For systems using a microprocessor with no built-in trace port, Green Hills Software’s TraceEdge trace collection solution makes it possible to tap in to the advantages of the MULTI TimeMachine Suite without switching microprocessors.
TraceEdge requires requires a minimal amount of software instrumentation to enable full TimeMachine Suite support. As a result, TraceEdge offers several unique benefits over
other instrumentation solutions:
- Minimizes impact on run-time performance. In the common case, TraceEdge inserts only a single instruction at each instrumentation point
- Can be quickly added to production code without timeconsuming recompilation since it merely requires you to re-link your program
- Is independent of programming language and does not require source code
- Enables your entire system to be traced, including application software, libraries, the operating system, and device drivers
|
| MULTI TimeMachine Suite |
To get the full benefit from trace data collected with the SuperTrace probe, Green Hills Software offers the MULTI TimeMachine Suite of analysis tools. This innovative collection of tools extends the range of Green Hills Software popular MULTI IDE by providing a window into complex software interactions that can result in bugs, performance problems, and testing nightmares. The TimeMachine Suite presents this information in easy-tounderstand displays that enable developers to quickly navigate through trace data and produce better code in less time. |
|
|
Helpful throughout the development cycle
The MULTI TimeMachine Suite offers benefits throughout the development cycle:
- Development proceeds more quickly when using
TimeMachine’s innovative debugging capabilities to help find and fix bugs faster
- Integration flows smoothly when TimeMachine’s powerful performance analysis tools are put to work uncovering system bottlenecks
- Testing is more thorough when test code is verified using TimeMachine’s advanced code coverage tools
|
|
Fix bugs faster
At the heart of the TimeMachine Suite, the TimeMachine Debugger combines a familiar debugger interface with innovative functions that enable developers to step and run forward and backward through their code. In addition, all standard debugging tools are available, such as execution
and data breakpoints. Using this capability, developers can more easily track down the root cause of bugs—even if they occurred long before the problem manifested itself. With TimeMachine, developers can more readily explore the execution of their processor and, as a result, fix bugs faster.
|
|
|
The TimeMachine Debugger also provides functions for:
- Viewing register and memory values at any point in the trace data
- Finding specific points in the trace using complex execution and data breakpoints
- Examining RTOS data structures to quickly debug task interaction bugs
- Debugging virtual memory applications, such as those in systems running under the INTEGRITY RTOS
Optimize with ease
To get the maximum performance from your embedded applications, the MULTI TimeMachine Suite offers these unique capabilities:
- Easily measure important system metrics such as interrupt latency, context switch time, and boot time with great accuracy and precision
- Perform detailed performance analysis over long periods of time
- Quickly debug a performance problem once it’s located
Test with confidence
The MULTI TimeMachine Suite’s powerful code coverage tools make it easy to determine which particular instructions in the application were executed. This information makes it easy to analyze the quality of your test suite and identify segments of code that have not been thoroughly tested.
|
| MULTI-based analysis tools |
To help application programmers efficiently locate key execution points in large SuperTrace datasets, Green Hills Software’s MULTI Integrated Development Environment supports the following features:
|
 |
Powerful and Easy-to-use Triggering—While SuperTrace probe’s extreme trace depth minimizes the need to devise complex trigger conditions, MULTI still offers a powerful interface for setting trace triggers. SuperTrace probe supports a target CPU’s triggers with an intuitive graphical interface that uses combinatorial logic statements and state machine diagram and eliminates the need to look up register bit definitions or manually type tedious command-line instructions.
Commonly-used triggers can be set from the MULTI debugger to enable developers to immediately use the SuperTrace probe without first becoming an expert in a particular target’s triggering intricacies. Example triggers provided include tracing a particular function, triggering
about a particular line of source code, and tracing a global data variable.
|
| |
|
Complex Data Searching—Searching through a gigabyte of trace data is a daunting task and requires special tools to quickly find trouble spots. Using the same graphical interface as the trigger tools, developers can create complex combinatorial logic statements and state machine diagrams to precisely locate and filter triggering. Search filters can be saved for future use.
Once found, locations can be bookmarked and highlighted or hidden to better organize the large trace data collection.
|
|
|
Statistical Analysis—To help developers optimize code, MULTI calculates and displays execution statistics such as the memory locations most read and written, the number of branches executed, the number of branches taken, and, if available on the target CPU, the number of cycles executed per instruction, function, task, and address space.
|
|
|
Streamline Navigation—Because very large trace buffers can take a while to upload, MULTI can display incremental portions of trace data for analysis while the remaining data continues to upload in the background.
|
|
|
Trace Display Interface—MULTI offers a traditional trace data display that lists a sequence of executed instructions and correlates them to source code.
|
|
|
RTOS-aware display— When debugging a real-time operating system such as INTEGRITY and velOSity, the trace window displays the task that executed the trace instruction and can be organized by tasks rather than instructions. This feature enables developers to quickly
scan a long display list to find specific tasks and expand selected tasks to see only those tasks’ instructions.
|
|
|
)
)
