Chetan Choppali Sudarshan,Nikhil Matkar,Sarma Vrudhula,Sachin S. Sapatnekar,Vidya A. Chhabria
Abstract:Decades of progress in energy-efficient and low-power design have successfully reduced the operational carbon footprint in the semiconductor industry. However, this has led to an increase in embodied emissions, encompassing carbon emissions arising from design, manufacturing, packaging, and other infrastructural activities. While existing research has developed tools to analyze embodied carbon at the computer architecture level for traditional monolithic systems, these tools do not apply to near-mainstream heterogeneous integration (HI) technologies. HI systems offer significant potential for sustainable computing by minimizing carbon emissions through two key strategies: ``reducing" computation by reusing pre-designed chiplet IP blocks and adopting hierarchical approaches to system design. The reuse of chiplets across multiple designs, even spanning multiple generations of integrated circuits (ICs), can substantially reduce embodied carbon emissions throughout the operational lifespan. This paper introduces a carbon analysis tool specifically designed to assess the potential of HI systems in facilitating greener VLSI system design and manufacturing approaches. The tool takes into account scaling, chiplet and packaging yields, design complexity, and even carbon overheads associated with advanced packaging techniques employed in heterogeneous systems. Experimental results demonstrate that HI can achieve a reduction of embodied carbon emissions up to 70\% compared to traditional large monolithic systems. These findings suggest that HI can pave the way for sustainable computing practices, contributing to a more environmentally conscious semiconductor industry.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to evaluate and reduce the carbon footprint (CFP) of heterogeneous integration (HI) systems based on chiplets throughout their entire life cycle in order to achieve sustainable VLSI (very - large - scale integration) design. Specifically, the paper aims to:
1. **Propose a new CFP estimation tool**: Develop a tool named ECO - CHIP for estimating the carbon footprint of systems based on the chiplet architecture, including carbon emissions in the design, manufacturing, packaging, and operation stages.
2. **Evaluate the sustainability potential of HI systems**: Prove through experimental results that, compared with traditional monolithic systems, HI systems can significantly reduce the carbon footprint by "reducing" computational complexity and "reusing" pre - designed chiplet IP modules.
3. **Optimize multi - dimensional design decisions**: Provide a method to optimize the power, performance, area, cost, and carbon emissions of HI systems, considering factors such as different packaging architectures, process node selections, and chip utilization.
### Specific problem description of the paper
#### 1. Necessity of carbon footprint assessment
With the development of the semiconductor industry, although energy efficiency and low - power - consumption design have successfully reduced carbon emissions during operation, the embodied carbon emissions generated during the design, manufacturing, and packaging processes are increasing. Existing CFP analysis tools are mainly applicable to traditional monolithic systems and are not suitable for emerging heterogeneous integration technologies. Therefore, new tools and models are required to evaluate the environmental impacts of these new technologies.
#### 2. Advantages of HI systems
HI systems can significantly increase yield and reduce waste by decomposing large SoCs into multiple smaller chiplets. In addition, the mixed use of different technology nodes and the reuse of chiplets can further reduce carbon emissions. For example:
- **Higher yield**: A smaller die size increases the yield, thereby reducing carbon emissions per unit area.
- **Flexibility in technology node selection**: Older technology nodes with lower defect densities and carbon emissions can be selected for part of the design.
- **Reuse of chiplets**: Pre - designed chiplets can be reused in multiple designs, reducing the cost and carbon emissions of redesign and remanufacturing.
#### 3. Functions of the ECO - CHIP tool
The ECO - CHIP tool is able to:
- Estimate the total carbon footprint of the design, manufacturing, packaging, and operation stages of HI systems.
- Consider factors such as different packaging architectures, scaling ratios, yields, device energy efficiencies, and silicon wafer waste.
- Provide detailed CFP analysis to help designers make optimal choices among power, performance, area, cost, and carbon emissions.
### Conclusion
By introducing the ECO - CHIP tool, this paper demonstrates the great potential of HI systems in reducing the carbon footprint, especially in the current situation where Moore's law is slowing down and the expansion of technology nodes is limited. HI systems provide a viable path for sustainable computing. Experimental results show that HI systems can reduce embodied carbon emissions by as much as 70%, which provides an important reference for future green VLSI design.
### Formula summary
The formulas involved in the paper are as follows:
1. **Total carbon footprint estimation formula**:
\[
C_{\text{tot}} = C_{\text{emb}} + \text{lifetime} \times C_{\text{op}}
\]
where \( C_{\text{emb}} \) is the embodied carbon footprint and \( C_{\text{op}} \) is the operational carbon footprint.
2. **Embodied carbon footprint estimation formula**:
\[
C_{\text{emb}} = C_{\text{mfg}} + C_{\text{des}} + C_{\text{HI}}
\]
where \( C_{\text{mfg}} \) is the manufacturing carbon footprint, \( C_{\text{des}} \) is the design carbon footprint, and \( C_{\text{HI}} \) is the additional carbon footprint of the HI system.
3. **Operational carbon footprint estimation formula**:
\[
C_{\text{op}} = C_{\text{src, use}} \times