# Telling neuronal apples from oranges: analytical performance modeling of neural tissue simulations

## TL;DR

This paper presents an analytical performance modeling approach to identify bottlenecks in neural tissue simulations, focusing on how model abstractions and hardware limitations affect scalability and performance.

## Contribution

It introduces a systematic performance analysis method for brain models, highlighting the impact of synaptic formalism and hardware bottlenecks on simulation scalability.

## Key findings

- Synaptic formalism significantly affects memory bandwidth saturation.
- Memory bandwidth and network latency are key bottlenecks as models scale.
- Performance landscape characterization guides future hardware and software development.

## Abstract

Computational modeling and simulation have become essential tools in the quest to better understand the brain's makeup and to decipher the causal interrelations of its components. The breadth of biochemical and biophysical processes and structures in the brain has led to the development of a large variety of model abstractions and specialized tools, often times requiring high performance computing resource for their timely execution. What has been missing so far was an in-depth analysis of the complexity of the computational kernels, hindering a systematic approach to identifying bottlenecks of algorithms and hardware, and their combinations. If whole brain models are to be achieved on emerging computer generations, models and simulation engines will have to be carefully co-designed for the intrinsic hardware tradeoffs. For the first time, we present a systematic exploration based on analytic performance modeling. We base our analysis on three in silico models, chosen as representative examples of the most widely employed modeling abstractions. We identify that the synaptic formalism, i.e. current or conductance based representations, and not the level of morphological detail, is the most significant factor in determining the properties of memory bandwidth saturation and shared-memory scaling of in silico models. Even though general purpose computing has, until now, largely been able to deliver high performance, we find that for all types of abstractions, network latency and memory bandwidth will become severe bottlenecks as the number of neurons to be simulated grows. By adapting and extending a performance modeling approach, we deliver a first characterization of the performance landscape of brain tissue simulations, allowing us to pinpoint current bottlenecks in state-of-the-art in silico models, and make projections for future hardware and software requirements.