[lnkForumImage]
TotalShareware - Download Free Software
Confronta i prezzi di migliaia di prodotti.
Asp Forum
 Home | Login | Register | Search 


 

Forums >

comp.programming

Please reread...

Ramine

3/18/2015 12:10:00 AM

Hello..


Finally i think i have understood something very important...

Because i have just read more PhD papers about concurrent
datastructures, and i have finally understood something very important,
what you have seen me doing in my previous post is trying to
use my scalable SeqlockX in a high level manner, but that's not the way
it works and that's not the way it works with SMR or RCU or with
Transactional memory, with transactional memory you can not just put
brackets of the transactions around the transactions of the insert() and
delete() and search() methods of the AVL tree or Red-Black tree or
Skiplists to make them concurrent, that's not the way it works, you have
to reason from inside the datastructures algorithms and use the
transactions in a finer grained manner, so this means that my scalable
SeqlockX algorithm is really a powerful algorithm, but how can you use
scalable SeqlockX to design a scalable AVL tree or scalable
Red-Black tree or a scalable Skiplists on read-mostly workloads ?
that's the same way as RCU is used or SMR is used or Transactional
memory is used, you can not wrap the reader section and writer section
with a RLock() and RUnlock() and Wlock() Wunlock(), instead
you have to do it a finer grained manner from inside the algorithms
around the memory parts like pointers or arrays or structs , that's the
way it works with my scalable SeqlockX and that's the way it works with
SMR and RCU and transactional memory. So finally my scalable SeqlockX is
really a powerful mechanism that is smart and useful, so finally i can
also say that my scalable SeqlockX can replace RCU and replace SMR.


You can download my scalable SeqlockX from:

https://sites.google.com/site/aminer68/scalabl...


Thank you,
Amine Moulay Ramdane.