Holistic Integration of Emerging Supercomputing Technologies (Hi-EST) is a 5-year project funded by an ERC Starting Grant (GA 639595) awarded to Dr. David Carrera.
Hi-EST aims to address a new class of the placement problem, a challenge for computational sciences that consists in mapping workloads on top of hardware resources with the goal to maximize the performance of workloads and the utilization of resources. The objective of the placement problem is to perform a more efficient management of the computing infrastructure by continuously adjusting the number and type of resources allocated to each workload.
Description
Key Duties
• Co-advise PhD students and follow their work on a daily basis
• Define the requirements of the performance models used across workloads, paying special attention to HPC applications, Data-driven jobs, interactive applications in the supercomputing domain, Data Analytics tasks and IoT stream processing pipelines.
• Define the architecture, properties, and requirements of the key/value store layer to be used to leverage different storage technologies.
• Develop the performance models used for each workload considered in Hi-EST, and study how to make them extensive to other types of workloads that may appear in the future
• Design the key/value store layer to be used to leverage different key/value storage technologies across the storage hierarchy
• Study approaches to develop new algorithms for making task placement decisions. The algorithms will be guided by the adaptive online learning algorithms developed by other team members, which will behave as oracles that provide guidance in the process to build task placement maps.
• Study approaches to develop new algorithms for making data placement decisions. The algorithms will be guided by the adaptive online learning algorithms developed by other team members, which will behave as oracles that provide guidance in the process to build data placement maps.
• Study the opportunities to perform data-prefetching and data-driven workflow orchestration on selected workloads
• Performance testing
Requirements
• Education
o Holding or pursuing a PhD in Computer Sciences, Telecommunications or related discipline
• Knowledge
o Strong knowledge about performance modeling techniques
o Desirable knowledge about learning techniques for better integration with other efforts in the project
o Knowledge about FPGAs and their integration with GPU accelerators and NVMe devices.
o Excellent computing skills in high level computer languages (especially Java and C++ and Web Technologies) and experience with UNIX/LINUX environments and scripting languages
o Strong Knowledge of Network Protocols and High Performance Networks (Infiniband, RDMA)
o Strong Knowledge of CUDA programming model or similar
o Desirable knowledge on NVMe or similar interfaces
o Knowledge of NoSQL databases Fluency in English essential: regular written and spoken communication in English will be required
o Excellent written and verbal communication skills
• Professional Experience
o Capacity to interact and build strong relations with a diverse members/stakeholder/staff base.
o Willing to travel
• Competences
o Ability to take initiative, prioritize and work under set deadlines and pressure
o Ability to work independently and in a team
Conditions
The position will be located at BSC within the Computer Sciences department in collaboration with the specific program coordinator. The contract will be for three years.
Nr of positions available : 1
Research Fields
Computer science - Autonomic computing
Career Stage
Experienced researcher or 4-10 yrs (Post-Doc)
Research Profiles
Recognised Researcher (R2)