Subj : Re: Challenge: Multithreading & Synchronization To : comp.programming.threads From : doug Date : Sat May 21 2005 01:27 pm "doug" wrote in message news:wqFje.29083$2k2.8043@fe1.news.blueyonder.co.uk... > > "Uenal Mutlu" <520001085531-0001@t-online.de> wrote in message > news:d6n5bd$31f$04$1@news.t-online.com... >> "doug" wrote >>> >>> Uenal, you keep coming up with new magic to make your code work. Do as >>> David asks - present a *complete* implementation of a recursive mutex >>> that >>> is not slower than a non-recursive one. (A win32 CriticalSection is not >>> an >> >> The statement is not correct. What I said is that my implementation >> of mutex (both recurs and non-recurs variants) is at least 2x faster >> than the standard method of Windows (ie. the CriticalSection). >> >>> example - it's recursive to begin with.) No magic, please - the full >>> implementation. >> >> I have neither the intent to be a free code deliverer for you, nor am I >> going >> to disclose to you the internals of my code. >> My code works. If you doubt it then I can't help you. I can give you only >> compiled demo code (lib for MSVC++) for testing. Contact via email if >> interessted and specify what you would like to test, describe as much as >> possible. >> >> > > Here we go again. > > All I wanted you to do was prove your assertion that a recursive mutex can > be as fast as a non-recursive one in the 'initial uncontested case'. > > You can't and/or won't do this. That's fine. I just wanted to make sure > we're on the same page. > > As for being a free code deliverer - don't worry. I don't think anyone > here will be touching your stuff anyway. > Peter and Giancarlo have actually done this, just a short time ago in this thread. So, I'll admit I was wrong in thinking it couldn't be done! Ta to them for proving me wrong. Right, I'm off to watch a footy game in this glorious English weather! Doug .