SuperCode: Sustainability PER AI-driven CO-DEsign

P. Chris Broekema,Rob V. van Nieuwpoort

2024-12-11

Abstract:Currently, data-intensive scientific applications require vast amounts of compute resources to deliver world-leading science. The climate emergency has made it clear that unlimited use of resources (e.g., energy) for scientific discovery is no longer acceptable. Future computing hardware promises to be much more energy efficient, but without better optimized software this cannot reach its full potential. In this vision paper, we propose a generic AI-driven co-design methodology, using specialized Large Language Models (like ChatGPT), to effectively generate efficient code for emerging computing hardware. We describe how we will validate our methodology with two radio astronomy applications, with sustainability as the key performance indicator. This paper is a modified version of our accepted SuperCode project proposal. We present it here in this form to introduce the vision behind this project and to disseminate the work in the spirit of Open Science and transparency. An additional aim is to collect feedback, invite potential collaboration partners and use-cases to join the project.

Instrumentation and Methods for Astrophysics,Artificial Intelligence

What problem does this paper attempt to address?

### What problems does this paper attempt to solve?

This paper aims to solve the problem of huge computational resource requirements and significant environmental impacts in data - intensive scientific applications (such as radio astronomy). Specifically, the paper proposes a project named **SuperCode**, whose goal is to optimize hardware and software through an AI - driven co - design method, thereby reducing the environmental footprint of these applications and improving their energy efficiency.

#### Main problem statements:

1. **Huge energy consumption in data - intensive science**:
- Data - intensive sciences (such as radio astronomy, high - energy physics, etc.) require a large amount of computational resources to process massive amounts of data, which leads to very high energy consumption.
- The current climate crisis makes it no longer possible for us to ignore the environmental impacts in scientific research. We must find a method to maximize scientific output while minimizing environmental impacts.

2. **Limitations of traditional co - design**:
- The traditional hardware and software co - design process is complex and time - consuming, and it is difficult to quickly adapt to the changes of emerging technologies.
- Traditional co - design usually only considers hardware and software, ignoring the opportunities that emerging technologies can promote new scientific discoveries.

3. **Application challenges of emerging technologies**:
- Although emerging hardware architectures (such as GPUs, neuromorphic chips, etc.) have higher energy efficiency, migrating existing code to these new architectures is very time - consuming and expensive.
- It is necessary to develop an efficient method to evaluate and utilize these emerging technologies to achieve more sustainable scientific computing.

#### Solutions:

- **AI - driven co - design**: Utilize generative AI technologies such as large language models (LLMs) to accelerate the hardware and software co - design process. Through AI, automatically generate efficient code suitable for emerging hardware platforms, reducing the time and cost of manual optimization.

- **Sustainability as a key performance indicator (KPI)**: Different from the traditional optimization of computing performance or efficiency, the SuperCode project takes "sustainability" as the main performance indicator. This includes environmental impacts in multiple aspects such as energy consumption, carbon emissions, and water resource usage.

- **Verification and application**: Verify the effectiveness of this method through two specific radio astronomy application cases (ground - based and space - based radio telescopes), and explore its potential applications in other fields (such as climate change research, remote sensing, earth observation, etc.).

#### Formula representation:

To quantify sustainability, the paper proposes an optimization formula for relative scientific value:

\[
MS=\frac{TVO}{TCO}
\]

where:
- \(MS\) represents the maximum scientific output per unit of environmental impact.
- \(TVO\) represents the Total Value of Ownership, that is, the value of scientific output.
- \(TCO\) represents the Total Cost of Ownership, that is, the cost of consumed resources.

In this way, the SuperCode project aims to create a process that can significantly reduce the time and effort required to evaluate and test hardware and software combinations, thereby achieving more sustainable scientific computing.