First, I want to introduce a little bit about the Rican background since some of you may

not know Rican.

I know that there are a bunch of students in the audience.

So just to give you a little bit of an overview, Rican was founded quite long ago, over 100

years ago.

It is one of the biggest research institutes in Japan.

It currently has 10 different campuses.

We are right here, one of the campuses in Kobe, which hosts the supercomputer Fugaku.

But the main office is near Tokyo in Wako.

We have over 3,000 researchers and administrators.

So quite a big research institute, quite similar to one of the DOE labs in general, Max Planck

from Germany for example.

So there's a bunch of stuff.

So we don't do only computer science.

There's biology, chemistry, physics.

We have Spring 8 for example, which is a cyclotron accelerator to do stuff like DESI in Germany

for example and other things.

And at RCCS, we basically do computer science.

So the science of computing, by computing and for computing.

So we have various research teams.

My team is only one of that, which was founded quite recently.

So we have computer science teams, we have computational science teams, which take care

of disaster mitigation, fluid dynamics and so on.

We have drug discovery teams, AI teams, and some of the teams are mainly for supercomputer

Fugaku for management operation stuff.

So there are a substantial amount of teams and researchers inside of RCCS.

And you may know a few of the recent achievements.

So Gordon Bell Prize was awarded to Rican for the fight against COVID-19.

And then we have data assimilation stuff going on where we combine real-time inputs into

climate simulations and disaster mitigation simulations.

And Fugaku also won and is currently still number one for the HPCG and the Graph 500

list.

But in the beginning, it was number one on all four titles.

So top 500, Graph, HPCG and HPL AI.

So that changed recently with the start of Frontier.

So let's look at the outline.

So after going a little bit over our center, let's look at the outline for the large cache

study we have done recently.

The motivation why we want to look at this.

Then I will introduce our hypothetical log processor and then go a little bit over the

different evaluation strategies we have developed.

And we are trying to basically see what it would bring if we would increase the amount

of cache and maybe reduce the latency as well.

And then a little bit of an outlook.

So why is biologic cache interesting?

Basically at the moment, we are a little bit at the end of a more era.

And so there are different potential paths we can go.

We can go quantum, neuromorphic.

People are looking into reconfigurable computing like CGIs, FPGAs, different pathways, how

we can combat the lack of transistor shrinking in the future.