Prev: Fastest DFA recognizer
Next: A A
From: Ben Bacarisse on 21 May 2010 07:40
Srinu <sinu.nayak2001(a)gmail.com> writes:

> You seems to be correct. I very much appreciate your polite reply. I
> am sorry incase I have unknowingly hurt you in any manner.

No, I am not in any way hurt. I'm sorry if you got that idea.

> Truly speaking, this question was bit general. The algorithm I have
> written witeen for the last four type of systems seems to be correct
> in very strict sence. But are not efficient/suitable in general for
> all these types. So, if we consider each of them again, we may have
> some ideas of improvement on top of the basic algorithm. I was looking
> for those improvement ideas, for example, giving writer a high
> priority, allowing mutiple reads simultaneously etc. If possible also
> small psudo codes as given by me with simple locking mechanism, that
> will be very simple and basic yet elegant. Not about highlevel
> primitives, (counting) semaphores, condition variables, monitors,
> rendezvous etc.

The "standard" solutions (Coutrois, Heymans and Parnas, "Concurrent
Control with 'Readers' and 'Writers'", CACM 1971, vol 14 no 10) allow
multiple concurrent readers and need only mutex locks. Two versions are
presented: one that favours readers and one that favours writers, but
both permit starvation.

Other solutions exist but I can't site them at the moment (too lazy to
go and look them up).

<snip>
--
Ben.
From: Srinu on 21 May 2010 10:34

> No, I am not in any way hurt.  I'm sorry if you got that idea.

It is my pleasure.

> The "standard" solutions (Coutrois, Heymans and Parnas, "Concurrent
> Control with 'Readers' and 'Writers'", CACM 1971, vol 14 no 10) allow
> multiple concurrent readers and need only mutex locks.  Two versions are
> presented: one that favours readers and one that favours writers, but
> both permit starvation.

Many thanks for your answer. I'll try to find the paper and read.
It will be great if you could please let me know some books where
these topics along with concurrent programming has been presented in
details.

Sincerely,
Srinivas Nayak
First  |  Prev  | 
Pages: 1 2
Prev: Fastest DFA recognizer
Next: A A