comp.programming

Ramine

3/17/2015 8:30:00 PM

Hello,


This scalability prediction was done for read-mostly workloads.


Thank you,
Amine Moulay Ramdane.


On 3/17/2015 11:52 AM, Ramine wrote:
>
> Hello,
>
>
> I have done a scalability prediction for a concurrent AVL tree that uses
> the following Distributed reader-writer mutex:
>
> https://sites.google.com/site/aminer68/scalable-distributed-reader-wr...
>
>
> Since the GetCurrentProcessorNumber() function that is used
> by the distributed reader-writer mutex is parallelized on multicores,
> and since the LockedExchangeAdd() of my LW_RWlockX that is used
> by the distributed reader-writer mutex is taking few CPU cycles,
> so i have done a scalability prediction for a concurrent AVL tree that
> uses keys as strings using this distributed reader-writer mutex and it
> has given 50X speedup on multicores, you will get 50x if the data
> corresponding to each key is small, but if the data corresponding to
> each key is bigger you will get much more speedup than 50X on multicores.
>
>
>
>
> Thank you,
> Amine Moulay Ramdane.
>
>
>