R8B PRE-RELEASES
Kostis Sagonas
kostis@REDACTED
Wed Oct 10 12:03:12 CEST 2001
Per Bergqvist asked:
> Any plans to add this support ?
Yes. The plan is to provide this support "really soon now",
meaning that we (the HiPE group) plan to do this probably
before the end of the year.
At this point, i.e., before the R8 release, we are concentrating
our efforts in making sure that HiPE is efficient and gets released
as bug-free as possible.
> Unified heap and hipe are both two very good things that give
> great performance boosts and I would hate to have to choose ...
Thanks!
Here is a more detailed version of our plan:
As mentioned, we have quite good reasons to believe that the
unified heap architecture is a better memory architecture than
the current one. However, we would like to be sure of this
before committing to one of them (supporting both of them is
a nightmare we would very much like to avoid).
We urge the Erlang community to test *applications* on the
unified heap architecture and report its findings either to us
or directly in this list as e.g. Pascal Brisset did yesterday;
(Pascal we were excited when we saw your mail - merci!).
In the interest of the community, it is important to try to get
data that show both the pros and cons of the two architectures.
So please, if you have an interesting application that you have
reasons to believe it will behave better in the one or the other
architecture, bench it and preferably send its _kernel_ to us.
Once we are finished with this benchmarking, we will choose one
of the architectures as the supported one and we will make sure
that HiPE runs on it. This will happen before R9.
In the meantime, the choice (for me) is obvious:
- configure with --enable-hipe and use this system for everyday use;
- if you suspect that a specific program behaves better (or worse)
with unified heap, or you are simply curious, try this architecture
and if your findings are interesting, please do send them to us.
Kostis Sagonas (for the HiPE group).
More information about the erlang-questions
mailing list