Products Downloads Try/Buy Support News and Events Contact

There is no limit imposed by Gedae on the size of the graph or the number of function boxes contained in them. The memory on the development system, and ultimately on the target hardware, will limit the size of the application.


return to top

Gedae does not provide extra support for evaluating precision. If desired, the user can create new functions that act on data of different precision. Gedae is intended for use as an application development and software generation tool for multiprocessor embedded systems. Gedae applications are executed on existing COTS hardware in which bit precision and base data type (for example, 32-bit floating point or 32-bit integer) are already defined.


return to top

Gedae does not display "bus loading" directly. Communication activity information is displayed in the "Trace Table" in both the software view and the hardware view. However, statistical information is now available from the Trace Table. Hardware statistical data, including the percentage of time that a processor spends in primitive execution, sending data, receiving data and in Gedae overhead functions is provided. In addition, the burst data rate and throughput rate is also provided. The statistical data is calculated over the time period for which the Trace Table information has been gathered.


return to top

Running multiple processes on the same processor is frequently desirable as it simplifies the design by allowing logically concurrent threads of execution to run independently. This can avoid complicated polling logic that would be necessary if these concurrent tasks needed to run in the same process. The effectiveness of running multiple processes on the same processor depends to some extent on the services provided by the native operating system.

Gedae provides many of the benefits of running multiple processes by mapping independent sections of a graph to the same Gedae process. Within that process the Gedae dynamic scheduler switches between the different subgraphs that were mapped to that processor. This is quite efficient as there is no context switch involved - Gedae's dynamic scheduler merely calls the static scheduler that executes the subgraph that is to be scheduled next. The developer has a great deal of control over the dynamic scheduling properties. (See FAQ: Real Time Control in Gedae).

Multiple processes running on the same processor will occur if a user chooses to map several partitions members to the same processor. They can also occur if processes independent of Gedae need to run on the processor. The efficiency of switching these processes depends much on the native operating system. Gedae is not directly involved in the switching strategy between processes running on the same processor though Gedae Board Support Packages (BSPs) can be designed to allow the user to tune the switching properties of the processes within the constraints of the services provided by the native OS.

An example is the Gedae Embedded Solaris (ESolaris) BSP which is a BSP that allows Gedae to run in parallel on a network of Sun processors. This BSP has been generated to allow processes to explicitly yield control when they are blocked in reads or writes. By explicitly yielding control the application does not need to wait for a round-robin timeout to switch the process out.


return to top


return to FAQ list

 
Copyright © Gedae, 2007.