gatk555

Yes, that one is more than 10 times slower when run with gdb. And my guess is also that making a system call for every AVR instruction is what slows...

Calling gdb_network_handler() once every 100 instructions makes the speed reduction about 30%, which seems acceptable. I tried to post a diff, but this site chewed it up as usual.

I tidied up my initial change and pushed to by fork as branch "gdb_burst". I would make a PR but the only validation is firmware.zip, which seems meagre. If anyone...

A probable work-round would be to convert the ELF file to ihex and load that. That does lose debug symbols. The problem sounds similar to #462, with the same cause:...

I recently did something similar and found the current manual to be less than ideal. I wrote up what I learned as a "Getting Started" guide and I would be...

My impression is that none of the communication peripherals are implemented at the "hardware" level: they do not change or read I/O pins. I think they just simulate the internal...

I think that the simulation of communication peripherals to the pin level was seen as being a low priority. So long as you are free to simulate the other end...

> I have got 'already raised' because ... I think it would help to include example code. It is not clear to me what a "fifo item" is here. Is...

This is a strange Pull Request: the change is described as tentative, but a PR should present a change that is considered ready for integration. I suggest closing it, and...

... and it was integrated [here](https://github.com/buserror/simavr/commit/d6ba83c1acf1b3c7ecabf83bf95cd7b1ceaeb35c). While that is recent enough that it is likely not in binary distributions, OP has forked the code, so should be using the current...