latest inline times

Stephen Weeks sweeks@research.nj.nec.com
Tue, 23 Feb 1999 17:02:35 -0500


> Times taken on Mitas.  (threshold 100 -- rounds 3)
> ---------------------
> 
>                          time       ratio with current inliner (SmallOnly/Recursive)
>                        -------      -------------------------------------------------
>             fft           41.9                  .92
>      barnes-hut           17.9                  .72
> matrix-multiply           13.5                  .98
>    knuth-bendix           38.5                  .80
>          lexgen           33.4                 1.40
>    count-graphs           14.2                  .87
>            life           64.3                  .88
>           logic           59.5                  .83
>      mandelbrot           15.9                 1.02
>          mlyacc           23.3                 1.00
>         nucleic           33.7                  .63
>          simple           17.4                  .75
>             tsp           34.0                  .92

I.E. only on lexgen does recursive inlining win?