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

Groups > comp.programming.threads > #2773 > unrolled thread

I think i have made a mistake

Back to article view | Back to comp.programming.threads

Contents

  I think i have made a mistake Ramine <ramine@1.1> - 2014-12-08 17:25 -0800
    Re: I think i have made a mistake "Chris M. Thomasson" <no@spam.invalid> - 2014-12-10 21:12 -0800

#2773 — I think i have made a mistake

Hello,


I think i have made a mistake about the follwing reader-writer lock of
Joe Duffy:

http://joeduffyblog.com/2009/02/20/a-more-scalable-readerwriter-lock-and-a-bit-less-harsh-consideration-of-the-idea/ 


Since a cache-line transfer between cores is not at all expensive,
so the serial part of the Amdahl's law for this reader-writer
above will still be smaller  with more and more cores,
so this will make this reader-writer lock above of Joe Duffy
really scalable on read-mostly scenarios.



Thank you,
Amine Moulay Ramdane.

[toc] | [next] | [standalone]

#2775

> "Ramine"  wrote in message news:m658e8$vpp$2@dont-email.me...

> Hello,


> I think i have made a mistake about the follwing reader-writer lock of
> Joe Duffy:

> http://joeduffyblog.com/2009/02/20/a-more-scalable-readerwriter-lock-and-a-bit-less-harsh-consideration-of-the-idea/

No, it is a bit of garbage. All of the readers have to spin on that global
writers counter. Yuck!

;^/ 

[toc] | [prev] | [standalone]

Back to top | Article view | comp.programming.threads

csiph-web