Atmel is obviously a modern company promoting bloat :~) On Fri, 12 Sep 2008 21:17:04 +0200 Ruud Vlaming <[EMAIL PROTECTED]> wrote:
> Hi all, > > Just too share some thoughts. I used to work with the > avr's like avr ATtiny861/461/261 and ATmega16. Now i > am porting a program to other chips like the ATmega > 48P - 328P series and ATmega 164P-644P series. > > When i look at the register summary, it really makes > me sad. Vast area's of reserved space and precious > registers are push beyond the 0x1F or 0x3F boundaries > for no obvious reason. Now i understand you cannot > squeeze everything into the lowest region. But why so > much empty space? The use of sbi/cbi and in/out could > be a lotter better. It is not just the number of instructions > that goes up needlessly. Also, sbi/cbi makes interrupt > immune switching of bits possible. > > For example, if at least the TIMSK registers had been > below 0x1F (makes atomic enable/disable of > timerinterrupts possible) or there been a few extra GPIO > present below 0x1F. > > The way this is designed does not honor the efficient way > the instructionset was meant to operate imho. > > I'dd like to hear oppinions and hope someone of the > Atmel hardware team cares to react. > > Regards, > Ruud. > > > > > > _______________________________________________ > AVR-chat mailing list > AVR-chat@nongnu.org > http://lists.nongnu.org/mailman/listinfo/avr-chat -- Unna dig ett bättre liv, dumpa Windows! _______________________________________________ AVR-chat mailing list AVR-chat@nongnu.org http://lists.nongnu.org/mailman/listinfo/avr-chat
