Hello, 

I want to clear something about  my Distributed sequential lock algorithm, 
you can download it and read about it here: 

https://sites.google.com/site/aminer68/scalable-distributed-sequential-lock 


If you read carefully the source code you will noticed that i am using 
this: "myid3:=FCount6^.fcount6;" inside the RLock() 
method in the reader side, so when  my algorithms switches from a Seqlock 
mode to a Distributed lock mode, the 
"myid3:=FCount6^.fcount6;" will be "serialized" when the writer side 
increment FCount6^.fcount6, so in the writer side i am calling the 
distributed reader-writer lock when the algorithm swithes to a Distributed 
mode, so if you have noticed since the "myid3:=FCount6^.fcount6;" is 
serialized in the reader side when "FCount6^.fcount6" is incremented in the 
writer side, 
so this will look like a probalistic mechanism cause the reader side is 
serializing the the transfer of FCount6^.fcount6 
on the bus, and the writer side is serializing the transfer of a variable 
in the bus also when it is calling the WLock() of the 
distributed lock, so this is a kind of a probalistic mechanism cause this 
can generate contention and the reader side can call 
RLock() on a distributed lock's  reader-writer mutex that is entered before 
the reader by the writer, but even though it is a probalistic mechanism, 
this probabilistic mechanism  will give a good result and it eliminates the 
"livelock" situation of the Seqlock when there is more writers. 




Thank you for your time. 


Amine Moulay Ramdane. 

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"Scalable Synchronization Algorithms" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/lock-free/95669533-eb4b-42d6-af93-c1c7b5d44f98%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to