Loading ...
Sorry, an error occurred while loading the content.

Re: crazy optimization stuff

Expand Messages
  • angelovjasen
    Hi Jeff, I am also facing the same issue. Very, very strange optimization when we have such a good instructions as BRCLR .... Did you found some answers
    Message 1 of 4 , Oct 3, 2007
    • 0 Attachment
      Hi Jeff,

      I am also facing the same issue.
      Very, very strange "optimization" when we have such a good
      instructions as BRCLR ....

      Did you found some answers about this issue?

      Regards,
      Yasen
    • Jeff Smith
      ... The only answer right now is that developers either don t have time, or don t know how to fix it.
      Message 2 of 4 , Oct 3, 2007
      • 0 Attachment
        --- In gnu-m68hc11@yahoogroups.com, "angelovjasen" <angelovjasen@...>
        wrote:
        >
        > Hi Jeff,
        >
        > I am also facing the same issue.
        > Very, very strange "optimization" when we have such a good
        > instructions as BRCLR ....
        >
        > Did you found some answers about this issue?

        The only answer right now is that developers either don't have time,
        or don't know how to fix it.
      • angelovjasen
        ... Hm, I am very sad to hear this ... To be honest, this issue was the first think I ve noticed when testing with gcc - and it really disappointed me. I ve
        Message 3 of 4 , Oct 3, 2007
        • 0 Attachment
          --- In gnu-m68hc11@yahoogroups.com, "Jeff Smith" <imajeff84663@...> wrote:
          >
          > --- In gnu-m68hc11@yahoogroups.com, "angelovjasen" <angelovjasen@>
          > wrote:
          > >
          > > Hi Jeff,
          > >
          > > I am also facing the same issue.
          > > Very, very strange "optimization" when we have such a good
          > > instructions as BRCLR ....
          > >
          > > Did you found some answers about this issue?
          >
          > The only answer right now is that developers either don't have time,
          > or don't know how to fix it.
          >

          Hm, I am very sad to hear this ... To be honest, this issue was the
          first think I've noticed when testing with gcc - and it really
          disappointed me. I've compared the code generated with other compilers
          and it looks very good ... except this, which makes gcc code to be larger.

          I've also noticed that the compiler do not generate BRCLR in the
          case "if (PTT.bit.BIT0)" ... while it generates "bclr" when
          "PTT.bit.BIT0=0" ... But the code in this case is better, but not optimal:
          if (PTG.bit.BIT0)
          f0015: f6 00 02 ldab 2 <.Ldebug_info0+0x2>
          f0018: c4 01 andb #1
          f001a: 27 05 beq f0021 <.LM12>


          It seems I will abandon the gcc and gdb (with TBDML) experiments ...
          although I've managed to make TBDML to work at some level ...
        Your message has been successfully submitted and would be delivered to recipients shortly.