User Tools

Site Tools



September 25, 2021

SAFARI Live Seminar (27.09.2021, 5:30pm): Efficient Synchronization Support for Near-Data-Processing Architectures

Dear CompArch Students,

We would like to invite you to the next SAFARI Live Seminar, which will take place on Monday. Please find the details below.

We would strongly recommend you to attend, since this Seminar is very much related to many of the contents of the course.




Efficient Synchronization Support for Near-Data-Processing Architectures

Christina Giannoula, National Technical University of Athens

Livestream at 5:30 pm Zurich time (CEST) on YouTube:

Abstract: Recent advances in 3D-stacked memories have renewed interest in Near-Data Processing (NDP). NDP architectures perform computation close to where the application data resides, and constitute a promising way to alleviate data movement costs. These architectures can provide significant performance and energy benefits to parallel applications. Typical NDP architectures support several NDP units, each including multiple simple cores placed close to memory. To fully leverage the benefits of NDP and achieve high performance for parallel workloads, efficient synchronization among the NDP cores of a system is necessary. However, supporting synchronization in many NDP systems is challenging due to three architectural characteristics: (i) most NDP architectures lack shared caches that can enable low-cost communication and synchronization among NDP cores of the system, (ii) hardware cache coherence protocols are typically not supported in NDP systems due to high area and traffic overheads, (iii) NDP systems are non-uniform, distributed architectures, in which inter-unit communication is more expensive (both in performance and energy) than intra-unit communication.

In this seminar, we comprehensively examine the synchronization problem in NDP systems, and propose SynCron, an end-to-end synchronization solution for NDP systems. SynCron is designed to achieve the goals of performance, cost, programming ease, and generality to cover a wide range of synchronization primitives through four key techniques. First, SynCron adds low-cost hardware support near memory for synchronization acceleration. Second, SynCron includes a specialized cache memory structure to avoid memory accesses for synchronization and minimize latency overheads. Third, it implements a hierarchical message-passing communication protocol to minimize expensive communication across NDP units of the system. Fourth, SynCron integrates a hardware-only overflow management scheme to avoid performance degradation when hardware resources for synchronization tracking are exceeded.

Our work is the first one to analyze synchronization primitives in NDP systems using a variety of parallel workloads, covering various contention scenarios, and evaluating various NDP configurations. We demonstrate that SynCron achieves significant performance and energy improvements both under high-contention and low-contention scenarios, while it also has low hardware area and power overheads. We conclude that SynCron is an efficient synchronization mechanism for NDP systems, and hope that this work encourages further research on the synchronization problem in heterogeneous systems, including NDP systems.

Bio: Christina Giannoula is a Ph.D. student in the School of Electrical and Computer Engineering at the National Technical University of Athens (NTUA). She is working in the Computing Systems Laboratory, and is an affiliated Ph.D. researcher in the SAFARI research group at ETH Zürich, which is led by Prof. Onur Mutlu. She received a 5-year Diploma degree (Masters equivalent) in Electrical and Computer Engineering from NTUA in 2016, being awarded with several distinctions including the ‘Paris Kanellakis’ NTUA award, and graduating in the top 2% of her class. Since 2017, she has been working toward a Ph.D. degree at NTUA, and in 2019 she was a visiting PhD researcher in the SAFARI research group at ETH Zürich advised by Prof. Onur Mutlu and mentored by Prof. Nandita Vijaykumar. Her research interests lie in the intersection of computer architecture and high-performance computing. Specifically, her research focuses on the hardware/software co-design of emerging applications, including graph processing, pointer-chasing data structures, machine learning workloads, and sparse linear algebra, with modern computing paradigms, such as large-scale multicore systems and near-data processing architectures. She has several publications and awards for her research on these topics.

Paper: Christina Giannoula, Nandita Vijaykumar, Nikela Papadopoulou, Vasileios Karakostas, Ivan Fernandez, Juan Gómez-Luna, Lois Orosa, Nectarios Koziris, Georgios Goumas, and Onur Mutlu, “SynCron: Efficient Synchronization Support for Near-Data-Processing Architectures”, Proceedings of the 27th International Symposium on High-Performance Computer Architecture (HPCA), Virtual, February-March 2021.

[Slides ( (]

[Short Talk Slides ( (]

[Talk Video (21 minutes)]

[Short Talk Video (7 minutes)]

September 21, 2021

Welcome to Computer Architecture Fall Semester (HS) 2021

Dear Computer Architecture Students,

First of all, I would like to welcome you to our Computer Architecture course (Fall 2021).

We have an exciting semester ahead of us, where we will cover the most recent trends and exciting ideas in Computer Architecture, including in critically important topics like memory systems, energy, security and reliability, emerging memory technologies, novel execution paradigms (like processing in memory), heterogeneous architectures, interconnects, multiprocessors, and so on. Many ideas we will cover are fundamental and they have had significant impact on the computing systems that are designed today (as we will discuss with examples in lectures and readings in the course). I very much hope that this course opens your eyes and minds to new ways of thinking about computing platforms and hopefully enables you to become the architects of much more efficient and powerful and overall fundamentally-better computing systems of the future.

Even though this course will be an intense learning experience, feedback from past students consistently mention the value of this course in their future careers, and I hope you will also benefit from the course greatly. I would suggest taking all aspects of this course as an experience in learning and self-growth - this positive learning- and growth-oriented mindset will enable you to make the most out of the course. We will for example release many extra credit assignments. Such assignments are not only excellent opportunities for learning, but also excellent opportunities for getting a strong grade.

Second, I would like to let you know that the course website is ready:

Please get familiar with its different sections as soon as possible. The course website will be the main synchronization point of the course. The course Moodle is also ready:

Third, I would like to announce your first assignments of the semester for this week:

1. Read and review the inspiring “You and your research” by Richard Hamming. You can access the transcript of this talk from our course website, and in the following link:

We will ask you to enter your review online for this transcript as part of paper reviews, which will be a regular occurrence in this course, so please construct your review. Please follow the guidelines for reviewing papers and talks that we also include on the course website:

2. Watch and review a recent overview talk I delivered on “Intelligent Architectures for Intelligent Machines”:

This talk provides an overview of some of the topics that we will cover in this course. This talk is a great approach to getting familiar with many of the contents of this course. You can access the slides of the talk from my website:

We will ask you to enter your review online for this talk as part of paper reviews, which will be a regular occurrence in this course, so please construct your review. Please follow the guidelines for reviewing papers and talks that we also include on the course website (link at the end of point 1.).

3. (Optional) Watch last year's Lecture 1 in preparation for our first week of lectures next week. You can find last year's Lecture 1 as well as all other lectures here:

Watching this lecture will help prepare you even more for the course. Also, our first lecture next week will be a revised version of last year's Lecture 1. You can interact better and potentially ask more and deeper questions if you prepare by watching last year's Lecture 1.

Fourth, the lectures will be livestreamed online this semester on both Zoom and Youtube. We will start the first lecture of the course next Thursday, 30 September 2021. We will provide you with a Zoom link soon and Youtube links will be available on the course website as well as on our Youtube channel: The lectures will also be available on Youtube after they are delivered.

You will also have the option to attend and watch the lectures from our lecture room (HG D 16.2). Our course staff will set up the room for the lecture livestream. You will be able to interact with me and the TAs, and ask questions and receive answers, regardless of where you attend the lecture from (either the lecture room or anywhere else via Zoom or Youtube).

If you come to the lecture room, recall that you need to have a valid Covid-19 certificate ( We will do the necessary certificate checkings. It is important that everyone, including the most vulnerable among us, are kept safe and none of us act as agents unnecessarily spreading a dangerous virus that has already wreaked too much havoc. My strong recommendation to you is to avoid attending the lectures in person if you are not fully vaccinated or if you feel sick for any reason.

Fifth, your HW0 is already released:

Please submit this sheet with your information via the following Moodle assignment (due 30.09):

Sixth, in the following days, we will release “Lab 1: Caching”. You can start working on it as soon as it is released. You will get a refresher on cache behavior by creating your own timing simulator. Note that we will have 5 labs in this course as in past years, and they will all help you learn practical aspects of Computer Architecture. Labs will operate asynchronously with lectures, and you will have time to complete them even after lectures end. However, I would strongly suggest that you do the labs during their suggested timeframes (i.e., deadlines) so that they do not pile up on you.

Seventh, our course staff will have office hours available to you starting next Monday, 27.09. We will announce the exact times and Zoom links soon.

I wish you an excellent start to the semester.

Do not hesitate to email me and our TAs (, if you have any questions.

Best regards,

Professor Onur Mutlu

announcements.txt · Last modified: 2021/10/04 17:26 by juang