[Gc] Re: Problems with GC_size_map
ludo at gnu.org
Tue Feb 2 09:40:37 PST 2010
Juan Jose Garcia-Ripoll
<juanjose.garciaripoll at googlemail.com> writes:
> On Tue, Feb 2, 2010 at 1:37 AM, Bruce Hoult <bruce at hoult.org> wrote:
>> On Tue, Feb 2, 2010 at 12:51 PM, Juan Jose Garcia-Ripoll
>> <juanjose.garciaripoll at googlemail.com> wrote:
>> > In general I am worried that the garbage collector is performing rather
>> > for a simple problem in which one builds and quickly discarding bignums
>> > pointerless memory, by the way --. Profiling shows that 80% of the time
>> > spent in the mark phase.
>> You are not allocating the bodies of the bignums using GC_malloc_atomic()?
> Of course I am. The whole bignum is a huge chunk of atomic memory, including
> a few words for the structure of the datatype and the rest full of bits that
> represent the bignum, using GMP. Other people have verified this behavior,
> not only in OS X but also in Linux. I have tried both with and without
> incremental garbage collection.
If you have a simple test case, I’d be happy to adapt it and test it on
Guile since have a similar setup, using GMP as well.
More information about the Gc