Manos Pavlidakis

Manos Pavlidakis

R&D Software Engineer

FORTH, ICS, CARV

Biography

Hi, my name is Manos Pavlidakis and I am a software engineer with a Ph.D. in Computer Science and Engineering from the University of Crete, Greece. My doctoral research on the “Transparent Spatial Sharing of Multiple and Heterogeneous Accelerators” provided extensive hands-on experience in various accelerator programing models and runtimes, including OpenCL, CUDA, and ROCm. Additionally, I am well-versed in popular machine learning frameworks such as PyTorch, Caffe, and TensorFlow, showcasing a comprehensive skill set at the forefront of technology.

Interests
  • Machine Learning
  • Hardware Accelerators
  • Accelerator Runtimes
Education
  • PhD in Computer Science (Scalable Architectures and Parallel Programming), 2024

    University of Crete

  • MSc in High Performance Distributed Computing, 2016

    Vrije Universiteit (VU) Amsterdam

  • BSc in Information and Communication Systems Engineering, 2012

    University of Aegean, Samos

Skills

Technical
C/C++
CUDA
ROCm
OpenCL
Hobbies
Padel

Experience

 
 
 
 
 
CARV, ICS-FORTH
Graduate Research Assistant
September 2017 – Present Heraklion, Crete
Design and develop Arax, a runtime that manages and shares transparently multiple and heterogeneous accelerators.
 
 
 
 
 
CSD, UoC
Teaching Assistant
September 2017 – Present Heraklion, Crete

Responsibilities include:

  • Marking assignments and exams
  • Give tutorials
  • Answering student’s questions
  • Examine students at digital design labs
 
 
 
 
 
CARV, ICS-FORTH
Research Engineer
January 2017 – September 2017 Heraklion, Crete
Design and develop Vinetalk, a framework that simplifies FPGA access while enabling FPGA sharing across different applications
 
 
 
 
 
CARV, ICS-FORTH
Intern
May 2016 – September 2016 Heraklion, Crete
Implementing an HDFS library over NoSQL DB
 
 
 
 
 
PavlidakisDev
Free launcher
PavlidakisDev
June 2013 – July 2014 Heraklion, Crete
Android application and website development

Accomplish­ments

NVIDIA certification
Fundamentals of Accelerated Computing with CUDA C/C++
See certificate
NVIDIA certification
Scaling CUDA C++ Applications to Multiple Nodes
See certificate

Talks

Programming Languages for Accelerators.
Introduction to CUDA and other programming models for accelerators.
Programming Languages for Accelerators.
Why accelerators?
Introduction to hardware accelerators and their usage.
Why accelerators?

Publications

Quickly discover relevant content by filtering publications.
(2024). G-Safe: Safe GPU Sharing in Multi-Tenant Environments. Arxiv.

PDF Cite

(2023). Flexy: Elastic Provisioning of Accelerators in MultiGPU Servers.

PDF Cite

(2022). Arax: a runtime framework for decoupling applications from heterogeneous accelerators. Proceedings of the 13th Symposium on Cloud Computing.

PDF Cite Poster Slides

(2020). AManifesto FOR FUTURE GENERATION HETEROGENEOUS COMPUTING: RESEARCH DIRECTIONS.

Cite