Hi guys,
I need to do some optimisation on a piece of code and wanted to try out the
profiling functionality of EW. Now I activated it like in the manual. This
means:
- C/C++ Compiler: Output>Generate debug information
- Linker: Format>Debug information for C-SPY
- Debugger: Plugins>Profiling
I can now start and stop the profiler by its on/off button.
But when I start my application it either says:
WITHOUT SOFTWARE BREAKPOINTS:
Mon Dec 21 17:00:32 2009: Profiler: No cycle counter could be found. Profiler
will only count function calls.
Mon Dec 21 17:00:33 2009: Failed to set hardware breakpoint. All available
resources are already in use.
Mon Dec 21 17:00:33 2009: Virtual breakpoints can be enabled in Options dialog
under Debugger->FET Debugger->Breakpoints
Mon Dec 21 17:00:33 2009: Profiler: Could not set enough breakpoints.
WITH SOFTWARE BREAKPOINTS:
Mon Dec 21 17:11:36 2009: Profiler: No cycle counter could be found. Profiler
will only count function calls.
Mon Dec 21 17:11:36 2009: Virtual breakpoint is used. This will force single
stepping on 'Go'.
Mon Dec 21 17:11:39 2009: Profiler activated
Mon Dec 21 17:11:45 2009: Performance warning: Use of virtual breakpoints forces
singlestepping.
Mon Dec 21 17:11:45 2009: Performance warning: Use of virtual breakpoints forces
singlestepping.
Any of you guys have seen similar issues?
Thanks
Zjeriet
Profiling with IAR EW
Started by ●December 21, 2009
Reply by ●December 21, 20092009-12-21
Zjeriet,
You just posted two groups of messages, but no 'issue' to be solved. All I can say is that the Enabling/Disabling of software breakpoints is in accordance with the difference between both groups of messages.
What's the problem?
Michael K.
--- In m..., "zjeriet" wrote:
>
> Hi guys,
>
> I need to do some optimisation on a piece of code and wanted to try out the profiling functionality of EW. Now I activated it like in the manual. This means:
> - C/C++ Compiler: Output>Generate debug information
> - Linker: Format>Debug information for C-SPY
> - Debugger: Plugins>Profiling
>
> I can now start and stop the profiler by its on/off button.
> But when I start my application it either says:
>
> WITHOUT SOFTWARE BREAKPOINTS:
> Mon Dec 21 17:00:32 2009: Profiler: No cycle counter could be found. Profiler will only count function calls.
> Mon Dec 21 17:00:33 2009: Failed to set hardware breakpoint. All available resources are already in use.
> Mon Dec 21 17:00:33 2009: Virtual breakpoints can be enabled in Options dialog under Debugger->FET Debugger->Breakpoints
> Mon Dec 21 17:00:33 2009: Profiler: Could not set enough breakpoints.
> WITH SOFTWARE BREAKPOINTS:
> Mon Dec 21 17:11:36 2009: Profiler: No cycle counter could be found. Profiler will only count function calls.
> Mon Dec 21 17:11:36 2009: Virtual breakpoint is used. This will force single stepping on 'Go'.
> Mon Dec 21 17:11:39 2009: Profiler activated
> Mon Dec 21 17:11:45 2009: Performance warning: Use of virtual breakpoints forces singlestepping.
> Mon Dec 21 17:11:45 2009: Performance warning: Use of virtual breakpoints forces singlestepping.
>
> Any of you guys have seen similar issues?
>
> Thanks
>
> Zjeriet
>
You just posted two groups of messages, but no 'issue' to be solved. All I can say is that the Enabling/Disabling of software breakpoints is in accordance with the difference between both groups of messages.
What's the problem?
Michael K.
--- In m..., "zjeriet" wrote:
>
> Hi guys,
>
> I need to do some optimisation on a piece of code and wanted to try out the profiling functionality of EW. Now I activated it like in the manual. This means:
> - C/C++ Compiler: Output>Generate debug information
> - Linker: Format>Debug information for C-SPY
> - Debugger: Plugins>Profiling
>
> I can now start and stop the profiler by its on/off button.
> But when I start my application it either says:
>
> WITHOUT SOFTWARE BREAKPOINTS:
> Mon Dec 21 17:00:32 2009: Profiler: No cycle counter could be found. Profiler will only count function calls.
> Mon Dec 21 17:00:33 2009: Failed to set hardware breakpoint. All available resources are already in use.
> Mon Dec 21 17:00:33 2009: Virtual breakpoints can be enabled in Options dialog under Debugger->FET Debugger->Breakpoints
> Mon Dec 21 17:00:33 2009: Profiler: Could not set enough breakpoints.
> WITH SOFTWARE BREAKPOINTS:
> Mon Dec 21 17:11:36 2009: Profiler: No cycle counter could be found. Profiler will only count function calls.
> Mon Dec 21 17:11:36 2009: Virtual breakpoint is used. This will force single stepping on 'Go'.
> Mon Dec 21 17:11:39 2009: Profiler activated
> Mon Dec 21 17:11:45 2009: Performance warning: Use of virtual breakpoints forces singlestepping.
> Mon Dec 21 17:11:45 2009: Performance warning: Use of virtual breakpoints forces singlestepping.
>
> Any of you guys have seen similar issues?
>
> Thanks
>
> Zjeriet
>
Reply by ●December 22, 20092009-12-22
Hey Michael,
Sorry about not being completly clear on the problem
I'm not able to get the profiling info as my app can't run or can't run normally due to the described scenarios.
I'm running an RTOS and this might complicate the matter.
EW is also popping up a message:
Profiler: No cycle counter could be found. Profiler will only count function calls.
But even that info is not available.
Greets,
Zjeriet
--- In m..., "Michael" wrote:
>
> Zjeriet,
> You just posted two groups of messages, but no 'issue' to be solved. All I can say is that the Enabling/Disabling of software breakpoints is in accordance with the difference between both groups of messages.
> What's the problem?
>
> Michael K.
>
Sorry about not being completly clear on the problem
I'm not able to get the profiling info as my app can't run or can't run normally due to the described scenarios.
I'm running an RTOS and this might complicate the matter.
EW is also popping up a message:
Profiler: No cycle counter could be found. Profiler will only count function calls.
But even that info is not available.
Greets,
Zjeriet
--- In m..., "Michael" wrote:
>
> Zjeriet,
> You just posted two groups of messages, but no 'issue' to be solved. All I can say is that the Enabling/Disabling of software breakpoints is in accordance with the difference between both groups of messages.
> What's the problem?
>
> Michael K.
>