An Agile Network-Based Key-Value Storage System with Robust Load Orchestration

In this paper, we propose OctoKV, an innovative network-based key-value storage system. OctoKV addresses the repetitive address translation overhead associated with traditional key-value stores running on file systems on the client side. To mitigate this overhead, we implemented the key-value store on the server side using NVMe-oF and a user-level NVMe driver. In particular, we employed fine-grained resource monitoring and load balancing based on heuristics to optimize I/O performance. OctoKV is deployed on a Linux cluster with Intel SPDK. The extensive evaluation shows that OctoKV achieves lower I/O response times in comparison to traditional approaches where key-value stores run on the client side. Also, the proposed load balancing strategies efficiently enhance I/O response times by equally distributing the workload from overloaded cores to other cores.

Khan, Awais↗
DOE OSTI2022

DenKv: Addressing Design Trade-offs of Key-value Stores for Scientific Applications

High-performance computing (HPC) facilities have employed flash-based storage tier near to compute nodes to absorb high I/O demand by HPC applications during periodic system-level checkpoints. To accelerate these checkpoints, proxy-based distributed key-value stores (PD-KVS) gained particular attention for their flexibility to support multiple backends and different network configurations. PD-KVS rely internally on monolithic KVS, such as LevelDB or RocksDB, to exploit the KV interface and query support. However, PD-KVS are unaware of the high redundancy factor in checkpoint data, which can be up to GBs to TBs, and therefore, tend to generate high write and space amplification on these storage layers. In this paper, we propose DenKv which is deduplication-extended node-local LSM-tree-based KVS. DenKv employs asynchronous partially inline dedup (APID) and aims to maintain the performance characteristics of LSM-tree-based KVS while reducing the write and space amplification problems. We implemented DenKv atop BlobDB and showed that our proposed solution maintains performance while reducing write amplification up to 2× and space amplification by 4× on average.

Khan, Awais↗
DOE OSTI2022

Hvac: Removing I/O Bottleneck for Large-Scale Deep Learning Applications

Scientific communities are increasingly adopting deep learning (DL) models in their applications to accelerate scientific discovery processes. However, with rapid growth in the computing capabilities of HPC supercomputers, large-scale DL applications have to spend a significant portion of training time performing I/O to a parallel storage system. Previous research works have investigated optimization techniques such as prefetching and caching. Unfortunately, there exist non-trivial challenges to adopting the existing solutions on HPC supercomputers for large-scale DL training applications, which include non-performance and/or failures at extreme scale, lack of portability and generality in design, complex deployment methodology, and being limited to a specific application or dataset. To address these challenges, we propose High-Velocity AI Cache (HVAC), a distributed read-cache layer that targets and fully exploits the node-local storage or near node-local storage technology. HVAC seamlessly accelerates read I/O by aggregating node-local or near node-local storage, avoiding metadata lookups and file locking while preserving portability in the application code. We deploy and evaluate HVAC on 1,024 nodes (with over 6000 NVIDIA V100 GPUS) of the Summit supercomputer. In particular, we evaluate the scalability, efficiency, accuracy, and load distribution of HVAC compared to GPFS and XFS-on-NVMe. With four different DL applications, we observe an average 25 % performance improvement atop GPFS and 9% drop against XFS-on-NVMe, which scale linearly and are considered the performance upper bound. We envision HVAC as an important caching library for upcoming HPC supercomputers such as Frontier.

Khan, Awais↗
DOE OSTI2022

DENOVA: Deduplication Extended NOVA File System

This paper shows mathematically and experimentally that inline deduplication is not suitable for file systems on ultra-low latency Intel Optane DC PM devices in terms of performance, and proposes DeNova, an offline deduplication specially designed for log-structured NVM file systems such as NOVA. DeNova offers high-performance and low-latency I/O processing and executes deduplication in the background without interfering with foreground I/Os. DeNova employs DRAM-free persistent deduplication metadata, favoring CPU cache line, and ensures failure consistency on any system failure. We implement DeNova in the NOVA file system. Evaluation with DeNova confirms a negligible performance drop of baseline NOVA of less than 1%, while gaining high storage space savings. Extensive experiments show DeNova is failure consistent in all failure scenario cases.

Khan, Awais↗
DOE OSTI2021

MOSIQS: Persistent Memory Object Storage With Metadata Indexing and Querying for Scientific Computing

Scientific applications often require high-bandwidth shared storage to perform joint simulations and collaborative data analytics. Shared memory pools provide a chance to satisfy such needs. Recently, a high-speed network such as Gen-Z utilizing persistent memory (PM) offers an opportunity to create a shared memory pool connected to compute nodes. However, there are several challenges to use scientific applications on the shared memory pool directly such as scalability, failure-atomicity, and lack of scientific metadata-based search and query. In this paper, we propose MOSIQS, a persistent memory object storage framework with metadata indexing and querying for scientific computing. We design MOSIQS based on the key idea that memory objects on PM pool can live beyond the application lifetime and can become the sharing currency for applications and scientists. MOSIQS provides an aggregate memory pool atop an array of persistent memory devices to store and access memory objects to accelerate scientific computing. MOSIQS uses a lightweight persistent memory key-value store to manage the metadata of memory objects, which enables memory object sharing. To facilitate metadata search and query over millions of memory objects resident on memory pool, we introduce Group Split and Merge (GSM), a novel persistent index data structure designed primarily for scientific datasets. GSM splits and merges dynamically to minimize the query search space and maintains low query processing time while overcoming the index storage overhead. MOSIQS is implemented on top of PMDK. We evaluate the proposed approach on many-core server with an array of real PM devices. Experimental results show that MOSIQS gains a 100% write performance improvement and executes multi-attribute queries efficiently with 2.7× less index storage overhead offering significant potential to speed up scientific computing applications.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

An Analysis of System Balance and Architectural Trends Based on Top500 Supercomputers

Supercomputer design is a complex, multi-dimensional optimization process, wherein several subsystems need to be reconciled to meet a desired figure of merit performance for a portfolio of applications and a budget constraint. However, overall, the HPC community has been gravitating towards ever more Flops, at the expense of many other subsystems. To draw attention to overall system balance, in this paper, we analyze balance ratios and architectural trends in the world’s most powerful supercomputers. Specifically, we have collected the performance characteristics of systems between 1993 and 2019 based on the Top500 lists and then analyzed their architectures from diverse system design perspectives. Notably, our analysis studies the performance balance of the machines, across a variety of subsystems such as compute, memory, I/O, interconnect, intra-node connectivity and power. Our analysis reveals that balance ratios of the various subsystems need to be considered carefully alongside the application workload portfolio to provision the subsystem capacity and bandwidth specifications, which can help achieve optimal performance.

Khan, Awais↗
DOE OSTI2020

Persistent Memory Object Storage and Indexing for Scientific Computing

This paper presents Mosiqs, a persistent memory object storage framework with metadata indexing and querying for scientific computing. We design Mosiqs based on the key idea that memory objects on shared PM pool can live beyond the application lifetime and can become the sharing currency for applications and scientists. Mosiqs provides an aggregate memory pool atop an array of persistent memory devices to store and access memory objects. Mosiqs uses a lightweight persistent memory key-value store to manage the metadata of memory objects such as persistent pointer mappings, which enables memory object sharing for effective scientific collaborations. Mosiqs is implemented atop PMDK. We evaluate the proposed approach on many-core server with an array of real PM devices. The preliminary evaluation confirms a 100% improvement for write and 30% in read performance against a PM-aware file system approach.

Khan, Awais↗
DOE OSTI2020

ScaleML: Machine Learning based Heap Memory Object Scaling Prediction

Memory subsystem contributes 28-40% of total energy consumption. Several studies investigated energy prediction and consumption via profiling memory object access patterns. However, such profiling leads to higher energy consumption due to intense memory object-level profiling to achieve high prediction accuracy. Further, memory object access pattern prediction has been considered through analyzing the variation between memory object access patterns, referred to as scaling rate. The existing techniques for scaling rate prediction, such as Linear Scaling Rate (LSR), suffer from a high error rate in prediction with changes in access patterns, which leads to a high error rate of energy consumption prediction. In this paper, we compare and evaluate several memory object access pattern prediction models including LSR and machine learning (ML) models. Further, we propose SCALEML, a heap memory object scaling rate prediction mechanism that employs an ML model to achieve high access pattern prediction accuracy with variations in memory object access patterns. We evaluate SCALEML using various application benchmarks. The experimental results show that SCALEML achieves about 20% higher accuracy than the LSR model for predicting the scaling rate of object access patterns and energy estimation.

Park, Joongeon↗
DOE OSTI2020

An Analysis of System Balance and Architectural Trends Based on Top500 Supercomputers

Supercomputer design is a complex, multi-dimensional optimization process, wherein several subsystems need to be reconciled to meet a desired figure of merit performance for a portfolio of applications and a budget constraint. However, overall, the HPC community has been gravitating towards ever more FLOPS, at the expense of many other subsystems. To draw attention to overall system balance, in this paper, we analyze balance ratios and architectural trends in the world's most powerful supercomputers. Specifically, we have collected performance characteristics of systems between1993 and 2018 based on the Top500 lists, and then analyzed their architectures from diverse system design perspectives. Notably, our analysis studies the performance balance of the machines, across a variety of subsystems such as compute, memory, I/O, interconnect, intra-node connectivity and power. Our analysis reveals that balance ratios of the various subsystems need to be considered carefully alongside the application workload portfolio to provision the subsystem capacity and bandwidth specifications, which can help achieve optimal performance.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

An Integrated Indexing and Search Service for Distributed File Systems

Data services such as search, discovery, and management in scalable distributed environments have traditionally been decoupled from the underlying file systems, and are often deployed using external databases and indexing services. However, modern data production rates, looming data movement costs, and the lack of metadata, entail revisiting the decoupled file system-data services design philosophy. In this article, we present TagIt, a scalable data management service framework aimed at scientific datasets, which can be integrated into prevalent distributed file system architectures. A key feature of TagIt is a scalable, distributed metadata indexing framework, which facilitates a flexible tagging capability to support data discovery. Furthermore, the tags can also be associated with an active operator, for pre-processing, filtering, or automatic metadata extraction, which we seamlessly offload to file servers in a load-aware fashion. We have integrated TagIt into two popular distributed file systems, i.e., GlusterFS and CephFS. Our evaluation demonstrates that TagIt can expedite data search operation by up to 10× over the extant decoupled approach.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

NUMA-Aware Thread Scheduling for Big Data Transfers over Terabits Network Infrastructure

The evergrowing trend of big data has led scientists to share and transfer the simulation and analytical data across the geodistributed research and computing facilities. However, the existing data transfer frameworks used for data sharing lack the capability to adopt the attributes of the underlying parallel file systems (PFS). LADS (Layout-Aware Data Scheduling) is an end-to-end data transfer tool optimized for terabit network using a layout-aware data scheduling via PFS. However, it does not consider the NUMA (Nonuniform Memory Access) architecture. In this paper, we propose a NUMA-aware thread and resource scheduling for optimized data transfer in terabit network. First, we propose distributed RMA buffers to reduce memory controller contention in CPU sockets and then schedule the threads based on CPU socket and NUMA nodes inside CPU socket to reduce memory access latency. We design and implement the proposed resource and thread scheduling in the existing LADS framework. Experimental results showed from 21.7% to 44% improvement with memory-level optimizations in the LADS framework as compared to the baseline without any optimization.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

Influence of Plasticizers on Mechanical and Thermal Properties of Methyl Cellulose-Based Edible Films

Methyl cellulose (MC) films plasticized with malic acid (MA) and sorbitol (SOR) were prepared by casting solution and the influence of plasticizers at different concentrations (10, 20 and 30/100 g dry MC) on the mechanical and thermal properties were investigated. A decrease in tensile strength of the plasticized films was observed with increase in quantity of plasticizer. On the contrary, percentage elongation at break increased with an increase in plasticizer concentration that was more pronounced in SOR plasticized films. Films prepared with SOR exhibited higher water affinity and resulted in greater moisture absorption. Thermal stability was reduced with the addition of plasticizers and the SOR film samples were more thermally resistant as compared to MA-added films. FTIR analysis demonstrated the hydrogen bonding interactions between the functional groups of MC and plasticizers. The SEM micrographs of all the films showed a smooth and homogeneous surface morphology. The X-ray diffraction analysis suggested a semi-crystalline nature of the plasticized films.

36 MATERIALS SCIENCE↗