You have the power!

Electronics Technical ArticlesSouth-East European INDUSTRIAL Мarket - issue 6/2007

Deeply embedded applications can now enjoy leading edge processing power, thanks to 32bit developments in microcontrollers.

Coming from a background of desktop applications, it would seem strange to consider using a processor that offers anything less than leading edge performance. For an engineer more entrenched in embedded applications, processing power has always been measured not just against the benefits it brings but by the resources it consumes.

Within the embedded domain, the irresistible force that is ‘faster, smaller, cheaper’ has always been stymied by the immovable object of available resources. This long-standing stalemate rests on the notion that too much processing power consumes not just system resources but financial ones too. An overabundance of performance or peripherals merely erodes an already beleaguered budget. However, it seems that notion is changing…

There can be little doubt that the demand for processing power is doing anything but increasing and so long as the benefits outweigh the penalties that scenario will continue, even within the embedded world. But within that world, it is only recently that the balance of power has swung significantly in favour of more processing power, thanks largely to the proliferation of 32bit cores within microcontrollers.

It is important to point out that any positive change in embedded engineers’ attitudes towards 32bit devices is an effect, as opposed to the cause of their penetration into the embedded domain.

For many years the deeply embedded – if such a term is still relevant given the level of connectivity in everyday objects – have favoured smaller, leaner devices underpinned by the mantra of ‘more than enough isn’t good enough’, for the reasons outlined above. That doesn’t just apply to processing power - in the deeply embedded domain, pin count is king, so any unnecessary I/O is as much a sin as unused horsepower.

Embedded engineers may have coveted the freedom a 32bit device offers, but they simply couldn’t afford it, until now. The latest fabrication technologies, coupled with aggressive design optimisation of processor cores, means the embedded engineer in general and the microcontroller user in particular can now afford – in all senses of the word – to dive into the 32bit pool.

For a 32bit microcontroller to be successful it needs to observe this history of frugality. If a family of devices comes with a hidden cost of adoption, that cost is unlikely to go unnoticed because even though budgets may now stretch further, they still have a limit.

As design complexity increases, that cost may relate more to the cost of development than the bill of materials, so any 32bit device introduced to the deeply embedded market needs to offer the right blend of performance, peripherals, tools and design solutions. Microchip believes it has found that blend, with the introduction of the PIC32 family of 32bit microcontrollers.

The launch of PIC32 comes at a time when the two leading IP providers in this field are seeing unprecedented growth. For its part, Microchip has chosen the MIPS32 M4K core to power its family of 32bit devices, instead of choosing to develop or extend its own technology (see block diagram).

The reasons for this are both technical and commercial. The MIPS core offers a best in class 1.5DMIPS/MHz operation, thanks to its proven 5-stage pipeline architecture, it also provides support for the MIPS16e code compression technology, which enables code size reductions of up to 40%. But in addition to these clear advantages, the MIPS architecture is recognised and supported by a growing ecosystem of Third Party tools and software vendors. This ecosystem becomes almost instantly available to all PIC32 users, allowing complex systems to be designed without relying on one particular vendor or the need to develop everything in-house, which can drastically reduce the time to market pressures faced by engineering teams in all sectors.

According to MIPS’ Vice President of Engineering, Pat Hays, the M4K has been used in low power applications for some time. This is largely thanks to its high performance and code compression technology, aided by shadow register sets which enable a fast interrupt or context switch response, which help make it applicable to applications where real time performance may be needed.

In addition, for Microchip’s family of 32bit devices, further enhancements were developed, as Hays explained: “We worked closely with Microchip to further extend software debugging capabilities, with the addition of complex hardware breakpoints and the new lightweight iFlowtrace technology.”

The addition of enhanced debug features will clearly help in the development of ever-more complex functions but perhaps more beneficial is the seamless migration from Microchip’s 8- and 16-bit PIC microcontrollers thanks to full compatibility within its MPLAB IDE tool chain.

Derek Carlson, Microchip’s Vice President of Development Tools, said: “As we near the 600 000th development system shipment, Microchip is extending the MPLAB toolsuite beyond our 8 and 16bit portfolio, in to the high performance 32bit microcontroller market.” The IDE will now support the PIC32’s advanced debugging features, including iFlowtrace technology, trace debugging and complex breakpoints. (see migration diagram).

The significance of this announcement is that with full API compatibility between the C compilers, engineers will have highly optimised code that is portable across Microchip’s 16bit and 32bit portfolios.

In the spirit of true multiple vendor support, however, Microchip’s IDE is not the only option for developers. At launch, the PIC32 is already supported by a large number of Third Party tools vendors, including HI-TECH Software who is also entering the 32bit market for the first time.

The tools vendor has been developing C compilers for Microchip devices since 1995 and enjoys the largest market share in Microchip compilers over all other Third Parties, thanks in part to its proprietary ‘Omniscient Code Generation’ technology (OCG). According to the company, OCG returns as much as 35% greater code density over competing compilers and it is expected to be particularly useful in the 32bit domain.

The company’s CEO, Clyde Stubbs, stated they chose to support Microchip for its first 32bit compiler because it is ‘the leader in the 8 and 16bit microcontroller market’. He went on to say: “They are offering a very easy migration path from their 16bit devices to the PIC32, with a unified development environment and pin compatibility between their 16bit and 32bit families.”

This highlights an important element in the ‘Microchip Mix’ for 32bit microcontrollers, not merely an easy software migration path but an even simpler hardware migration path. With pin compatibility across 16bit and 32bit devices, coupled with the unified software environment, the decision to use a 32bit device is made even simpler.

Patrick Johnson, director of Microchip’s High Performance Microcontroller Division, also welcomed Green Hills Software into the Microchip development tools program at the launch of the PIC32. He stated: “Green Hills Software’s embedded development solutions are recognised industry-wide as leading edge, and its MULTI integrated development environment enables PIC32 software developers to maximise performance while minimising code size, development and product unit costs.”

Of course, being MIPS based, the M4K Software Development Environment (SDE), available from MIPS, may also appeal to engineers already familiar with it. Supporting both Microchip’s MPLAB IDE and MIPS’ SDE, Ashling Microsystems has also made an announcement that its Pathfinder Debugger, AsIDE and emulation tools are available for the PIC32. Engineers using Ashling’s IDE are free to switch between Microchip’s and MIPS’ compilers.

Microchip’s Johnson stated that Ashling has over 10 years experience of the MIPS architecture, adding: “Combining the MPLAB C32 C compiler with Ashling’s integrated peripheral library gives developers everything needed to deliver a product to market.”

Low level software and middleware can often be a significant part of any 32bit development and here, again, Microchip’s choice in processor core brings a wealth of Third Party assistance.

A real time operating system (RTOS) is often essential in higher end applications and to support these devices, Microchip has brought in Express Logic’s ThreadX RTOS which, importantly, already supports the 16bit Microchip families.

In addition, CMX Systems has announced its support for PIC32, in the form of an RTOS and TCP/IP stack (CMX-RTX and CMX-MicroNet respectively).

Patrick Johnson explained: “These products provide a powerful suite of embedded software for designers using the PIC32 family. Our customers can benefit from the expanded compatibility in real time and networking applications.”

Thanks to the low-level compatibility, two other existing Microchip partners will also be supporting the PIC32 family from its introduction. Germany’s Segger Microcontroller Systeme GmbH brings its range of RTOS, graphical user interface and file system products, while Micrium brings its embedded software components – including an RTOS, file system, and TCP/IP and USB stacks.

While there will be continued demand for 8 and 16bit solutions, it is clear that the time for 32bit microcontrollers has arrived. While without the right mix of products and services the transition could prove challenging, Microchip believes it has achieved that mix and will help bring to market a new and exciting breed of embedded devices. Coming from a background of desktop applications, it would seem strange to consider using a processor that offers anything less than leading edge performance. For an engineer more entrenched in embedded applications, processing power has always been measured not just against the benefits it brings but by the resources it consumes.

Within the embedded domain, the irresistible force that is ‘faster, smaller, cheaper’ has always been stymied by the immovable object of available resources. This long-standing stalemate rests on the notion that too much processing power consumes not just system resources but financial ones too. An overabundance of performance or peripherals merely erodes an already beleaguered budget. However, it seems that notion is changing…

There can be little doubt that the demand for processing power is doing anything but increasing and so long as the benefits outweigh the penalties that scenario will continue, even within the embedded world. But within that world, it is only recently that the balance of power has swung significantly in favour of more processing power, thanks largely to the proliferation of 32bit cores within microcontrollers.

It is important to point out that any positive change in embedded engineers’ attitudes towards 32bit devices is an effect, as opposed to the cause of their penetration into the embedded domain.

For many years the deeply embedded – if such a term is still relevant given the level of connectivity in everyday objects – have favoured smaller, leaner devices underpinned by the mantra of ‘more than enough isn’t good enough’, for the reasons outlined above. That doesn’t just apply to processing power - in the deeply embedded domain, pin count is king, so any unnecessary I/O is as much a sin as unused horsepower.

Embedded engineers may have coveted the freedom a 32bit device offers, but they simply couldn’t afford it, until now. The latest fabrication technologies, coupled with aggressive design optimisation of processor cores, means the embedded engineer in general and the microcontroller user in particular can now afford – in all senses of the word – to dive into the 32bit pool.

For a 32bit microcontroller to be successful it needs to observe this history of frugality. If a family of devices comes with a hidden cost of adoption, that cost is unlikely to go unnoticed because even though budgets may now stretch further, they still have a limit.

As design complexity increases, that cost may relate more to the cost of development than the bill of materials, so any 32bit device introduced to the deeply embedded market needs to offer the right blend of performance, peripherals, tools and design solutions. Microchip believes it has found that blend, with the introduction of the PIC32 family of 32bit microcontrollers.

The launch of PIC32 comes at a time when the two leading IP providers in this field are seeing unprecedented growth. For its part, Microchip has chosen the MIPS32 M4K core to power its family of 32bit devices, instead of choosing to develop or extend its own technology (see block diagram).

The reasons for this are both technical and commercial. The MIPS core offers a best in class 1.5DMIPS/MHz operation, thanks to its proven 5-stage pipeline architecture, it also provides support for the MIPS16e code compression technology, which enables code size reductions of up to 40%. But in addition to these clear advantages, the MIPS architecture is recognised and supported by a growing ecosystem of Third Party tools and software vendors. This ecosystem becomes almost instantly available to all PIC32 users, allowing complex systems to be designed without relying on one particular vendor or the need to develop everything in-house, which can drastically reduce the time to market pressures faced by engineering teams in all sectors.

According to MIPS’ Vice President of Engineering, Pat Hays, the M4K has been used in low power applications for some time. This is largely thanks to its high performance and code compression technology, aided by shadow register sets which enable a fast interrupt or context switch response, which help make it applicable to applications where real time performance may be needed.

In addition, for Microchip’s family of 32bit devices, further enhancements were developed, as Hays explained: “We worked closely with Microchip to further extend software debugging capabilities, with the addition of complex hardware breakpoints and the new lightweight iFlowtrace technology.”

The addition of enhanced debug features will clearly help in the development of ever-more complex functions but perhaps more beneficial is the seamless migration from Microchip’s 8- and 16-bit PIC microcontrollers thanks to full compatibility within its MPLAB IDE tool chain.

Derek Carlson, Microchip’s Vice President of Development Tools, said: “As we near the 600 000th development system shipment, Microchip is extending the MPLAB toolsuite beyond our 8 and 16bit portfolio, in to the high performance 32bit microcontroller market.” The IDE will now support the PIC32’s advanced debugging features, including iFlowtrace technology, trace debugging and complex breakpoints. (see migration diagram).

The significance of this announcement is that with full API compatibility between the C compilers, engineers will have highly optimised code that is portable across Microchip’s 16bit and 32bit portfolios.

In the spirit of true multiple vendor support, however, Microchip’s IDE is not the only option for developers. At launch, the PIC32 is already supported by a large number of Third Party tools vendors, including HI-TECH Software who is also entering the 32bit market for the first time.

The tools vendor has been developing C compilers for Microchip devices since 1995 and enjoys the largest market share in Microchip compilers over all other Third Parties, thanks in part to its proprietary ‘Omniscient Code Generation’ technology (OCG). According to the company, OCG returns as much as 35% greater code density over competing compilers and it is expected to be particularly useful in the 32bit domain.

The company’s CEO, Clyde Stubbs, stated they chose to support Microchip for its first 32bit compiler because it is ‘the leader in the 8 and 16bit microcontroller market’. He went on to say: “They are offering a very easy migration path from their 16bit devices to the PIC32, with a unified development environment and pin compatibility between their 16bit and 32bit families.”

This highlights an important element in the ‘Microchip Mix’ for 32bit microcontrollers, not merely an easy software migration path but an even simpler hardware migration path. With pin compatibility across 16bit and 32bit devices, coupled with the unified software environment, the decision to use a 32bit device is made even simpler.

Patrick Johnson, director of Microchip’s High Performance Microcontroller Division, also welcomed Green Hills Software into the Microchip development tools program at the launch of the PIC32. He stated: “Green Hills Software’s embedded development solutions are recognised industry-wide as leading edge, and its MULTI integrated development environment enables PIC32 software developers to maximise performance while minimising code size, development and product unit costs.”

Of course, being MIPS based, the M4K Software Development Environment (SDE), available from MIPS, may also appeal to engineers already familiar with it. Supporting both Microchip’s MPLAB IDE and MIPS’ SDE, Ashling Microsystems has also made an announcement that its Pathfinder Debugger, AsIDE and emulation tools are available for the PIC32. Engineers using Ashling’s IDE are free to switch between Microchip’s and MIPS’ compilers.

Microchip’s Johnson stated that Ashling has over 10 years experience of the MIPS architecture, adding: “Combining the MPLAB C32 C compiler with Ashling’s integrated peripheral library gives developers everything needed to deliver a product to market.”

Low level software and middleware can often be a significant part of any 32bit development and here, again, Microchip’s choice in processor core brings a wealth of Third Party assistance.

A real time operating system (RTOS) is often essential in higher end applications and to support these devices, Microchip has brought in Express Logic’s ThreadX RTOS which, importantly, already supports the 16bit Microchip families.

In addition, CMX Systems has announced its support for PIC32, in the form of an RTOS and TCP/IP stack (CMX-RTX and CMX-MicroNet respectively).

Patrick Johnson explained: “These products provide a powerful suite of embedded software for designers using the PIC32 family. Our customers can benefit from the expanded compatibility in real time and networking applications.”

Thanks to the low-level compatibility, two other existing Microchip partners will also be supporting the PIC32 family from its introduction. Germany’s Segger Microcontroller Systeme GmbH brings its range of RTOS, graphical user interface and file system products, while Micrium brings its embedded software components – including an RTOS, file system, and TCP/IP and USB stacks.

While there will be continued demand for 8 and 16bit solutions, it is clear that the time for 32bit microcontrollers has arrived. While without the right mix of products and services the transition could prove challenging, Microchip believes it has achieved that mix and will help bring to market a new and exciting breed of embedded devices.

Paul Garden

Microchip Technology




LATEST issue 4/2022

issue 4-2022

  READ THE ISSUE ONLINE

ALL ARTICLES | ARCHIVE



Top