If users of embedded development tools are asked what features they need to meet the needs of their next generation of ultimate processing design, they will inevitably answer: “performance and advanced processing capabilities.” In order to meet the requirements of embedded environment, engineers must try to dig out the performance from most of the existing cost-effective devices, and they need convenient tools in both design and debugging. Therefore, it is very important for engineers to choose a development tool with powerful performance and advanced processing functions.
Xilinx platform studio (XPS) is a widely used embedded development kit (EDK). Its upgraded version xps8.2 adds unique functions and performance, which can further improve the efficiency of embedded system design. Through this new version, engineers can obtain embedded processing support for FPGA products that are currently very fast in the industry, such as virtex-5lx devices. Here are some new features and enhanced performance of xps8.2.
Xps8.2 adds a new highly optimized MicroBlaze soft core, supports integrated PowerPC floating point unit (FPU), and adds support for the latest virtex-5lx devices. Instead of just using the clock frequency of a single processor to measure the degree of performance improvement, embedded design engineers should take advantage of the XPS suite that integrates all the advantages of faster hard processing core, flexible soft processing core and additional CO processing engine, whether they are FPU, other IP or additional processing core for unloading and distributing workload. In addition, the new XPS adds a lot of enhanced functions, debugging visibility, and improves the convenience of use.
New devices and processing support
Virtex-5lx device integrates advanced system functions on the 65nm triple gate oxide architecture, so it can fully support embedded processing. Compared with the previous generation FPGA, Virtex-5 technology improves the performance of semiconductor devices by 30%, while reducing the power consumption by more than 30%. Flexible processing of IP peripherals allows engineers to build customized solutions precisely without adding additional device costs or wasting expensive but unnecessary device choices.
In order to make the most effective use of embedded design in high-performance virtex-5lx devices, Xilinx optimizes the soft core of MicroBlaze for 65nm process, and adds some special enhancements to achieve less cycles per instruction (CPI). Xilinx’s microblaze5.0 processor is equipped with 5-level pipeline and optional 4 or 8-word cache memory to improve performance. The new MicroBlaze embedded processor is backward compatible with the code written for the initial instance, so engineers won’t have trouble with processor degradation. Engineers who use Spartan series FPGAs for embedded design can continue to use the existing microblaze4.0 core, which has been optimized for high-performance applications with minimal footprint.
In fact, in the design wizard of base system builder, platform studio tool suite automatically illustrates the optimized MicroBlaze core for the appropriate FPGA device. Unless the engineer wants to do this manually, let the design wizard optimize the core configuration for you to speed up the design. Engineers don’t need to “port” pre stored designs or code just to take advantage of a faster Virtex-5 device and a higher performance MicroBlaze processor.
Performance and debugging visibility
In addition, one way to maximize the overall performance of embedded systems is to use the concept of CO processing. Using virtex-4 devices with powerpc405 processor cores, platform studio 8.2 introduces support for FPU implementation. By illustrating the performance of FPU in parallel FPGA architecture, significant performance improvement can be achieved.
As embedded systems become more complex, debugging tools need to provide a higher level of visibility. Because the traditional partition between hardware and software components is fuzzy, engineers can no longer only rely on hardware tools to debug hardware and software tools to debug code, but also make use of the progress in hardware implementation of devices, processing, coprocessing and software instructions.
As the first step in introducing “platform debugging”, Xilinx has integrated hardware and software debuggers to provide system level observation and reduce conflicts between hardware and software engineers. The integrated chipscopepro hardware debugger cross triggers GNU software debugger, and vice versa. Engineers can now repeatedly find system faults from two directions. If a problem is found at a hardware address or after a hardware event, simply cross trigger the software debugger to see what the software is doing; if the system crashes after entering some software function “foo”, then cross trigger the hardware debugger to see what the hardware is doing.
In the evolution process of platform debugging, the second step is to use the results of Agilent and Xilinx alliance. Agilent company provides an e9524amicroblaze tracking tool set, which can completely decompose traces through Agilent’s Windows XP based logic analyzer. It is an easy-to-use solution to track and reconstruct MicroBlaze program execution history by using data flow
The MicroBlaze tracking core (MTC) provided by xilinxedk version 8.2 is exemplified in the design to provide visibility for Agilent’s tracking toolset. You can capture real-time code flows that are related to measurements from surrounding systems. By turning on the cache, powerful triggering and deep memory tracking are unaffected, as probes appear in the execution state of the MicroBlaze pipeline. The system enables the hardware and software engineers to work together to find and locate the design defects in the system as soon as possible. For more information about Agilent products, please visit.
Improvement of productivity and ease of use
In addition to the powerful new trace debugging performance, xps8.2 also includes many enhancements and ease of use improvements. These improvements further realize the new IP core, RTOS support and installation, as well as new functions and other productivity improvements.
Better integration between components within the XPS tool suite, linker scripts, and data table generation can improve overall productivity. Version 8.2 improves the visibility of IP core, which is helpful for version control, and accelerates the performance of calling internal tools and loading large-scale designs.
The new configuration wizard (see Figure 2) helps to set up debug time periods and make them more quickly and easily reproduced in subsequent time periods. This version includes support for PowerPC cache memory burst mode, hard three mode Ethernet MAC, dispersion / aggregation DMA, and numerous new memory controller interfaces. In addition, version 8.2 is the first platform studio version to support VxWorks 6.3 of WinDriver systems and embedded Linux 2.6 of MontaVista.
Version 8.2 extends the foundation of the platform studio tool suite, adding new options for processing and device support. The support for FPU enhances the already fast PowerPC processing solution. In addition to general XPS enhancements and improvements to accelerate productivity, version 8.2 brings platform debugging to a new stage. The combined use of Xilinx’s MicroBlaze tracking core and Agilent’s tracking toolset brings new and unique visibility to the complex embedded system. To learn more about the new features in the full version 8.2, please visit.