The Multicore Software Development Kit (MSDK) is a set of tools that developers can use to test, debug and analyze applications targeted for multicore hardware systems. Now day, multicore architectures have become very popular as it helps developers in parallelism software developing of processor architectures. In this exclusive article we have bought to you some latest advancement in MSDK from experts around the world.
The challenges in designing a modern System-on-Chip (SoC) have stretched both Electronic Design Automation (EDA) tools, as well as traditional Embedded Software methodologies, to breaking point. A new generation of technologies is necessary to address development issues beyond the RTL and C code design levels, targeting “Electronic System Level,” or ESL. Design architectures are changing. With the increase in integrated circuit fabrication costs and complexities, a movement towards standardized platforms, with greater levels of specific functionality coded in software, is apparent. To maintain the growth in performance required for these modern, software focused electronic systems, the use of multi-core processors has become prevalent. To effectively develop software for multi-core devices, new development technologies, methodologies and tools are required. The requirement has evolved to build new solutions for multi-core embedded software development. The Multicore Software Development Kit (MCSDK) provides the productivity necessary to accelerate engineering schedules, while increasing code quality, in the light of multi-core complexity.
A Multicore Software Development Kit (MCSDK) provides the core foundational building blocks for customers to quickly start developing embedded applications on TI high performance multicore DSPs. MCSDK tools are capable of analyzing various concurrency related problems in the applications which arise due to different types of memory and synchronization errors. Also, the MSDK contains tools to do lock analysis and synchronization coverage analysis. A MCSDK uses the SYS/BIOS or Linux real-time operating system, accelerates customer time to market by focusing on ease of use and performance and provides multicore programming methodologies. Basic features of a typical MSDK can be summarized as:
Standard set of APIs to configure and utilize peripherals, accelerators, and other hardware resources
Compliant implementation of packet‐ and network‐based protocols
Utilities to boot, test, debug, and monitor execution
Individual core‐based real‐time operating system
Communication and facilitated cooperation between processes, cores, and devices, as well as between peripherals and cores
Optimized, ready‐to‐use algorithm libraries, example code, and demonstration applications, provides the core foundational building blocks for customers to quickly start developing embedded applications on multicore DSPs: Uses either the SYS/BIOS or Linux real‐time operating system or accelerates time‐to‐market by focusing on ease of use and performance
Provides multicore programming methodologies and utilities
Simplifies porting of applications: to a standard evaluation platform, from a standard evaluation platform to customer’s target platform and to next generation platform hardware
The ever-increasing global appetite for computing power and performance has put embedded computing squarely on the path to multicore systems. Although the move to multicore began in high-end applications such as networking and wireless infrastructure, multiple cores are appearing in less-traditional areas such as automotive, the safety-critical aspects of industrial applications and even printer and imaging applications any application that could use more speed and power. The move toward global network connectivity is straining networks and increasing the need for more processing power. Single-threaded performance can’t continue to improve without geometrically increasing power, which is counterproductive in terms of system cost and reliability. To keep power at manageable levels, applications must move to multicore instead of continuously increasing frequency.
Achieving multicore performance levels is not as simple as putting more cores into a system. There are complex issues surrounding the migration of legacy software from single-core to multicore hardware. Software engineers lack training in this kind of migration, and in multithreaded programming in general. Virtualization, which is necessary to get the most out an investment in multicore hardware, is a completely different way of thinking about operating systems and resource partitioning.
Two big roadblocks on the path to multicore involve software. The first problem is one of migration. Millions of lines of legacy code for single-processor sequential programs must be migrated to preserve investments.Until that mapping and transition becomes easier and more efficient, it will be somewhat painful to move from single-core to multicore systems. The second problem is a lack of basic knowledge about programming for multicore systems. It’s fairly easy to install multiple cores from a hardware perspective, but programming for multicore requires a different mindset, and different education, than traditional single-core programming.
