Performance and Efficiency – A Timeless Quest with Evolving Requirements

Performance and efficiency are the central pillars that define the entire microprocessor business. The “Holy Grail” has always been to increase one without sacrificing the other, and probably always will.   Exponentially increasing demands of applications and deployment venues continuously challenge architects and designers to balance these pillars as they work to deliver ever-improving user experiences. Today’s users in particular see no reason for any arbitrary dividing lines between their experiences at home, on their smart devices, at work or in the car. These increasingly powerful and interconnected experiences cannot be allowed to mean short battery life or high cost to deliver.

Design engineers in long lead-time industries such as automotive, embedded, industrial, enterprise networking, servers and medical devices in particular face important decisions as they seek to provide that balance of performance and efficiency.  It can take years for an automotive company to design and deploy a cutting edge Infotainment or digital cluster design into the market.   As microprocessor architectures evolve rapidly it is critical that solutions which enhance performance and efficiency be available well in advance.

It seems like only yesterday we were discussing designing for the future and how to get the most from our top of the line i.MX applications processors with ARMv7 architecture (based on the 32-bit ARM® Cortex®-A8 and –A9 cores).  NXP is preparing for tomorrow’s need today with a new generation of i.MX processors based on the ARMv8 architecture.  With Arm’s recent introduction of the 64-bit Cortex-A72 core, a critical piece of that new generation is in place.

In addition to the raw specs, which are impressive enough, there are several reasons the Cortex-A72 is compelling to NXP and our customers:

• Earlier access to an advanced platform – NXP has been working with the Cortex-A72 from the beginning.   Again, especially in long lead-time applications, the design cycle demands use of the most advanced architecture possible today, in order to be ready for tomorrow.  In automotive applications, for example, the driver information systems and other compute- and media-intensive applications a few years from now will make today’s in-dash systems appear primitive.
• Design flexibility – the Cortex-A72 core is designed to work in conjunction with the Cortex-A53 core, allowing for more flexibility in overall design and performance efficiency delivered.
• Software “future-proofing” – the new ARMv8 architecture is backward-compatible with earlier 32-bit applications, allowing for a smooth transition and a clear path to the future of 64-bit computing.  NXP considers all our designs going forward as being part of the Internet of Things (IoT).  In this world no device or application is an island, and ubiquitous connectivity is enhanced with an advanced architecture.
• Memory addressability – memory requirements go only in one direction – upward.  The 64-bit architecture of the Cortex-A72 provides addressability for terabytes of DRAM, vs. the 4 gigabytes of previous generation 32-bit architecture.
• Environmental flexibility – environments such as those found in-dash are thermally challenging, and the improved power efficiency of the Cortex-A72 core, especially when combined with the Cortex-A53, provides additional flexibility without risking onerous temperature increases.

It seems that stringent market demands always arrive sooner than expected.  The time to anticipate the future is now, in order to capture the imaginations and experiences of tomorrow’s customers.  The transition from a 32-bit world to a 64-bit one is a natural step, made easier by the Cortex-A72 and one to be taken sooner, rather than later.

Contributed by Kyle Fox, i.MX Product Line Manager and Dan Loop, i.MX Automotive Business Development