--max-allocs-per-node 300 took 6 minutes to compile, caused a bank overflow and can not generate a binary file to test.
Woody http://palmmicro.com/woody/ ----- Original Message ----- From: "Lin Rongrong" <wo...@palmmicro.com> To: <sdcc-user@lists.sourceforge.net>; <supp...@palmmicro.com> Sent: Wednesday, July 13, 2011 10:37 AM Subject: Re: [Sdcc-user] Z80 code size reduction > --max-allocs-per-node 500 took 7 minutes to compile, and generated another > different run time error pattern. > > Woody > > http://palmmicro.com/woody/ > > ----- Original Message ----- > From: "Lin Rongrong" <wo...@palmmicro.com> > To: <sdcc-user@lists.sourceforge.net>; <supp...@palmmicro.com> > Sent: Wednesday, July 13, 2011 10:24 AM > Subject: Re: [Sdcc-user] Z80 code size reduction > > >> --max-allocs-per-node 100 reduced compile time to 5 minutes. But it >> generated larger code size than 3.0.1, caused a bank overflow in my case >> and thus can not generate a binary file to test whether the result is >> correct or wrong. >> >> Woody >> >> http://palmmicro.com/woody/ >> >> ----- Original Message ----- >> From: "Lin Rongrong" <wo...@palmmicro.com> >> To: <sdcc-user@lists.sourceforge.net>; <supp...@palmmicro.com> >> Sent: Wednesday, July 13, 2011 10:09 AM >> Subject: Re: [Sdcc-user] Z80 code size reduction >> >> >>> --max-allocs-per-node 1000 reduced compile time from 34 minutes to 9 >>> minutes. The result is still wrong, but in a different pattern now. >>> >>> Woody >>> >>> http://palmmicro.com/woody/ >>> >>> ----- Original Message ----- >>> From: "Philipp Klaus Krause" <p...@spth.de> >>> To: <sdcc-user@lists.sourceforge.net> >>> Sent: Tuesday, July 12, 2011 11:30 PM >>> Subject: Re: [Sdcc-user] Z80 code size reduction >>> >>> >>>> Am 12.07.2011 15:38, schrieb Lin Rongrong: >>>>> After another half hour compile with --no-peep option, I sadly found >>>>> the >>>>> same problem. Seems that the peephole optimizer is not the source of >>>>> error. >>>>> >>>>> Woody >>>>> >>>> >>>> While I have no idea what kind of program results in such long compile >>>> times using the defaults setting, you might try setting >>>> --max-allocs-per-node 1000 or so during bug-hunting to save time. >>>> >>>> Sometimes it helps even doing a kind of bisect on the >>>> --max-allocs-per-node parameter, since there are bugas that will not >>>> occour at low values, and when a bug appears e.g. at 102 and above , >>>> but >>>> not at 101 and below, diffing the asm output using 102 vs. 101 can help >>>> finding the exact location in the code that is affected by the bug. >>>> >>>> Philipp >>>> >>>> ------------------------------------------------------------------------------ >>>> All of the data generated in your IT infrastructure is seriously >>>> valuable. >>>> Why? It contains a definitive record of application performance, >>>> security >>>> threats, fraudulent activity, and more. Splunk takes this data and >>>> makes >>>> sense of it. IT sense. And common sense. >>>> http://p.sf.net/sfu/splunk-d2d-c2 >>>> _______________________________________________ >>>> Sdcc-user mailing list >>>> Sdcc-user@lists.sourceforge.net >>>> https://lists.sourceforge.net/lists/listinfo/sdcc-user >>> >> > ------------------------------------------------------------------------------ AppSumo Presents a FREE Video for the SourceForge Community by Eric Ries, the creator of the Lean Startup Methodology on "Lean Startup Secrets Revealed." This video shows you how to validate your ideas, optimize your ideas and identify your business strategy. http://p.sf.net/sfu/appsumosfdev2dev _______________________________________________ Sdcc-user mailing list Sdcc-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sdcc-user
