STM with multiverse

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

STM with multiverse

Johannes Link
Hi,

As promised a while ago I've just tried to add a few demos to GPars STM support. Since some kind of collections will be needed for most non-trivial examples, I started by using TxnList and TxnSet just to realize that they will throw a TodoException for most useful cases, e.g. when trying to iterate through the elements. As far as I can see, this has not changed in Multiverse 0.8 SNAPSHOT (which itself has not been touched in over a year).

Therefore, my personal impression is that GPar's current STM support is not even sufficient for my simple demonstration purposes. That's a pity. I guess we should either branch Multiverse or swap it out for a different STM lib - if there is any.

cheers, Johannes
Reply | Threaded
Open this post in threaded view
|

Re: STM with multiverse

Russel Winder-3
On Tue, 2013-07-09 at 15:37 +0200, Johannes Link wrote:

> Hi,
>
> As promised a while ago I've just tried to add a few demos to GPars STM
> support. Since some kind of collections will be needed for most non-trivial
> examples, I started by using TxnList and TxnSet just to realize that they
> will throw a TodoException for most useful cases, e.g. when trying to
> iterate through the elements. As far as I can see, this has not changed in
> Multiverse 0.8 SNAPSHOT (which itself has not been touched in over a year).
>
> Therefore, my personal impression is that GPar's current STM support is not
> even sufficient for my simple demonstration purposes. That's a pity. I
> guess we should either branch Multiverse or swap it out for a different STM
> lib - if there is any.
If Multiverse is not up to the task, then I suspect we are looking for
another library since I doubt there is resource to implement something
specific to GPars.  The usual suspect would appear to be:

        Multiverse
        Clojure
        Akka
        Deuce STM
        AtomJava
        JVSTM
        stmlib

I have no personal experience of any of these; I am not a fan of STM as
an application level tool, it should be an implementing technique for
higher level architectures.

Václav had reasons for choosing Multiverse, but I am not sure what they
were just now.


--
Russel.
=============================================================================
Dr Russel Winder      t: +44 20 7585 2200   voip: sip:[hidden email]
41 Buckmaster Road    m: +44 7770 465 077   xmpp: [hidden email]
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc (205 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: STM with multiverse

Vaclav
Administrator
Thank you, Johannes, for the analysis. Although Peter is reading this list, I'm adding him in cc explicitly, in case he has a comment.

Cheers,

Vaclav




On Tue, Jul 9, 2013 at 4:31 PM, Russel Winder <[hidden email]> wrote:
On Tue, 2013-07-09 at 15:37 +0200, Johannes Link wrote:
> Hi,
>
> As promised a while ago I've just tried to add a few demos to GPars STM
> support. Since some kind of collections will be needed for most non-trivial
> examples, I started by using TxnList and TxnSet just to realize that they
> will throw a TodoException for most useful cases, e.g. when trying to
> iterate through the elements. As far as I can see, this has not changed in
> Multiverse 0.8 SNAPSHOT (which itself has not been touched in over a year).
>
> Therefore, my personal impression is that GPar's current STM support is not
> even sufficient for my simple demonstration purposes. That's a pity. I
> guess we should either branch Multiverse or swap it out for a different STM
> lib - if there is any.

If Multiverse is not up to the task, then I suspect we are looking for
another library since I doubt there is resource to implement something
specific to GPars.  The usual suspect would appear to be:

        Multiverse
        Clojure
        Akka
        Deuce STM
        AtomJava
        JVSTM
        stmlib

I have no personal experience of any of these; I am not a fan of STM as
an application level tool, it should be an implementing technique for
higher level architectures.

Václav had reasons for choosing Multiverse, but I am not sure what they
were just now.


--
Russel.
=============================================================================
Dr Russel Winder      t: <a href="tel:%2B44%2020%207585%202200" value="+442075852200">+44 20 7585 2200   voip: [hidden email]
41 Buckmaster Road    m: <a href="tel:%2B44%207770%20465%20077" value="+447770465077">+44 7770 465 077   xmpp: [hidden email]
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder



--
E-mail: [hidden email]
Blog: http://www.jroller.com/vaclav
Linkedin page: http://www.linkedin.com/in/vaclavpech
Reply | Threaded
Open this post in threaded view
|

Re: STM with multiverse

Eric MacAdie
In reply to this post by Russel Winder-3
Adding Akka or Clojure to GPars to get STM seems like overkill.

- Eric MacAdie

On Tue, Jul 9, 2013 at 9:31 AM, Russel Winder <[hidden email]> wrote:


If Multiverse is not up to the task, then I suspect we are looking for
another library since I doubt there is resource to implement something
specific to GPars.  The usual suspect would appear to be:

        Multiverse
        Clojure
        Akka
        Deuce STM
        AtomJava
        JVSTM
        stmlib

I have no personal experience of any of these; I am not a fan of STM as
an application level tool, it should be an implementing technique for
higher level architectures.

Václav had reasons for choosing Multiverse, but I am not sure what they
were just now.


--
Russel.
=============================================================================
Dr Russel Winder      t: <a href="tel:%2B44%2020%207585%202200" value="+442075852200">+44 20 7585 2200   voip: [hidden email]
41 Buckmaster Road    m: <a href="tel:%2B44%207770%20465%20077" value="+447770465077">+44 7770 465 077   xmpp: [hidden email]
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder