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

Groups > comp.programming.threads > #2103

I correct please read again...

Cross-posted to 2 groups.

Show all headers | View raw

I correct please read again...

Hello,


I have previously given you a benchmark that uses a small reader and 
writer sections, but if you make the reader and writer sections a little 
bit bigger the cache-lines transfers time will become less and less 
bigger in the Amdahl equation and it will make all my RWLock
variants scale much better.

For exemple i have made the read and write sections a little bit bigger 
and the execution time inside the read and write sections was equal
and i have taken an average number of cache-lines tranfers and the
instructions inside the reader section was scalable, and it has given 
the following calculation results:


On a computer with 32 cores LW_RWLockX scales up to 29X

On a computer with 32 cores LW_RWLock scales up to 29X


And on a computer with 32 cores my RWLock and my RWLockX scales up to 9X.


So this is much better.

So i advise you to use LW_RWLockX and LW_RWLock that scales better.



You can download all the variant of my scalable RWLock 3.05 from:


http://pages.videotron.com/aminer/



Thank you,
Amine Moulay Ramdane.

Back to comp.programming.threads | Previous | Next — Next in thread | Find similar

Thread

I correct please read again... aminer <aminer@toto.net> - 2014-03-18 15:02 -0700
  Re: I correct please read again... aminer <aminer@toto.net> - 2014-03-18 15:08 -0700
  Re: I correct please read again... aminer <aminer@toto.net> - 2014-03-18 15:16 -0700

csiph-web