I wrote: > Could you try running this through cachegrind, both with and without > the list+pushmark optimisation, and perhaps with different numbers > of lexicals? > > I disabled the list+pushmark optimisation on the > sprout/nolistpushmarkopt branch to facilitate this. BTW, a simplistic benchmark like this shows a 10% speed increase with the list+pushmark optimisation disabled: $ time ./miniperl -e 'for(1..20000000){() = (0,my($b,$c))}'Thread Previous | Thread Next