Closed seanmcilroy29 closed 6 months ago
Attended
attended
Attended
Attended
Attended!
Attended
Attended
Attended
attended
Meeting summary
AI project scope and article development. Thomas and Henry discussed the green AI project, including prioritising and scoping the project's goals and timeline. The SC provided feedback and insights on the project, including questions and concerns about tooling and content. Thomas: Cross-working group committee formed to develop new standards, with input from different membership levels and cultures. Thomas suggests bringing in other experts to support the project and ensure everyone is involved and engaged. Sean provided an update on the article for the standards WG, which will be publicized and shared with other WGs.
Benchmarking large language models for efficiency. Chris and Sean ‘O’ discussed collaborating on an AI initiative. They explored the existing solutions in the market and identified the gaps that need to be filled. Chris also talked about the potential use of SCI standards in the energy sector and how they could focus on Iran to implement them. Additionally, Chris proposed using a standardised benchmarking workload to measure the carbon footprint of large language models, which could be a valuable tool for organisations to track their impact on the environment.
Energy consumption and carbon footprint of AI models. Thomas and Gadhu talked about a new AI model that focuses on energy consumption by using large language models. They also discussed the challenges of training and deploying AI models on edge devices. Thomas pointed out the limitations of current language models, such as slow performance and the inability to run on edge devices due to GPU limitations. Chris mentioned the high upfront cost of building and training a language model and the ongoing cost of running it.
AI model development and benchmarking. Gadhu and Chris discussed the standardisation of measuring AI models on reference platforms. Gadhu shared a list of benchmark workloads and tasks for AI research during the meeting, which will be available to the public. Chris and Henry expressed their interest in using this information for their research, and Gadhu agreed to share the link with them.
Project scope and technical feasibility. Thomas highlights the challenge of scoping milestones for the project, focusing on technical feasibility and sustainability.
Developing a framework for green software in organisations. Pindy provided an overview of a project aimed at assisting organisations in using data to make better decisions and achieve net-zero targets. Pindy has also provided detailed information regarding project timelines and milestones for new audiences, which includes delivery managers, change managers, and project managers. In addition, Pindy and Henry have discussed the use of a maturity matrix for organizational change management and have focused on the target audience and messaging approach.
Standardising software efficiency measurement. Sean O proposed labelling software with energy efficiency ratings to help consumers make informed choices. WG aims to create a standard for AI initiatives, with a proposed next step of joining hands on the sheer standard for Iran.
Software carbon footprint measurement. Sean ‘O’ emphasizes the importance of understanding the unit of software function when measuring energy consumption, as different categories have different energy requirements. The group discusses the idea of creating a mock label to inform consumers about software efficiency and reward producers for creating efficient software within their category. Sean ‘O’ discusses the importance of measuring both relative intensity and absolute intensity in software development, and how they plan to prototype different ways to present this information to users. Sean ‘O’ shares their research on the carbon footprint of different programming languages, including Java, Python, and Rust, and how this could impact software development in the future.
Software carbon footprint and labelling. Chris highlights the benefits of customizable labels in software certification, driving better behaviour and efficiency. The discussion revolves around the use of AI in software development, focusing on the importance of understanding the context and requirements of the project. Chris discussed the relativity of time in AI language models and how it impacts their performance.
Software efficiency testing and benchmarking. Sean ‘O’ discusses defining workload and environment for rating tests to ensure constant factors. Navveen adds comments to the spec, highlighting standard components across categories.
2024.02.15 Agenda/Minutes
Time 1600 (GMT) - See the time in your timezone
Antitrust Policy
Joint Development Foundation meetings may involve participation by industry competitors, and the Joint Development Foundation intends to conduct all of its activities in accordance with applicable antitrust and competition laws. It is, therefore, extremely important that attendees adhere to meeting agendas and be aware of and not participate in any activities that are prohibited under applicable US state, federal or foreign antitrust and competition laws.
If you have questions about these matters, please contact your company counsel or counsel to the Joint Development Foundation, DLA Piper.
Recordings
WG agreed to record all Meetings. This meeting recording will be available until the next scheduled meeting.
Roll Call
Please add 'Attended' to this issue during the meeting to denote attendance.
Any untracked attendees will be added by the GSF team below:
Agenda
GreenAI - Discussion Continued - Henry / Thomas
WG Project - Dashboard
Project Review updates
[ ] Transforming Organisational Software Sustainability - Pindy
[ ] Software Carbon Efficiency Rating - Chris
Project Updates
ISO - Sean
Use cases
Article R&A peer review
Note: WG use case template submission - After submitting this issue, your use case will be submitted to the WG Agenda for discussion. Article submission - Once you submit this issue, it will be assigned to the GSF Editor to review.
Future meeting Agenda submissions
Next Meeting
Adjourn
[ ] Motion to adjourn
Meeting Action Items / Standing Agenda / Future Agenda submissions