Which are the BIG questions that you expect to discuss at Multicore World 2023?

Every year we invite our attendees to write their expectations (through the registration form -completely optional) so we can discuss them collectively during the sessions, and speakers have an idea a priori of which are the expectations from the audience.

Please take a few minutes and share your vision on what should we be discussing in February 2023 at the 10th Multicore World. And if you cannot make it to New Zealand in February 2023, feel free to send us your BIG Questions to info@multicoreworld.com and we’ll aim to cover them through the sessions.

---

BIG Questions for Multicore World 2023

Updated 17 January 2023

• How do I lead people through the IT future landscape?
• Especially given that the human race is notoriously poor at figuring out what it wants, how do we produce software that does what we want?
• What is missing from recent architectures? What is important to Community : Memory latency / BW, Re-coding efforts, Reliability, Energy / Power consumption?
• What is the next big challenge for High Performance Computing?   [Perhaps to inject some debate/controversy – I don’t think the HPC community should target ZettaFLOPS]
• What does it all mean to the rest of the tech industry?
• What does the world need from us and how can we best meet that need?
• What are the challenges to “crossing the valley of death” for future microelectronics and computing technology?
• How can next-gen computing community and food and fibre systems community work together to enable the future of an equitable, robust and sustainable food and fibre system in NZ and internationally?
• How do we create a common interconnected environment for complex scientific workflows that span heterogeneous experimental instruments and computational resources?
• Is Moore’s Law actually dead beyond 2-3 nm, or is there much promise in mitigating the effects of quantum tunneling and in 3D stacking (e.g. AMD’s “3D V-Cache”)?
• How could hardware vulnerabilities affect the HPC landscape from both a security and a performance perspective? What are some useful metrics in determining the security: performance balance when deploying mitigations?
• The 1966 Burt Bacharach song “Alfie” contains the line “Without true love we just exist”.  My question to any human-level AI researchers is whether it is also true that “If we don’t exist, then we just love”, and, if so, how best to argue this point against a sentient AI. (Note of the Editor: this was sent before ChatGPT 😉

---

BIG Questions from Multicore World 2020

Updated 14 February 2020

• While we had the DOT-com boom, nobody seemed to connect those dots. Meanwhile we’re transitioning into a world where the technology allows us to connect all the dots. We all know how certain technologies changed the world forever. What do you think is the biggest breakthrough a fully connected world can
  provide to humanity? And why do you think this could be a bad thing?
• What’s the value and feasibility of extreme heterogeneity?
• How long will the AI hype last?, or is it really fundamental this time?
• TSMC is ready with 5nm, is Moore’s law dead now?
• When can I stop renewing my drivers license? 🙂
• Security
• State and directions of secure software, especially with regard to infrastructure and IoT
• What are some of the biggest impacts that have been felt from the onslaught of recently discovered hardware-level vulnerabilities?
• Bridging the chasm between [1] BIG Governments’ funding of ‘hero’ ‘arms race’ HPC installations and [2] the realities of relatively small government / small companies’ smart use of ‘cloud’ technologies
• What are the expectations for 5G in NZ, AU?
• How long before OpenCL is a serious challenge to CUDA?
• Value and feasibility of extreme heterogeneity
• I am looking forward to Bruce Jacob’s talk
• How AI can help HPC, not just HPC for AI
• Should our HPC site integrate with external cloud providers, or remain independent?
• Hardware (in)security
• Mistakes at scale: business and legal implications of “decisions” made by AI models trained on exotic architectures from massive data
• Resilience
• Is the OS stack dealing accordingly with the massive potential offered by the new architectures?
• How can we now actually leverage the huge computational power already available?
• Would be feasible to “re-invent” the whole “Information Technology (IT)” industry solving everything by brute force?
• Is Von Neumann architecture an outdated model?

Learn about colleagues who will were at Multicore World 2020:

• Main professional activity / position
• Main fields of interest
• Any specific expectation from Multicore World 2020?
• Where did your colleagues come from?

---

BIG Questions from Multicore World 2019

Updated 10 February 2019

• How to adjust computer architectures in a world where traditional technology is not scaling in a sufficient way?
• How do you know you’re getting the “right answer”? And how much accuracy do you actually need in your answers (not in the method of computing the answer)?
• How would companies run NS3 applications on the Cloud using a variety of vendor’s networking models in a secure fashion?
• What are the biggest challenges currently facing artificial intelligence?
• Do we need to be concerned that the knowledge about processors and systems seems to go downhill? Are we doing enough to make the next generation of developers to “think parallel”?
• How to guide and Drive semiconductor innovation to maximize efficiencies of large scale Data Centers?
• Virtualisation of heterogenous computing elements for HPC e.g. steps to make different architectures accessible via VMMs for on-demand interactive computing needs. Something like that at least. 🙂
• With CPU progress continuing to remain relatively stagnant, what are some of the potentially big disruptions that are poised to shake up the HPC space?
• How do we cope with the extreme complexity being thrust on us with the end of Moore’s Law?
• Keeping New Zealand a place of possibilities in IT, how to be leaders like NZ was with women voting, antiracism & the Rainbow Warrior?
• How to apply novel architectures for accelerators (networking, crypto and others) in the datacenter while not breaking current models of deployment and allowing everything to be controlled by the operating system in a logical manner?
• Future for ageing Geeks?
• The future of GPU-based storage?
• What are the coming hardware limitations for computing that need to be overcome with future materials choices, or that require new modalities of computing and logic (e.g. superconducting, quantum, neuromorphic computing)? What are the anticipated timelines for reaching these limits of performance, power, energy use…? Where can advanced materials research in New Zealand and elsewhere partner to address these problems?

• Is the Age of Linux coming to an end? Linux is everywhere, from embedded systems and smartphones up to hyperscale datacenters, but there are signs the big players may be contemplating a switch away from it (e.g. Google’s Fuchsia project –https://www.osnews.com/story/30923/android-emulator-picks-up-support-for-fuchsias-zircon-kernel/).

• Security is a cat-and-mouse game, and ever more sophisticated attacks require more and more resources to mitigate. With the end of Moore’s Law and cost scaling, how will we cope?

• Are we as a research and engineering community doing enough to uphold privacy and ethics in the face of mass data collection and surveillance capitalism?

Learn about colleagues who were at Multicore World 2019:

• Main professional activity / position
• Main fields of interest
• Any specific expectation from Multicore World 2019?
• Where did your colleagues come from?

---

Multicore World 2018 – BIG Questions

Updated 6 February 2018

• How can we restore reproducibility?
• How do we take potential into reality…
• HPC and Business Versus Fidelity
• What are the requirements and implications of HPC/BD/ML’s convergence for system software?
• What are the implications of widely heterogeneous computing on system software
• How will future HPC hardware look like?
• Scalability of compute and storage resources within New Zealand versus cloud-based resources
• The end of Moores Law (or Moores Law going backwards)
• Death of the POSIX file-system
• How does IoT and Computing on the Edge influence HPC?
• Blockchain will be an interesting topic even if it’s a short-lived one.
• The state of AI and how it influences developments in HPC
• “Blockchain” -a huge buzzword, interested on how much of it is hype versus how much lasting potential the technology will have.
• Panel on Meltdown/Spectre: for some workloads, these chip design flaws effectively dialed back a modest amount of progress.
• Privacy, Human Rights, and Ethics in the Big Data era
• Discuss specific niche workloads where Meltdown resulted in a large performance penalty as well, or if the impact is most severely limited to custom-tailored, worst-case-scenario benchmarks.

Learn about your colleagues who were at Multicore World 2018:

• Main professional activity / position
• Main fields of interest
• Any specific expectation from MW2018?

Featured Image (Top):

Multicore World 2018 – Mohamad S. El Zein (John Deere Fellow, USA)