Follow-up Comment #2, bug #32551 (project avr-libc):
Frank Zhao: >>>> An example of the problem is demonstrated by this forum post as well http://www.avrfreaks.net/index.php?name=PNphpBB2&file=printview&t=100131&start=0 I have seen bug 30363 but it seems unrelated <<<< Something sounds very contradictory here. Your comment says the delay is 4 times longer while the thread talks about delay being 4 times less. If the problem is related to the topic discussed in thread, then it is exactly referring to bug30363. The output is 4 less because previous version of avr-libc ( < 1.7.0) used some algorithm where iterating through a loop takes 4 cpu cycles. The idea is to iterate through the loop delay/4 times so that required delay is achieved. However in 1.7.0, the loop has been replaced with builtins which provide exact number of delay cycles. However the same parameter which was sent to the loop (double __tmp = ((F_CPU) / 4e3) * __ms;) was used. This was fixed in 1.7.1 _______________________________________________________ Reply to this item at: <http://savannah.nongnu.org/bugs/?32551> _______________________________________________ Message sent via/by Savannah http://savannah.nongnu.org/ _______________________________________________ AVR-libc-dev mailing list AVR-libc-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/avr-libc-dev
