Compare-and-swap semantics in GHC

Carter Schonwald carter.schonwald at
Thu Dec 5 08:52:10 UTC 2013

hrm, i'm reading your mesage a bit more closely, and i'm not sure i
understand the distinction you mean by strong vs weak.

Do you mean "strong" as in STM style semantics for reads and writes? (ie if
i'm doing a CAS on memory location x, it totally orders all reads and
writes to ANY location y!=x too?).

if you mean something like that, I think CAS isn't strong, though perhaps
the best way to answer you question is to read the source for CAS and look
up the semantics of the assembly instructions on various architectures!

cheers (and apologies for the confusion on my part)

On Thu, Dec 5, 2013 at 3:48 AM, Carter Schonwald <carter.schonwald at
> wrote:

> cas in ghc has sequential semantics.
> i'm not sure if every use of it needs that semantics, if you identify
> examples where weaker operations may suffice, please share!
> On Thu, Dec 5, 2013 at 2:58 AM, KwangYul Seo <kwangyul.seo at>wrote:
>> Hi,
>> java.util.concurrent.atomic package provides two flavors of compare and
>> set operations: compareAndSet and weakCompareAndSet. The latter does not
>> create any happens-before orderings, so we can use it where no guarantees
>> with respect to previous or subsequent reads and writes of any variables
>> other than the target of the weakCompareAndSet are required.
>> I'd like to ask if the compare-and-swap function provided by GHC runtime
>> (cas() in includes/stg/SMP.h) is strong or weak. If it is strong, do all
>> use cases of cas() require this semantics?
>> Regards,
>> Kwang Yul Seo
>> _______________________________________________
>> ghc-devs mailing list
>> ghc-devs at
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the ghc-devs mailing list