www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Thin Locks in D

reply Walter Bright <newshound1 digitalmars.com> writes:
http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

and of course on reddit:

http://www.reddit.com/comments/6tbzc/thin_locks_in_d/
Jul 24 2008
next sibling parent reply "Craig Black" <craigblack2 cox.net> writes:
"Walter Bright" <newshound1 digitalmars.com> wrote in message 
news:g6b1ae$up4$1 digitalmars.com...

 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

 and of course on reddit:

 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


Very cool stuff.  So is a future version of D going to use thin locks when 
the synchronize keyword is used?

-Craig
Jul 24 2008
parent reply "Jarrett Billingsley" <kb3ctd2 yahoo.com> writes:
"Craig Black" <craigblack2 cox.net> wrote in message 
news:g6ba04$1pfb$1 digitalmars.com...

 "Walter Bright" <newshound1 digitalmars.com> wrote in message 
 news:g6b1ae$up4$1 digitalmars.com...

 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

 and of course on reddit:

 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


 Very cool stuff.  So is a future version of D going to use thin locks when 
 the synchronize keyword is used?

 -Craig


It's something that's part of the runtime lib.  I think they can be 
implemented in Tango already.  I'm not sure how robust Phobos is in terms of 
monitor flexibility.
Jul 24 2008
next sibling parent superdan <super dan.org> writes:
Jarrett Billingsley Wrote:


 "Craig Black" <craigblack2 cox.net> wrote in message 
 news:g6ba04$1pfb$1 digitalmars.com...

 "Walter Bright" <newshound1 digitalmars.com> wrote in message 
 news:g6b1ae$up4$1 digitalmars.com...

 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

 and of course on reddit:

 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


 Very cool stuff.  So is a future version of D going to use thin locks when 
 the synchronize keyword is used?

 -Craig


 
 It's something that's part of the runtime lib.  I think they can be 
 implemented in Tango already.  I'm not sure how robust Phobos is in terms of 
 monitor flexibility. 


but i thot (hoped) that bartosh works on phobos already?
Jul 24 2008
prev sibling next sibling parent Walter Bright <newshound1 digitalmars.com> writes:
Jarrett Billingsley wrote:

 It's something that's part of the runtime lib.  I think they can be 
 implemented in Tango already.  I'm not sure how robust Phobos is in terms of 
 monitor flexibility. 


The way the locks work is independent of the compiler, it's a pure 
runtime library thing.
Jul 24 2008
prev sibling parent reply Sean Kelly <sean invisibleduck.org> writes:
Jarrett Billingsley wrote:

 "Craig Black" <craigblack2 cox.net> wrote in message 
 news:g6ba04$1pfb$1 digitalmars.com...

 "Walter Bright" <newshound1 digitalmars.com> wrote in message 
 news:g6b1ae$up4$1 digitalmars.com...

 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

 and of course on reddit:

 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


 Very cool stuff.  So is a future version of D going to use thin locks when 
 the synchronize keyword is used?

 -Craig


 
 It's something that's part of the runtime lib.  I think they can be 
 implemented in Tango already.  I'm not sure how robust Phobos is in terms of 
 monitor flexibility. 


 From what I understand, it's being redesigned to work a lot like Tango.


Sean
Jul 24 2008
parent bearophile <bearophileHUGS lycos.com> writes:
Sean Kelly:

  From what I understand, it's being redesigned to work a lot like Tango.


Ah, I see, the start of a convergence :-)
Is Phobos going to gain an actor-based concurrency module, beside the current
threading module?

Bye,
bearophile
Jul 25 2008
prev sibling next sibling parent Graham St Jack <Graham.StJack internode.on.net> writes:
On Thu, 24 Jul 2008 15:56:14 -0700, Walter Bright wrote:


 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/
 
 and of course on reddit:
 
 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


Very nice. I particularly like the part about not needing thread-safe and 
non-thread-safe versions of library code.
Jul 24 2008
prev sibling next sibling parent reply superdan <super dan.org> writes:
Walter Bright Wrote:


 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/
 
 and of course on reddit:
 
 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


very cool stuff that's gonna cross the t in threads for d. i'm so very glad
bartosz is on team. would he join the newsgroup, or is he a mimosa as well?
Jul 24 2008
parent reply Walter Bright <newshound1 digitalmars.com> writes:
superdan wrote:

 Walter Bright Wrote:
 

 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/
 
 and of course on reddit:
 
 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


 
 very cool stuff that's gonna cross the t in threads for d. i'm so
 very glad bartosz is on team. would he join the newsgroup, or is he a
 mimosa as well?


We are very lucky to have Bartosz working with us. He's an expert on 
multithreaded programming, and even better, he spends a lot of time 
researching the latest in what others are doing so that D will build on 
best practices rather than reinvent mistakes.
Jul 24 2008
next sibling parent reply JAnderson <ask me.com> writes:
Walter Bright wrote:

 superdan wrote:

 Walter Bright Wrote:


 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

 and of course on reddit:

 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


 very cool stuff that's gonna cross the t in threads for d. i'm so
 very glad bartosz is on team. would he join the newsgroup, or is he a
 mimosa as well?


 
 We are very lucky to have Bartosz working with us. He's an expert on 
 multithreaded programming, and even better, he spends a lot of time 
 researching the latest in what others are doing so that D will build on 
 best practices rather than reinvent mistakes.


That's really cool.  How long has Bartosz been working with you?  You've 
always said that multithreading is not your strongest point.  Its great 
to have someone on your team with that sort of experience.

-Joel
Jul 25 2008
parent reply superdan <super dan.org> writes:
JAnderson Wrote:


 Walter Bright wrote:

 superdan wrote:

 Walter Bright Wrote:


 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

 and of course on reddit:

 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


 very cool stuff that's gonna cross the t in threads for d. i'm so
 very glad bartosz is on team. would he join the newsgroup, or is he a
 mimosa as well?


 
 We are very lucky to have Bartosz working with us. He's an expert on 
 multithreaded programming, and even better, he spends a lot of time 
 researching the latest in what others are doing so that D will build on 
 best practices rather than reinvent mistakes.


 
 That's really cool.  How long has Bartosz been working with you?  You've 
 always said that multithreading is not your strongest point.  Its great 
 to have someone on your team with that sort of experience.
 
 -Joel


judging by his articles andrei also knows his shit. walter seems to have his
ass covered. so walter - any word on a true multiprocessing model for d coming
soon? u understand it's expected to be better than java's or c++'s.
Jul 25 2008
parent Walter Bright <newshound1 digitalmars.com> writes:
superdan wrote:

 judging by his articles andrei also knows his shit. walter seems to
 have his ass covered. so walter - any word on a true multiprocessing
 model for d coming soon? u understand it's expected to be better than
 java's or c++'s.


We hope to have something concrete by September.
Jul 25 2008
prev sibling parent reply Robert Fraser <fraserofthenight gmail.com> writes:
Walter Bright Wrote:


 superdan wrote:

 Walter Bright Wrote:
 

 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/
 
 and of course on reddit:
 
 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


 
 very cool stuff that's gonna cross the t in threads for d. i'm so
 very glad bartosz is on team. would he join the newsgroup, or is he a
 mimosa as well?


 
 We are very lucky to have Bartosz working with us. He's an expert on 
 multithreaded programming, and even better, he spends a lot of time 
 researching the latest in what others are doing so that D will build on 
 best practices rather than reinvent mistakes.


So, the STM stuff Bartoz was talking about at the conference... is that still
slated for inclusion?
Jul 25 2008
parent Sean Kelly <sean invisibleduck.org> writes:
== Quote from Robert Fraser (fraserofthenight gmail.com)'s article

 Walter Bright Wrote:

 superdan wrote:

 Walter Bright Wrote:


 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

 and of course on reddit:

 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


 very cool stuff that's gonna cross the t in threads for d. i'm so
 very glad bartosz is on team. would he join the newsgroup, or is he a
 mimosa as well?


 We are very lucky to have Bartosz working with us. He's an expert on
 multithreaded programming, and even better, he spends a lot of time
 researching the latest in what others are doing so that D will build on
 best practices rather than reinvent mistakes.


 So, the STM stuff Bartoz was talking about at the conference... is that still
 slated for inclusion?


From what I hear, yes.


Sean
Jul 25 2008
prev sibling next sibling parent reply "Lionello Lunesu" <lionello lunesu.remove.com> writes:
"Walter Bright" <newshound1 digitalmars.com> wrote in message 
news:g6b1ae$up4$1 digitalmars.com...

 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

 and of course on reddit:

 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


I remember reading somewhere that critical-sections on Windows do just that. 
It sure appears that it does from WinNT.h :

typedef struct _RTL_CRITICAL_SECTION {
    PRTL_CRITICAL_SECTION_DEBUG DebugInfo;

    //
    //  The following three fields control entering and exiting the critical
    //  section for the resource
    //

    LONG LockCount;
    LONG RecursionCount;
    HANDLE OwningThread;        // from the thread's ClientId->UniqueThread
    HANDLE LockSemaphore;
    ULONG_PTR SpinCount;        // force size on 64-bit systems when packed
} RTL_CRITICAL_SECTION, *PRTL_CRITICAL_SECTION;

(from WinBase.h: typedef RTL_CRITICAL_SECTION CRITICAL_SECTION;)

L.
Jul 24 2008
parent reply Sean Kelly <sean invisibleduck.org> writes:
Lionello Lunesu wrote:

 
 "Walter Bright" <newshound1 digitalmars.com> wrote in message 
 news:g6b1ae$up4$1 digitalmars.com...

 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

 and of course on reddit:

 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


 
 I remember reading somewhere that critical-sections on Windows do just 
 that.


They pretty much do.  And futexes are largely the same thing in Linux.


Sean
Jul 24 2008
parent reply Leandro Lucarella <llucax gmail.com> writes:
Sean Kelly, el 24 de julio a las 23:09 me escribiste:

 Lionello Lunesu wrote:

"Walter Bright" <newshound1 digitalmars.com> wrote in message 
news:g6b1ae$up4$1 digitalmars.com...

http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

and of course on reddit:

http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


I remember reading somewhere that critical-sections on Windows do just that.


 
 They pretty much do.  And futexes are largely the same thing in Linux.


So, why D must reinvent the wheel, duplicating a well known technique done
in most modern OS? This will only add overhead to D.

Please, at least make it conditional only to OSs that don't provide this
optimization by themselves.

-- 
Leandro Lucarella (luca) | Blog colectivo: http://www.mazziblog.com.ar/blog/
----------------------------------------------------------------------------
GPG Key: 5F5A8D05 (F8CD F9A7 BF00 5431 4145  104C 949E BFB6 5F5A 8D05)
----------------------------------------------------------------------------
"Lidiar" no es lo mismo que "holguear"; ya que "lidiar" es relativo a
"lidia" y "holguear" es relativo a "olga".
	-- Ricardo Vaporeso
Jul 25 2008
parent reply superdan <super dan.org> writes:
Leandro Lucarella Wrote:


 Sean Kelly, el 24 de julio a las 23:09 me escribiste:

 Lionello Lunesu wrote:

"Walter Bright" <newshound1 digitalmars.com> wrote in message 
news:g6b1ae$up4$1 digitalmars.com...

http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

and of course on reddit:

http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


I remember reading somewhere that critical-sections on Windows do just that.


 
 They pretty much do.  And futexes are largely the same thing in Linux.


 
 So, why D must reinvent the wheel, duplicating a well known technique done
 in most modern OS? This will only add overhead to D.
 
 Please, at least make it conditional only to OSs that don't provide this
 optimization by themselves.


there is no reinvention. the man quotes his sources. at most there is
reimplementation. and that's unavoidable as far as i understand. thin locks
must be integrated with the object model so they place that word right there
and do shit with it.
Jul 25 2008
next sibling parent reply Leandro Lucarella <llucax gmail.com> writes:
superdan, el 25 de julio a las 13:17 me escribiste:

 Leandro Lucarella Wrote:
 

 Sean Kelly, el 24 de julio a las 23:09 me escribiste:

 Lionello Lunesu wrote:

"Walter Bright" <newshound1 digitalmars.com> wrote in message 
news:g6b1ae$up4$1 digitalmars.com...

http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

and of course on reddit:

http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


I remember reading somewhere that critical-sections on Windows do just that.


 
 They pretty much do.  And futexes are largely the same thing in Linux.


 
 So, why D must reinvent the wheel, duplicating a well known technique done
 in most modern OS? This will only add overhead to D.
 
 Please, at least make it conditional only to OSs that don't provide this
 optimization by themselves.


 
 there is no reinvention. the man quotes his sources. at most there is
reimplementation. and that's unavoidable as far as i understand. thin locks
must be integrated with the object model so they place that word right there
and do shit with it.


Then please don't add this at all! It's a little lame add overhead to all
modern OS just to be a babysitter of poor ones.

-- 
Leandro Lucarella (luca) | Blog colectivo: http://www.mazziblog.com.ar/blog/
----------------------------------------------------------------------------
GPG Key: 5F5A8D05 (F8CD F9A7 BF00 5431 4145  104C 949E BFB6 5F5A 8D05)
----------------------------------------------------------------------------
41% of all people take people with curly hair less seriously
Jul 25 2008
parent reply Jesse Phillips <jessekphillips gmail.com> writes:
On Fri, 25 Jul 2008 15:06:22 -0300, Leandro Lucarella wrote:


 superdan, el 25 de julio a las 13:17 me escribiste:

 Leandro Lucarella Wrote:
 

 Sean Kelly, el 24 de julio a las 23:09 me escribiste:

 Lionello Lunesu wrote:

"Walter Bright" <newshound1 digitalmars.com> wrote in message
news:g6b1ae$up4$1 digitalmars.com...

http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

and of course on reddit:

http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


I remember reading somewhere that critical-sections on Windows do
just that.


 
 They pretty much do.  And futexes are largely the same thing in
 Linux.


 
 So, why D must reinvent the wheel, duplicating a well known technique
 done in most modern OS? This will only add overhead to D.
 
 Please, at least make it conditional only to OSs that don't provide
 this optimization by themselves.


 
 there is no reinvention. the man quotes his sources. at most there is
 reimplementation. and that's unavoidable as far as i understand. thin
 locks must be integrated with the object model so they place that word
 right there and do shit with it.


 
 Then please don't add this at all! It's a little lame add overhead to
 all modern OS just to be a babysitter of poor ones.


I'm pretty sure that just because your OS handles multi-threading this 
way that your program won't just automatically inherit it. When you right 
your program, you have to handle all the synchronous issues yourself, 
otherwise all this multi-threading business would be a snap.
Jul 26 2008
parent Leandro Lucarella <llucax gmail.com> writes:
Jesse Phillips, el 26 de julio a las 17:27 me escribiste:

 On Fri, 25 Jul 2008 15:06:22 -0300, Leandro Lucarella wrote:
 

 superdan, el 25 de julio a las 13:17 me escribiste:

 Leandro Lucarella Wrote:
 

 Sean Kelly, el 24 de julio a las 23:09 me escribiste:

 Lionello Lunesu wrote:

"Walter Bright" <newshound1 digitalmars.com> wrote in message
news:g6b1ae$up4$1 digitalmars.com...

http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

and of course on reddit:

http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


I remember reading somewhere that critical-sections on Windows do
just that.


 
 They pretty much do.  And futexes are largely the same thing in
 Linux.


 
 So, why D must reinvent the wheel, duplicating a well known technique
 done in most modern OS? This will only add overhead to D.
 
 Please, at least make it conditional only to OSs that don't provide
 this optimization by themselves.


 
 there is no reinvention. the man quotes his sources. at most there is
 reimplementation. and that's unavoidable as far as i understand. thin
 locks must be integrated with the object model so they place that word
 right there and do shit with it.


 
 Then please don't add this at all! It's a little lame add overhead to
 all modern OS just to be a babysitter of poor ones.


 
 I'm pretty sure that just because your OS handles multi-threading this 
 way that your program won't just automatically inherit it. When you right 
 your program, you have to handle all the synchronous issues yourself, 
 otherwise all this multi-threading business would be a snap.


Parse error =)

-- 
Leandro Lucarella (luca) | Blog colectivo: http://www.mazziblog.com.ar/blog/
----------------------------------------------------------------------------
GPG Key: 5F5A8D05 (F8CD F9A7 BF00 5431 4145  104C 949E BFB6 5F5A 8D05)
----------------------------------------------------------------------------
El discman vuelve locos a los controles, te lleva a cualquier lugar.
Ajústense pronto los cinturones, nos vamos a estrellar.
Evidentemente, no escuchaste el speech,
que dio la azafata, antes de despegar.
Jul 27 2008
prev sibling parent Sean Kelly <sean invisibleduck.org> writes:
== Quote from superdan (super dan.org)'s article

 Leandro Lucarella Wrote:

 Sean Kelly, el 24 de julio a las 23:09 me escribiste:

 Lionello Lunesu wrote:

"Walter Bright" <newshound1 digitalmars.com> wrote in message
news:g6b1ae$up4$1 digitalmars.com...

http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

and of course on reddit:

http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


I remember reading somewhere that critical-sections on Windows do just that.


 They pretty much do.  And futexes are largely the same thing in Linux.


 So, why D must reinvent the wheel, duplicating a well known technique done
 in most modern OS? This will only add overhead to D.

 Please, at least make it conditional only to OSs that don't provide this
 optimization by themselves.


 there is no reinvention. the man quotes his sources. at most there is


reimplementation. and that's unavoidable as far as i understand. thin locks must
be integrated with the object model so they place that word right there and do
shit with it.

Yup.  I think the point is trying to avoid a cache miss from having the monitor
in
a separate memory block.


Sean
Jul 25 2008
prev sibling next sibling parent JAnderson <ask me.com> writes:
Walter Bright wrote:

 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/
 
 and of course on reddit:
 
 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


Thanks a very interesting read!

-Joel
Jul 24 2008
prev sibling next sibling parent Alberto Bertogli <albertito blitiri.com.ar> writes:
On Fri, Jul 25, 2008 at 01:17:58PM -0400, superdan wrote:

 Leandro Lucarella Wrote:

 Sean Kelly, el 24 de julio a las 23:09 me escribiste:

 Lionello Lunesu wrote:

"Walter Bright" <newshound1 digitalmars.com> wrote in message
news:g6b1ae$up4$1 digitalmars.com...

http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

and of course on reddit:

http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


I remember reading somewhere that critical-sections on Windows do just that.


 They pretty much do.  And futexes are largely the same thing in Linux.


 So, why D must reinvent the wheel, duplicating a well known technique done
 in most modern OS? This will only add overhead to D.

 Please, at least make it conditional only to OSs that don't provide this
 optimization by themselves.


 there is no reinvention. the man quotes his sources. at most there is
 reimplementation. and that's unavoidable as far as i understand. thin
 locks must be integrated with the object model so they place that word
 right there and do shit with it.


No, it's not unavoidable.

If you just use Linux' futexes (or, from what I read on this threads,
Windows' too), you get those "thin locks" *for free*.

You don't have to implement the whole CAS-based mechanism with fallback
to the OS' locks, just use the locks directly: they implement the same
trick.


Obviously I think the trick is quite nice (and there are several
articles on lwn.net describing Linux' implementation in detail, besides
"man 7 futex", if you are interested), and if the underlying OS doesn't
support it, it's obviously worth it. But at least Linux has been using
it for years, and you gain nothing by using it on that platform.

So the only thing you are going to gain on the platforms that already do
it is a nice layer of overhead.

Thanks,
		Alberto
Jul 25 2008
prev sibling parent reply lurker1024 <lurker1024 mailinator.com> writes:
Walter Bright Wrote:


 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/
 
 and of course on reddit:
 
 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


That's really nice. Have you run some benchmarks on this? Also, are there any
plans to implement this on Tango too?

Thanks.
Jul 25 2008
parent Sean Kelly <sean invisibleduck.org> writes:
== Quote from lurker1024 (lurker1024 mailinator.com)'s article

 Walter Bright Wrote:

 http://bartoszmilewski.wordpress.com/2008/07/24/thin-locks-in-d/

 and of course on reddit:

 http://www.reddit.com/comments/6tbzc/thin_locks_in_d/


 That's really nice. Have you run some benchmarks on this? Also, are there any


plans to implement this on Tango too?

Not sure yet.


Sean
Jul 25 2008