[Haskell-cafe] Re: [16/16] SBM: Discussion and Conclusion

[Don Stewart]

simonmarhaskell:
> Peter Firefly Brodersen Lund wrote:
> 
> >Using top together with huge input files convinced me that -sstderr was
> >untrustworthy so I developed the pause-at-end preloading hack.  Paranoia 
> >paid
> >off big time there.
> 
> In what way did you find -sstderr untrustworthy?  Perhaps it is because the 
> "memory in use" figure only counts memory that the RTS knows about, and 
> doesn't include malloc()'d memory.  If it's anything else, I need to know 
> about it.
> 
> > o can the core/stg code be made easier to read?
> 
> By all means - submit tickets with specific suggestions, or better still 
> send patches.

It's also easier to read if you pipe it into hscolour. A little trick
I use to get dynamic syntax highlighting of core:


    ghc A.hs -O2 -ddump-simpl --make | HsColour -tty | less -r

will give you coloured, scrollable syntax highlighted core.

-- Don