User forums > Embedded development

Developing for AVR, IDE tuning.

<< < (2/3) > >>

Lost_Byte:
None of the above reasons...
Debugger: JTAG-ICE mkII
No any debugging windows.

And now quite strange things... Trace-cursor does not appears, "Programm is not being run" in debugger window, "Failed to synchronise with the JTAG ICE (is it connected and powered?)" on C::B log....
Are these tools efficient, in general? Or it's just experimental? I mean avrdude+avarice...

Does anyone uses these seriously? ...Under C::B IDE?

scarphin:
Believe it or not there is no faster way to debug with hardware whether it's 'jtag-ice' or 'dragon'. It's not related to CB, it's not even much related to 'gdb' and 'avarice', that's how things are. You can also try AVR Studio's simulator.

Lost_Byte:
Simulator???
There are:
- SD-card with FAT FS on SPI;
- 2x C328R Cameras on UARTS;
- mma7455 accelerometer (SPI);
- M2M modem connection (UART with hardware flow)...
- Dallas ds1338 RTC (I2C).
in my project...

Years ago I've tried to implement  C328R driver under both simulavr and IAR-sym, only for sport. It is awful!!!


--- Quote from: scarphin on March 26, 2012, 01:01:45 am ---Believe it or not there is no faster way to debug with hardware whether it's 'jtag-ice' or 'dragon'. It's not related to CB, it's not even much related to 'gdb' and 'avarice', that's how things are. You can also try AVR Studio's simulator.

--- End quote ---

Sorry, I do not believe. Hardware debugging under AVR Studio is so fast as fast you can press F5 button, whatever debug views you have opened...
It is unfortunate that this is not a open source project.

scarphin:
By 'faster' I mean 'same speeds as PC debugging'. AVR Studio has its own protocol for PC side for debugging but it uses the same hardware in the end. I've also tried it but there is no considerable speed difference imo, you can't still catch the 'zero-crossing' of the mains voltage (which I needed at the time) and run the corresponding portion of the program till the next 'zero-crossing'. Afaik Obfuscated is working on a MI interface for the gdb which I think will speed up things, I don't know how much increase in speed it will attain though. Maybe Obfuscated can give some ideas. ;/

oBFusCATed:
The new gdb/mi plugin is faster only if gdb produces lots of output, because the protocol is asynchronous and so multiple commands can be sent to gdb at once.
If you don't see lots of lines in the Debug log pane, when full logging is enabled the problem is not in C::B's integration. It is somewhere inside gdb (probably).

Navigation

[0] Message Index

[#] Next page

[*] Previous page