Multicore processors: Providing opportunities for embedded systems designers

October 01, 2008

Multicore processors: Providing opportunities for embedded systems designers

A highly integrated system can be constructed with real-time software components and human-directed elements running on separate cores in a single pro...

The big news in processor development is how major CPU manufacturers are now standardizing on multicore processor technology. While most of the software community has focused on server applications, developers across the wide spectrum of embedded computing applications can also benefit from the latest advances in multicore processors.

Multicore processors offer a solution to the need for mixing
new features with legacy code and combining multiple operating environments on
the same system. Compared to traditional embedded systems composed of multiple
subsystems, a highly integrated system can be constructed with real-time
software components and human-directed elements running on separate cores in a
single processing system, decreasing system manufacturing and maintenance costs
by eliminating redundant hardware.

The challenge is to implement software that efficiently
utilizes the new processor silicon. Today, systems are dedicating processor
cores to separate, distinct operating environments for both Real-Time Operating
Systems (RTOSs) and General-Purpose Operating Systems (GPOSs).

Sharing I/O at the expense of performance

Software that hosts multiple operating environments must
support virtualization of the processor's hardware interfaces so that multiple
software applications can share the multicore processor's I/O without conflict.
In this context, the concept of virtualization involves using software to allow
a single piece of hardware to service multiple OSs at the same time.

Historically, virtual machine management software has
emulated the entire underlying machine, including all the I/O devices. However,
using a completely virtualized machine imposes a performance penalty that the
guest OS does not have if it interacts directly with the hardware. For example,
graphics-intensive applications need access to real hardware for maximum
performance. A virtual frame buffer is too slow and lacks the adequate features
for an application that renders 3D moving images. This poses a major problem
for applications such as medical imaging systems or robotic assembly machines.
In such systems, the guest OS that renders the images needs direct access to
the physical frame buffer and its control I/O.

Direct access to I/O improves responsiveness

Given this performance setback, a different approach to
virtual machine management is needed to support the latest I/O hardware
enhancements and yield maximum performance in deterministic processing
environments. To address this problem, a Virtual Machine Manager (VMM, shown in
Figure 1) assigns specific devices directly to the I/O tasks that control them.
In this system, the VMM doesn't emulate the underlying machine's entire I/O
interface, only those devices that are shared. For all other devices, it
ensures that only authorized operating environments can access specific
performance-critical I/O. For example, as shown in the diagram, the VMM ensures
that the main operator display is only accessible to the GPOS, in this case


Figure 1




This notion of assigning I/O exclusively to a specific
virtual machine is essential to guaranteeing real-time responsiveness. Access
to response-critical hardware must be restricted to the RTOS that controls the
hardware; likewise, access to legacy I/O interfaces should be restricted to the
corresponding legacy application software.

Virtualization enables legacy code migration

Running a legacy RTOS in a virtual real-time machine on its
own processor core enables legacy real-time software to be migrated from
obsolete hardware to modern embedded platforms. Because I/O can be virtualized,
it is possible to simulate old hardware devices, which minimizes the need to
rewrite proven software. For example, a VMEbus system can be converted to a
less expensive SBC system by intercepting I/O requests to legacy VMEbus I/O and
redirecting them to equivalent onboard I/O devices.

An effective VMM system distinguishes resources that can be
multiplexed by the VMM from those that must be exclusive to a virtual machine.
For example, devices like the disk and an enterprise Ethernet interface can be
multiplexed and shared among all virtual machines. However, when determinism
and performance are more important than equal access, the virtualization
software should isolate resources for use by a specific virtual machine and its
guest OS.

Benefits of combining independent subsystems

Because a multicore chip can host multiple operating
environments, systems that previously required multiple discrete computing
modules can now be combined in a single hardware environment. By reusing proven
legacy applications and supporting faster communication and coordination
between RTOS and GPOS subsystems, this technology can decrease costs, improve
reliability and robustness, and save design, manufacturing, and maintenance

Paul Fischer is a
senior technical marketing engineer at TenAsys Corporation in Beaverton,
Oregon. He has more than 25 years of experience building and writing about
real-time and embedded systems in a variety of engineering and marketing roles.
Paul has an MSE from UC Berkeley and a BSME from the University of Minnesota.

[email protected]


Paul Fischer (TenAsys)