[gambit-list] Performance observations

Christian christian at pflanze.mine.nu
Fri Apr 28 20:36:01 EDT 2006


At 18:19 Uhr +0200 28.04.2006, I wrote:
>.. But it looks like it is only
>about as fast as the BoehmGC as being used by Bigloo.

I concluded this from former tests allocating lists. I've now also 
tried to run my stream benchmarks on Bigloo, and have realised that 
Bigloo (C backend, with -O or -Obench) is about 2 times slower 
processing them than Gambit, and worse, the stream heads are not 
reclaimed while walking the streams (*), so Bigloo runs out of memory 
on the same code which is processed in constant memory by Gambit or 
Chicken. I'm sending a copy of this mail to the Bigloo mailing list: 
please tell me if there's a solution to this problem.

Christian.

(*) with code like: (stream-length (stream-iota 10000000))
or more precisely: see code at
http://scheme.mine.nu/bigloo/div/cheapstream.scm



More information about the Gambit-list mailing list