Comment 4 for bug 727615
Revision history for this message
| Roman Marynchak (roman-marynchak) wrote | #4 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
1b1ed1a
(Get the code!)
Well, the idea itself is not so bad. Some guys have already implemented it: http://
In case there is a possibility to do something like this without serious performance issues... It is a nice feature. It should be very helpful when you have a ton of the legacy code which should be adapted to use SMP. I work with one CL system which has around 1M LOC, and I have seen a lot of very complex synchronization issues. Synchronized data structures allow better SMP problems analysis, and it is critical for large systems and large teams (teams which do not have he-knows-it-all-guy who can fix any SMP-related system bug in a week). When you have 50 hashtables and 5-10 working threads which access them with quite complex logic behind the scene, every smallest compiler feature towards SMP makes your life significantly easier. At least, it may help to debug the problem (Though, I don't say that it solves the issues magically).
Regarding the real-life example, imagine a hash table which holds a structure called 'relation' as a value, and a relation name as a key. The relation is something like '(A RELATED-TO B), and a name is a symbol. There are 10 rules, each of them running in its own thread, which are interested in this relation. Two of them can change it, other eight can only read it. All rules are launched every 1 second, they execute during 50ms and then sleep the rest time waiting for the next launch. In case hash-table is thread-safe, this stuff works perfectly, just because the maximum action delay upon a _final_ relation change is 1 second, which is not a problem (relation changes are allowed to overwrite one another without any action). But in case a table is not thread-safe, and a relation value becomes garbage, we are likely to end up in the debugger.
I can tell you about more examples, but anyway I don't push this feature very hard, I just think that it is _quite reasonable_.
In case we fight for the ultimate performance, just make this feature to be a build option. But don't throw it out all at once.