Seems fine, here. Thank you. On Dec 13, 2016 10:44 AM, "Juergen Sauermann" <juergen.sauerm...@t-online.de> wrote:
> Hi, > > I have reverted the change between *SVN 818 *and *819* and I hope > everything is back > to normal in *SVN 820*. > > /// Jürgen > > > On 12/13/2016 10:10 AM, Juergen Sauermann wrote: > > Hi, > > I see. This was supposed to fix some other problem, but apparently didn't > work. > > I will look into this; in the meantime please fall back to r818 (*svn up > -r818*). > > /// Jürgen > > > On 12/13/2016 06:14 AM, David B. Lamkins wrote: > > Revision 819 causes *severe* performance issues for my APL Packager. > > https://github.com/TieDyedDevil/apl-pkg > (Use the `percy` branch.) > > $ git clone https://github.com/TieDyedDevil/apl-pkg > # cd apl-pkg > $ git branch percy > $ ./install.sh > $ awe > > Normally the system comes up in a second or so; under r819 it takes tens > of minutes to complete, during which one CPU core is pegged. (It's slow > enough that I first thought that some altered edge condition must've caused > an infinite loop.) > > Memory usage is stable. > > FWIW, this happens in the packager's boot loader, so there's not a lot of > APL code involved. > > https://github.com/TieDyedDevil/apl-pkg/blob/percy/boot/gnu-apl-linux.apl > > Once past the boot loader (which took about 90 minutes on my Haswell i7 > notebook), virtually every packager operation is noticeably slow. If I had > to guess, I'd suspect that either branching or APL function calls are > eating a lot of cycles unnecessarily; my code uses lots of both. > > > > > >
