About lockfree queues

[aminer]

Hello,


I have come to an interresting subject about lockfree queues,
if you take a look at those lockfree queues, like in transactional
memory, if the data have changed those lockfree queue retries and loop
again, but this (lockfree) mechanism generates a lot of cache-coherence
traffic, so i will not advice this lockfree mechanism, so how
can we do it ? you can take a look at the MCS queue Lock , i
think they are doing it like a waitfree linklist that doesn't spin and
that reduces a lot the cache-coherence traffic, other than that
if your memory manager is not optimal and uses a lockfree mechanism and
generates a lot cache-coherence traffic, so you have to use a freelist
to lower the cache coherence traffic.


So be smart please and think about that.


Thank youi,
Amine Moulay Ramdane.

[Chris M. Thomasson]

> "aminer" wrote in message news:l1sh13$krl$1...@news.albasani.net... I have

> come to an interresting subject about lockfree queues,

> if you take a look at those lockfree queues[...]

This MPMC queue one scales very good for a damn anti-pattern:

I think I will post the algorithm that uses eventcounts to get around
spin-wait.

The code needs Relacy to run.


Some context:


https://groups.google.com/forum/#!searchin/comp.arch/random$20isa/comp.arch/aLVoxdQdRac/overview
(read all)

https://groups.google.com/forum/#!original/comp.arch/aLVoxdQdRac/9DtioMswSmUJ
(pseudo-code for spin-based algorithm)