Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
My explanation...
17 views
Skip to first unread message
Ramine
Oct 20, 2015, 7:50:14 PM10/20/15
to
Hello....
As you have noticed i have wrote my new algorithm versoin 1.31
of my scalable distributed reader-writer mutex, and just after that i
have made it sequential consistent, but you have to understand ladies
and gentlemen that i have implemented my new algorithm in Object Pascal
and i have ported it to FreePascal and Delphi, and they both (Delphi and
FreePascal compilers) respect the strong memory model of the x86
architecture, so it was much easier for me to make it sequential
consistent because you have only to take care of the following
non-guarantee listed in the x86 spec:
Loads may be reordered with older stores to different locations.
So i have made my new algorithm version 1.31 of my scalable distributed
reader-writer mutex sequential consistent with easy.
That's not the case for C++, C++ have made it difficult for sequential
consistency, because it respect the weak memory model that is much
difficult to reason about and that is prone to mistakes by programmers,
so becareful of C++ and languages that respect the weak memory model.
And also my new algorithm of a scalable reader-writer mutex version 1.31
take care of false-sharing and it is now sequential consistent and like
in Seqlock or RCU , my new scalable distributed reader-writer mutex
doesn't use any atomic operations and/or StoreLoad style memory barriers
on the reader side, so it's very fast and scalable..but you have to use
the define's option TLW_RWLockX or the define's option TRWLockX inside
the defines1.inc file for that.
You can download it from here:
https://sites.google.com/site/aminer68/scalable-distributed-reader-writer-mutex
Thank you,
Amine Moulay Ramdane.
As you have noticed i have wrote my new algorithm versoin 1.31
of my scalable distributed reader-writer mutex, and just after that i
have made it sequential consistent, but you have to understand ladies
and gentlemen that i have implemented my new algorithm in Object Pascal
and i have ported it to FreePascal and Delphi, and they both (Delphi and
FreePascal compilers) respect the strong memory model of the x86
architecture, so it was much easier for me to make it sequential
consistent because you have only to take care of the following
non-guarantee listed in the x86 spec:
Loads may be reordered with older stores to different locations.
So i have made my new algorithm version 1.31 of my scalable distributed
reader-writer mutex sequential consistent with easy.
That's not the case for C++, C++ have made it difficult for sequential
consistency, because it respect the weak memory model that is much
difficult to reason about and that is prone to mistakes by programmers,
so becareful of C++ and languages that respect the weak memory model.
And also my new algorithm of a scalable reader-writer mutex version 1.31
take care of false-sharing and it is now sequential consistent and like
in Seqlock or RCU , my new scalable distributed reader-writer mutex
doesn't use any atomic operations and/or StoreLoad style memory barriers
on the reader side, so it's very fast and scalable..but you have to use
the define's option TLW_RWLockX or the define's option TRWLockX inside
the defines1.inc file for that.
You can download it from here:
https://sites.google.com/site/aminer68/scalable-distributed-reader-writer-mutex
Thank you,
Amine Moulay Ramdane.
Christopher Pisz
Oct 20, 2015, 7:51:10 PM10/20/15
to
On 10/20/2015 9:52 PM, Ramine wrote:
> Hello....
>
>
> As you have noticed i have wrote my new algorithm versoin 1.31
One more spammy post and you're going on my blacklist.
> Hello....
>
>
> As you have noticed i have wrote my new algorithm versoin 1.31
You only need one post for one topic
--
I have chosen to troll filter/ignore all subthreads containing the
words: "Rick C. Hodgins", "Flibble", and "Islam"
So, I won't be able to see or respond to any such messages
---
0 new messages