DragonFly users List (threaded) for 2008-01
I may be starting a flamewar with this question, so I apologize in advance.
Some time ago there was a discussion about the status of SMP in Dfly.
As far as I can remember, the developers behind Dfly said that the SMP
implementation is currently not on their schedule, even though
progress was made in implementing various part of the system so that
it's not running under the big lock.
Then the news came out that FreeBSD completed their SMP implementation
as of version 7.0, and some benchmarks were performed showing it
beating Linux SMP. Then there was a discussion again on this list
hinting that Dfly would simply import the FreeBSD SMP implementation.
So my question is - has a decision been made on the SMP development?
In other words - is it still on the list of tasks to do, will it be
native Dfly implemenation, or will it be a port of FreeBSD's
IF SMP is no longer pursued in Dfly b/c of limited developer time,
then I'd like to get an estimate if possible on how much work remains
to be done to get SMP fully functional. It doesn't matter if it would
be a port from another *BSD or custom Dfly implementation. I just need
a rough estimate in terms of man-hours. If you can give an estimate
for each option and the preferred option then it's even better.
The reason I am asking is that I am a co-founder of several startups
and we plan on becoming exclusively Dfly-centric. We already use Dfly
extensively as a DB platform (both MySQL and recently switching to
PostgreSQL given that Sun bought the former). The HAMMER filesystem
will be a major feature that we will use extensively, but SMP is also
high on the needs list. Without going into too much detail, one of the
projects we are working is a very large scale distributed search
engine. You can see why HAMMER and SMP are important to us.
So, we may be able to sponsor some/all work on getting SMP implemented
in Dfly if it's a feature that Dfly still wants to implement.
Anyways, let me know. Thanks in advance.