CyberInfrastructure Facilitator

General Summary/Purpose:

The Maryland Advanced Research Computing Center (MARCC) is a state of the art High Performance Computing (HPC) facility that provides resources (HPC, storage and analytics) for researchers at Johns Hopkins University and The University of Maryland at College Park. TheSoftware Engineer(CyberInfrastructure Facilitator) provides direct support to all users on the effective utilization of resources, code development, debugging, optimization, installation and maintenance of open source scientific applications. The incumbent will enable faculty to advance research-computing agendas by providing direct technical support and training modules on HPC topics to the research community. There will be also many opportunities to establish scientific collaborations and partnerships with research groups.

Position’s roles & interactions:

The CyberInfrastructure Facilitator will interact directly with the user community to support their research computing activities. There will be many opportunities to teach, train and educate users in traditional HPC topics but perhaps more importantly in new techniques and methodologies in HPC. These topics may include programming, parallel and gpu programming, data intensive applications, scripting and best practices in advanced computing. The CyberInfrastructure Facilitator will also interact with the systems group to assist them in troubleshooting and fixing potential problems. This position will be subject to mentoring and training but eventually a degree of independence is expected. The director supervises projects, reviews progress and may assign projects or tasks as necessary.

Describe the specific systems, applications, projects for which the position is responsible:

MARCC staff (user and application support group) interact directly with the user community to sustain and expand a productive research computing environment conducive to moving research forward. CyberInfrastructure Facilitators will train and educate users in HPC best practices, efficient ways to utilize and balance resources, timely answering a variety of support tickets, and installing and supporting scientific applications. These tasks are critical to the user community and MARCC staff as we aim to promote collaborations, partnerships and collegiality. Most applications have different sets of requirements, either other applications, libraries or particular compilers that make these tasks complex and require special skills. These are a few examples of the responsibilities for this position: Strengthen the effective utilization of resources, foster code development, and apply best practices for debugging, optimization, installation and maintenance of open source scientific applications.

In addition, CyberInfrastructure Facilitators work in collaboration with the systems group to install and manage middleware applications in test or in production environments.

Describe scale/size of area, project and/or system supported:

MARCC provides services to over 1000 active researchers (approximately 450 research groups) from these schools: KSAS, SOM, WSE, BSPH, and UMCP. Currently, we have two HPC clusters. The Bluecrab consists of over 23,000 cores and the new Rockfish cluster of over 18,000 cores and expected to double its size within one year. Most of these researchers utilize different applications and conduct different types of computing. We expect a multitude of large parallel jobs (256 cores per job), and high throughput computing jobs that use job arrays (thousands of serial jobs like parameter sweeps), serial and parallel jobs that require large memory, and several users with applications that use GPUs. Although this is a shared facility, the diverse set of applications and users is definitely a factor in the complexity of the system, along with the different applications that need to be synchronized for seamless functioning of the facility.

Required & preferred skills specific to position: