Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.programming > #4612 > unrolled thread

Read again...

Back to article view | Back to comp.programming

Contents

  Read again... Ramine <ramine@1.1> - 2014-06-14 17:25 -0700

#4612 — Read again...

Hello,

I think we are safe with my scalable RWLocks...

Cause i have forgot one thing, if you are distributing the access
to the memory using multiple memory buses, like is doing NUMA systems,
so when in this configuration of distributed memory each threads will 
access the memory in its correcponding memory bus to its corresponding 
memory, this transfers from the memory to the cores is  much expensive 
than the cache-lines transfers from core to core of my writer side of my 
algorithm, hence this will make the parallel part of the Amdahl equation 
much bigger than the serial part, so  my scalable RWLocks will scale 
very well, so i think my scalable RWLocks are still useful
and good and scalable.

So as you have noticed i am a little bit smart now, or am i smart ?

A smile :)

So be happy with my scalable RWLocks.

You can download my scalable RWLocks from here:

https://sites.google.com/site/aminer68/scalable-rwlock

Thank you,
Amine Moulay Moulay.

[toc] | [standalone]

Back to top | Article view | comp.programming

csiph-web