Researcher's Dream

Making the world a better place through cutting-edge computing technology


Published on March 28, 2023

Turning computer knowledge into my own secret weapon

While I was in elementary school, I started to learn about computing at a class run by a friend of my mother.
This involved playing lots of games, which introduced us to basic computer operations and taught us to type quickly.
I used to struggle with writing characters neatly, and felt empowered by being able to output characters just by typing on a computer keyboard without having to write them all by hand!
This was a turning point for me, recognizing that computers would have a major impact on the world in the future. I knew that I wanted to use this knowledge as my own weapon!
The more I learned, the more interested I became in computing, with the ambition of working in this field in the future.

Learning from playing shogi

During my childhood, I was hooked on shogi for a time.
While it was fun to learn the tactics, what really stands out in my memory is the experience of teaching my friends what I had learned and enjoying shogi together.
When I entered high school, I realized that I was good at physics and chemistry.
I especially liked being able to work hands-on with practical applications based on theory.
For example, I was able to apply the technique of friction propulsion to run fast.
It really fired me up to translate this into my daily life.
When the time came to decide on my future direction, I found it more of a struggle than I anticipated to balance two particular avenues. I found teaching others very fulfilling but I also loved computer science.
So it was very difficult to decide whether to major in computer science or become a science teacher.
In the end, I majored in information engineering, but I have never lost my excitement for teaching, as first experienced when playing shogi and sharing my knowledge and skills to others. It still remains with me and is very much connected to the present.

A life-changing experience: presenting at an international academic conference

While I was in graduate school, lots of research was being done on the use of Graphics Processing Units (GPU)(*1), well beyond just improving the rendering speed of 3D games.
I was also involved in research on making GPU numerical calculations more general-purpose.
At the time, I was so focused on fulfilling my graduation requirements that I didn't give much thought to the social impact of my research. But when my paper (*2) was accepted at an international conference, I felt a real sense of pride in being at the forefront of my field.
I presented my paper to an audience of experts from both inside Japan and internationally, and the interest everyone paid to my work really boosted my self-confidence.
This experience prompted me to re-consider the direction of my future research and the impact it could have.

Focusing on High-Performance Computing (HPC) development

Since joining Fujitsu, I have been working in the field of High-Performance Computing (HPC) (*3).
When I first started, my knowledge was limited to a specific field, and I realized that I had a lot of gaps in the required knowledge for accelerating the applications of supercomputers.
So it was a major challenge for me initially to focus on HPC.

One key moment for me was the supercomputer contest TOP500 in 2016.
I had to run the High-Performance Linpack (HPL) on a supercomputer comprising many computers, but it took 13-15 hours from the start to the end of the HPL.
During this time, if anything went wrong with any of the related parts, we had to start all over again.
10 hours into the operation, an error occurred and we had to restart.
We just kept dealing with it until it was finally completed.

An error can mean many things. For example, the hardware cooling may be insufficient, leading to some of the computers overheating and shutting down.
I would say the supercomputer is like a "living creature". When you are sick, it affects your overall performance, similar to the operation of a supercomputer.
Of course, after that, we identified the source of the error and took the appropriate measures.
The HPL operation was successfully completed, and we were able to show the supercomputer to the world.
We received some very good evaluations and achieved a number of successful business negotiations, thanks to the fact that our supercomputer had higher execution efficiency than our competitors’ equivalent supercomputers (*4).

Since then, I have had several other opportunities to operate supercomputers as part of Fujitsu’s business negotiations.
In fact, I have had many chances to showcase my abilities, spreading and sharing my computing technology knowledge to those around me as well as turning the gears within projects.
After overcoming lots of challenges as a team, it has been a wonderful feeling to achieve milestones such as the Japan’s or world's fastest supercomputer(*5).
Achieving first place gave me a genuine feeling of joy and it has directly motivated me in my work.

Bringing cutting-edge computing technology to the masses

In research and development, I always like to make sure that I am moving forward, even if I have to take unexpected detours or can't see what's ahead.
I believe that even if I do take a detour, I might still reach the summit before anyone else and encounter new discoveries.
In the past, I also began to study AI, anticipating that the deep learning field would require faster computing.
I tried making a simple framework to run AI, but at that time, the programming reference was not well established, therefore the development of the framework was very difficult.
As a results, it ended up being shelves.
Nevertheless, by continuing to take this approach, I have gradually understood what was needed and been able to achieve my goals, even when I couldn't clearly see the direction.

Now, I am working on the realization of Computing Workload Broker (*6) and developing quantum and HPC hybrid computing technology (*7).
Specifically, I aim to challenge practical problems in the field of quantum computing and to make it possible for anyone to use the most advanced computing without any specialist knowledge.
My dream is to deliver the knowledge of the most advanced computing technology to as many people as possible in some form.
Therefore, I feel that this project aligns with my purpose.

Combining computing technology with AI, quantum, network technology, or others field can lead to innovative technologies.
I believe that having the strength and conviction to step boldly into unknown areas is an important element in research. I just want to keep looking ahead, and do my absolute best.

Akihiko Kasagi
Computing Laboratory
Graduate School of Engineering
Joined Fujitsu in 2015
My Purpose
Making the world a better place through cutting-edge computing technology
On my days off, I go for walks with my children. I love to cook, and my specialty is Chikuzenni - a dish that originated from northern Kyushu in Japan, made of braised chicken and vegetables.

Editor's note

Editor: Xiang Yi Peck, Communication Strategy Division

The Fujitsu Supercomputers “K computer” and “Fugaku” are developed using cutting-edge technologies. Akihiko has been in charge of computing technology development since joining the company, and continues to explore new areas such as AI and quantum computing, beyond his own specialized computing fields. In the past, he had a dream of becoming a teacher. Today, he is a reliable leader, who promotes knowledge sharing with team members in a different way. He looks forward to continuing his mission of bringing cutting-edge technology to people and we wish him all the very best in his work.


  • facebook
  • twitter
  • LinkedIn
  • LINE
  • Weibo



Connect with Fujitsu Research

Top of Page