The Next Generation Hardware

FountainBlue’s July 10 VIP roundtable was on the topic of ‘Next Generation Hardware’, conducted online with introductory remarks provided by AMD.

We were fortunate to have a diverse range of executives in attendance who shared a wide range of perspectives around the Next Generation of Hardware.

Chips have been powering not just servers, laptops and devices, but now progressively more the Cloud, Gaming, Automotive, AI/ML which requires intensive acceleration of performance, and response times, while also respecting the privacy and security of users.

The hardware – including CPUs, GPUs, Tensor Cores, Digital Signal processing devices, IoT devices – facilitates the generation of the data whereas the software ensures that the right data is captured to drive the application, to report and measure on specific outcomes, to enabled data-based decision-making.

Below is advice from our esteemed hardware executives.

• As performance and response times as safety-critical – particularly in auto and health-related solutions – OR business-critical – particularly in manufacturing and production – the physical design of all the hardware involved in each solution must be scalable and flexible, working seamlessly with the software.
• The hardware helps to collect the data, but must be designed so that AI and ML integrated into the software can efficiently collect the relevant data, and provide real-time information to relevant stakeholders.
• The hardware must be modular enough to work with other hardware units, small enough to fit within a device, powerful enough to meet the needs of the customer, durable enough to withstand intensive usage, and efficient enough to work with minimal power.
  • As an example, the hardware must become even smaller and more portable, so the functionality is provided for demanding customers, in small form factors such as the phones which fit in our pockets!
• IoT devices will increasingly need to do some processing on the edge, especially when performance is critical. This is the ‘Empowered Edge’.
• Sort the data in terms of what’s most relevant, most urgent and to what audience, and give actionable real-time reports which would help them make critical decisions.
• Design the hardware to keep up with the explosion of data, and design it to be flexible enough to work with the software. 

Below are examples of specific enterprise use cases involving augmented reality hardware:

• Remote assistance, so that the expert can support the user to do everything from monitor or fix or manage equipment or devices from a distant location
• Guided Workflow, which supports the adoption of efficient processes
• Digital Collaboration on design and implementation

Below are examples of proactive management solutions related to the production of hardware.

• Predictive Maintenance to proactively manage when equipment needs parts or service
• Proactive management of Supply chain to ensure no one part is a limiting factor for production

Below are some thoughts about future trends and things to think about:

• Much as there has been a consolidation of architectures and GPUs and DSPs, there will also be a consolidation of AI accelerators. Create a software ecosystem to support the AI accelerator, to increase the likelihood of becoming a hardware standard.
• Leverage biological constructs to design solutions which can store and process immense amounts of data.
• Hardware does everything from managing the batteries on your phone to navigating home. How can the hardware work with the software to increase performance and accuracy? to do it with a smaller, more powerful footprint? to integrate with other functionality?
• The Work-From-Home phenomena resulting from the pandemic is exacerbating the adoption of laptop and smartphone hardware innovations as well. With everyone working (or not working) from home, the volume of data is amplified, the adoption of unstructured video data is magnified, and the demand for immediate and accurate response and support is urgent. 
  • What does this mean for chip designers, manufacturers, retailers, distributors? What kinds of innovations would suit this immense and quickly growing WFH user base?
  • What does this mean for the executive who wants to maximize operational and minimize IT issues, while addressing privacy, access and security issues?

We close with some provocative thoughts which might not be too far in the future.

• What’s next after the smart phone?
• How do we create an electronic mask for protection?
• How do we sanitize our clothing between washes?
• How can we leverage Lidar to better navigate our surroundings?
• How do we make brain computing a reality?

The pendulum swings back and forth between the hardware and the software, and both will always be important. 

Advertisement