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

Groups > comp.programming.threads > #1770

About another subject...

Cross-posted to 2 groups.

Show all headers | View raw

Hello again,

I have implemented another Lock that is fair using an algorithm that 
look like my SemaCondvar algorithm, but there is still a problem with 
it, if you take a look
at my SemaCondvar algorithm it uses a critical section, but this 
critical section defeat the purpose of fairness cause it uses a 
spin-wait mechanism that introduce the starvation problem, but
if i use a Semaphore instead of a Windows Critical Section the
context switches will slow a lot the algorithm, other than that
i am using three locked regions in my algorithm that uses
the same Lock, so this will higher the delay in the backoff mechanism
of the Windows critical section or the Spinlock with an exponential
backoff , so this will slow a lot the algorithm, hence those problems 
have defeated the purpose of fairness of my algorithm and they will make
my algorithm slow.

But my Scalable Distributed Lock is good and i advice you to use it 
cause it is suited even for bigger locked regions, i think the Windows 
critical section is not suited for bigger locked regions and i have 
explained
to you why.



Thank you,
Amine Moulay Ramdane,.

Back to comp.programming.threads | Previous | Next | Find similar

Thread

About another subject... aminer <aminer@toto.net> - 2013-10-03 14:29 -0700

csiph-web