Hi,
I'm a student working with Andreas.
This isn't only a problem of buffered output as you supposed. The output of the
following code is mixed up the same way as the prior one. Indeed, it seems a "a
little less" mixed up, but is still not working as expected.
public class Clockbug {
public static def main(args: Array[String]) {
val runs : Int = 10;
val output : ArrayList[String] = new ArrayList[String]();
val outputRef = GlobalRef[ArrayList[String]](output);
clocked finish {
for (place in Place.places()) at (place) {
clocked async {
for (var i : Int = 0; i < runs; i++) {
val message1 = "Phase A at place " + here.id + " in
run:" + i;
at(outputRef) {
atomic { outputRef().add(message1); }
}
Clock.advanceAll();
val message2 = "Phase B at place " + here.id + " in
run:" + i;
at(outputRef) {
atomic { outputRef().add(message2); }
}
Clock.advanceAll();
}
}}}
for(s in output) {
Console.OUT.println(s);
}
}
}
This code produces output such as:
Phase B at place 2 in run:8
Phase A at place 3 in run:8
Phase B at place 0 in run:8
Phase A at place 1 in run:8
Phase A at place 2 in run:9
Phase B at place 3 in run:8
Phase A at place 0 in run:9
Phase B at place 1 in run:8
Phase B at place 2 in run:9
Phase A at place 3 in run:9
Phase B at place 0 in run:9
Phase A at place 1 in run:9
Phase B at place 3 in run:9
Phase B at place 1 in run:9
As you can see, the runs are still mixed up, let alone the phases. It seems to
get worse with every iterations of the loop. It would be great if you could
help.
Dominic
________________________________________
Von: Olivier Tardieu [tard...@us.ibm.com]
Gesendet: Freitag, 4. Mai 2012 21:02
An: Mailing list for users of the X10 programming language
Betreff: Re: [X10-users] Clock bug?
Hi,
I think you are expecting outputs from distinct places to show in order.
There is no such guarantee.
You code guarantees that Console.OUT.println(...) in phase A at place 1 will
happen before Console.OUT.printn(...) in phase B at place 0.
But the messages may appear on the console in a different order (as they take
time to travel across places).
Olivier
[cid:1__=0abbf367dff501ab8f9e8a93df...@us.ibm.com]Andreas Zwinkau ---05/04/2012
03:07:54 AM--------BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
From: Andreas Zwinkau <zwin...@kit.edu>
To: x10-users@lists.sourceforge.net,
Date: 05/04/2012 03:07 AM
Subject: [X10-users] Clock bug?
________________________________
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
The following program is not synchronized as I expect it to be:
public class Clockbug {
public static def main(args: Array[String]) {
val runs = 7;
clocked finish {
for (place in Place.places()) at (place) {
clocked async {
for (var i : Int = 0; i < runs; i++) {
atomic Console.OUT.println("Run "+i+" phase A at place " +
here.id);
Clock.advanceAll();
atomic Console.OUT.println("Run "+i+" phase B at place " +
here.id);
Clock.advanceAll();
}}}}}}
For each iteration i, the "phase A" output should appear before the
"phase B" output. However, I also observe output like this:
$ X10_NPLACES=4 ./a.out
Run 0 phase A at place 2
Run 0 phase A at place 0
Run 0 phase A at place 1
Run 0 phase A at place 3
Run 0 phase B at place 0
Run 0 phase B at place 2
Run 0 phase B at place 1
Run 1 phase A at place 2
Run 0 phase B at place 3
Run 1 phase A at place 0
Run 1 phase A at place 1
Run 1 phase A at place 3
Run 1 phase B at place 2
Run 1 phase B at place 0
Run 1 phase B at place 1
Run 1 phase B at place 3
Apparently, the iterations get mixed up. Is that a bug in Clock?
- --
Andreas Zwinkau
Karlsruhe Institute of Technology (KIT)
Institut für Programmstrukturen und Datenorganisation (IPD)
Lehrstuhl Prof. Snelting
Adenauerring 20a
76131 Karlsruhe
Phone: +49 721 608 48351
Fax: +49 721 608 48457
Email: zwin...@kit.edu
Web: http://pp.info.uni-karlsruhe.de/person.php?id=107
KIT – University of the State of Baden-Wuerttemberg and
National Research Center of the Helmholtz Association
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEcBAEBAgAGBQJPo4AhAAoJEKm3M8a3pi2iP34H/2/qS2QQ7Xdy3iF1PZVNZ83/
yMKg8K+YZARSiGz+EvN/MWAyHMh10qChWfcoZwhzgTrf9xvGeeN/b2eSCYv0ArcM
k6WdZ+IlWTz1qSZeuLAy5PdmNHpqSb2MO7D7YPK3G+eisq4YOg2KK0gqZsSRvIeI
xaDhTUNEXAX6TwI4cevcDTRrKaFYIbFUdiEslUBBKAlJ33/pjCI0C8UKzGXYXeAQ
IK2CdJIgOsjGuSIHZUFi7iil8TgT3VC+TaBhFTwdpHifuWDHQKsTnL17E+bQhUpr
XG9uD3SOE1nC0J1saJNhYmdTOpX3Lh/LfR244dVnjvqEQeayHhlxszf7IwcFdyc=
=qufe
-----END PGP SIGNATURE-----
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
X10-users mailing list
X10-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/x10-users
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
X10-users mailing list
X10-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/x10-users
