FONT SIZE : AAA
The best use of FPGAs and processors can be explained through a couple of simple applications.
One of the basic usages of soft processors is to act as a microcontroller which moni- tors a video pipeline engine and responds to interrupts when something unusual is observed in the video pattern. For example, a security camera can detect moving patterns and compare faces to a central database. On finding a match with a person of criminal background, it can set an alarm which can notify the right offi cials of the presence of such a person on premises. A system can be represented in a block dia- gram as shown in Fig. 6.4 . The block diagram shown in Fig. 6.4 is overly simplified to explain the concept and the role of the processor. In a real system, there would be an additional requirement of fl ow control as well as more connections.
Xilinx provides IPs such as HDMI controllers, memory controllers, interrupt controllers, and MicroBlaze for building systems, while you will have to provide the special secret sauce such as facial recognition and the database lookup and com- pare code. The solid arrows show the typical datafl ow from the external camera image and the memory lookup, while the dashed lines indicate the control fl ow. The MicroBlaze processor controls all the IPs in the system and is usually respon- sible for initialization and periodic status checks. This will depend on the software developed. Without a processor, you would have to write up complex state machines to ensure that the entire system works in tandem.
You could potentially decide to port the entire application to an SoC family of devices. The MPSoC has all the peripheral IPs necessary for realizing a system as shown in Fig. 6.5 . This would require the secret sauce (such as the facial recognition and database lookup) to be written in a software, which is compiled to the ARM processor. But even with a 1.5 GHz processor, it is hard to match the performance of dedicated computation in an FPGA.
The facial recognition and comparison algorithms could take thousands of clock cycles to detect an image and compare it to a picture in the database. If the job is done in a hardware (i.e., programmable logic), the entire algorithm could be paral- lelized and be done in a few clock cycles. Xilinx tools make it easy for embedded algorithm developers to take the compute-intensive functions through HLS tool
Fig. 6.5 Basic application
Fig. 6.6 Accelerated system with MPSoC
chain to create a hardware component which is a faster and a parallel version of the software. The tool chain also creates the necessary connection with the processor which can continue to take care of data acquisition and overall control. A simplifi ed view of the application after running through the tool chain would look similar to the one shown in Fig. 6.6 .
Manufacturer:Xilinx
Product Categories: FPGAs
Lifecycle:Active Active
RoHS:
Manufacturer:Xilinx
Product Categories:
Lifecycle:Obsolete -
RoHS: No RoHS
Manufacturer:Xilinx
Product Categories: Microcontrôlleur
Lifecycle:Obsolete -
RoHS: No RoHS
Manufacturer:Xilinx
Product Categories: CPLDs
Lifecycle:Active Active
RoHS: No RoHS
Manufacturer:Xilinx
Product Categories:
Lifecycle:Any -
RoHS: -
Support
