Parallel performance drops off a cliff

[Simon Marlow]

2009/4/20 Neil Mitchell <ndmitchell at gmail.com>
>
> Hi,
>
> Using one benchmark I have, which doesn't create any threads, I have:
>
> $ benchmark +RTS -Nx
>
> x             time (Seconds)
> 1              2
> 2              2
> 3              2
> 4              3
> 5              3
> 6              3
> 7              3
> 8              aborted after 2 minutes
>
> This is using GHC 6.10.2 on Windows XP, 2 processors. Is this a known
> bug, or should I try and replicate it? (benchmark is fairly big and
> very dependent on internal things, but I suspect the dramatic
> performance slowdown is unlikely to be related to these bits).

Yes, what's happening is this: GHC 6.10.2 contains some slightly bogus
heuristics about when to turn on the parallel GC, and it just so
happens that 8 processors tips it over the point where the parallel GC
is enabled for young-generation collections.  In 6.10.2 the parallel
GC really didn't help most of the time, but it has undergone a lot of
tuning since then, and in the HEAD things are much better (see the
results from our ICFP submission).

In the meantime you might get somewhere by disabling parallel GC
altogether (+RTS -g1), but as the results in our paper show, sometimes
the parallel GC is essential for retaining locality in parallel
programs.

Cheers,
   Simon