sablevm-developer Mailing List for SableVM (Page 43)
Brought to you by:
egagnon
Menu
▾
▴
sablevm-developer — SableVM developer mailing list
You can subscribe to this list here.
| 2000 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(27) |
Aug
(22) |
Sep
(1) |
Oct
|
Nov
(1) |
Dec
(30) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2001 |
Jan
(2) |
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(2) |
Sep
(32) |
Oct
|
Nov
|
Dec
|
| 2002 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(69) |
Sep
(10) |
Oct
(31) |
Nov
(15) |
Dec
(58) |
| 2003 |
Jan
(33) |
Feb
(81) |
Mar
(85) |
Apr
(24) |
May
(15) |
Jun
(14) |
Jul
(6) |
Aug
(9) |
Sep
(101) |
Oct
(59) |
Nov
(142) |
Dec
(34) |
| 2004 |
Jan
(107) |
Feb
(164) |
Mar
(181) |
Apr
(96) |
May
(81) |
Jun
(71) |
Jul
(1) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: David <db...@cs...> - 2003-06-30 22:19:06
|
Hello, FYI, About my subversion repository layout You will notice that I have 2 SableVM projects in my developer sandbox. One under ..../belanger/sandbox (branch 1) and another under ..../sandbox/sablejit/sablevm (branch 2). branch 1 - will contain only bug fixes that I am considering ready to be merged with SableVM trunk. In the next few days I will be adding them. branch 2 - sablejit/sablevm This is my development code. It contains also the fixes but also unstable code. If you need a fix that I did but cannot wait for it to get to branch 1, this is where you will find it. For people not familiar with svn, you can check the log to see the bug fixes done so far with: svn log -r 247:HEAD svn+ssh://svn.sablevm.org/developers/belanger/sandbox sablevm sablevm-class-library sablevm-native-library Have fun, David --- David B=E9langer Graduate Student School of Computer Science McGill University Office: MC226 Web page: http://www.cs.mcgill.ca/~dbelan2/ Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt |
|
From: David <db...@cs...> - 2003-06-29 03:33:50
|
On Sat, Jun 28, 2003 at 10:20:27PM -0400, Etienne Gagnon wrote: > Hi David, >=20 > David B=E9langer wrote: > >How about rw access? > > > >I still cannot get rw access to work. It doesn't get pass the public > >key authentication. Has everything been set up? I am assuming yes > >since there is the belanger directory. I am also assuming that my > >username is belanger. >=20 > I didn't get your public key, so you don't yet have write access. I sent it a few days ago, I will resend it to both addresses in a few minutes. David >=20 > Note: Apparently there's a routing problem from [sable|cs].mcgill.ca > to mail.uqam.ca, so you should temporarily CC: messages to > eg...@j-.... I guess the routing problem should get solved next > week. >=20 > Etienne >=20 > --=20 > Etienne M. Gagnon, Ph.D. http://www.info.uqam.ca/~egagnon/ > SableVM: http://www.sablevm.org/ > SableCC: http://www.sablecc.org/ >=20 --=20 --- David B=E9langer Graduate Student School of Computer Science McGill University Office: MC226 Web page: http://www.cs.mcgill.ca/~dbelan2/ Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt |
|
From: Etienne G. <eti...@uq...> - 2003-06-29 02:20:44
|
Hi David, David B=E9langer wrote: > How about rw access? >=20 > I still cannot get rw access to work. It doesn't get pass the public > key authentication. Has everything been set up? I am assuming yes > since there is the belanger directory. I am also assuming that my > username is belanger. I didn't get your public key, so you don't yet have write access. Note: Apparently there's a routing problem from [sable|cs].mcgill.ca to mail.uqam.ca, so you should temporarily CC: messages to eg...@j-.... I guess the routing problem should get solved next week. Etienne --=20 Etienne M. Gagnon, Ph.D. http://www.info.uqam.ca/~egagnon/ SableVM: http://www.sablevm.org/ SableCC: http://www.sablecc.org/ |
|
From: David <db...@cs...> - 2003-06-29 02:01:35
|
Got it! How about rw access? I still cannot get rw access to work. It doesn't get pass the public key authentication. Has everything been set up? I am assuming yes since there is the belanger directory. I am also assuming that my username is belanger. pinto:~/work/svn_sablejit/rw> echo $SVN_SSH /usr/bin/ssh -l belanger pinto:~/work/svn_sablejit/rw> svn co svn+ssh://svn.sablevm.org/developers/belanger Permission denied (publickey). svn: Network connection closed unexpectedly svn: Connection closed unexpectedly pinto:~/work/svn_sablejit/rw>=20 With debugging output: debug3: authmethod_is_enabled publickey debug1: next auth method to try is publickey debug1: try privkey: /home/research/ccl/dbelan2/.ssh/identity debug3: no such identity: /home/research/ccl/dbelan2/.ssh/identity debug1: try privkey: /home/research/ccl/dbelan2/.ssh/id_rsa debug3: no such identity: /home/research/ccl/dbelan2/.ssh/id_rsa debug1: try pubkey: /home/research/ccl/dbelan2/.ssh/id_dsa debug3: send_pubkey_test debug2: we sent a publickey packet, wait for reply debug1: authentications that can continue: publickey debug2: we did not send a packet, disable method debug1: no more auth methods to try Permission denied (publickey). debug1: Calling cleanup 0x8063a9c(0x0) svn: Network connection closed unexpectedly svn: Connection closed unexpectedly pinto:~/work/svn_sablejit/rw> = =20 David On Sat, Jun 28, 2003 at 05:12:38PM -0400, Etienne Gagnon wrote: > David B=E9langer wrote: > >I tried to get a read-only copy from the svn repository and it seems t= o > >to just hang. Moreover, viewcvs will not respond until the svn client > >is killed. > > > >I tried: > >svn export svn://svn.sablevm.org:7467/developers/belanger > >svn export svn://svn.sablevm.org:7467/sablevm >=20 > Of course it hangs! Try: > svn export svn://svn.sablevm.org/developers/belanger > instead (without ":7467"). :-) The svn server port is 3690. The 7467 = port > is used by viewcvs. [the svn:// protocol is registered as port 3690; se= e > /etc/services on your system.] >=20 > By the way, svn export is only for getting "snapshots". If you want to= get > a read-only "working-copy" (e.g. you want to be able to type "svn diff"= , > "svn status", etc.) try svn co svn://svn.sablevm.org/developers/belang= er > instead. >=20 > Have fun! >=20 > Etienne >=20 > --=20 > Etienne M. Gagnon, Ph.D. http://www.info.uqam.ca/~egagnon/ > SableVM: http://www.sablevm.org/ > SableCC: http://www.sablecc.org/ >=20 --=20 --- David B=E9langer Graduate Student School of Computer Science McGill University Office: MC226 Web page: http://www.cs.mcgill.ca/~dbelan2/ Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt |
|
From: Etienne G. <eti...@uq...> - 2003-06-28 21:36:58
|
I forgot to say: *NEVER* export svn export svn://svn.sablevm.org/sablevm as this will export a HUGE amount of data (all the historical snapshots[t= ags] and the trunk). Do this instead: svn export svn://svn.sablevm.org/sablevm/trunk sablevm Etienne Etienne Gagnon wrote: > David B=E9langer wrote: >=20 >> I tried to get a read-only copy from the svn repository and it seems t= o >> to just hang. Moreover, viewcvs will not respond until the svn client= >> is killed. >> >> I tried: >> svn export svn://svn.sablevm.org:7467/developers/belanger >> svn export svn://svn.sablevm.org:7467/sablevm >=20 >=20 > Of course it hangs! Try: > svn export svn://svn.sablevm.org/developers/belanger > instead (without ":7467"). :-) The svn server port is 3690. The 7467 = port > is used by viewcvs. [the svn:// protocol is registered as port 3690; se= e > /etc/services on your system.] >=20 > By the way, svn export is only for getting "snapshots". If you want to= get > a read-only "working-copy" (e.g. you want to be able to type "svn diff"= , > "svn status", etc.) try svn co svn://svn.sablevm.org/developers/belang= er > instead. >=20 > Have fun! >=20 > Etienne >=20 --=20 Etienne M. Gagnon, Ph.D. http://www.info.uqam.ca/~egagnon/ SableVM: http://www.sablevm.org/ SableCC: http://www.sablecc.org/ |
|
From: Etienne G. <eti...@uq...> - 2003-06-28 21:12:55
|
David B=E9langer wrote: > I tried to get a read-only copy from the svn repository and it seems to= > to just hang. Moreover, viewcvs will not respond until the svn client > is killed. >=20 > I tried: > svn export svn://svn.sablevm.org:7467/developers/belanger > svn export svn://svn.sablevm.org:7467/sablevm Of course it hangs! Try: svn export svn://svn.sablevm.org/developers/belanger instead (without ":7467"). :-) The svn server port is 3690. The 7467 po= rt is used by viewcvs. [the svn:// protocol is registered as port 3690; see /etc/services on your system.] By the way, svn export is only for getting "snapshots". If you want to g= et a read-only "working-copy" (e.g. you want to be able to type "svn diff", "svn status", etc.) try svn co svn://svn.sablevm.org/developers/belanger= instead. Have fun! Etienne --=20 Etienne M. Gagnon, Ph.D. http://www.info.uqam.ca/~egagnon/ SableVM: http://www.sablevm.org/ SableCC: http://www.sablecc.org/ |
|
From: David <db...@cs...> - 2003-06-28 20:26:38
|
Hi, I tried to get a read-only copy from the svn repository and it seems to to just hang. Moreover, viewcvs will not respond until the svn client is killed. I tried: svn export svn://svn.sablevm.org:7467/developers/belanger svn export svn://svn.sablevm.org:7467/sablevm I am using client svn 0.24.2. David On Sat, Jun 21, 2003 at 03:25:34PM -0400, Etienne Gagnon wrote: > Hi all, >=20 > I have completed the migration of the historical SableVM development > repository to a new Subversion server. Subversion is a follow-up > to CVS that is much better. (http://subversion.tigris.org/) >=20 > Read-only access is available for all through the web at: >=20 > http://svn.sablevm.org:7467/viewcvs/sablevm/trunk/ > http://svn.sablevm.org:7467/viewcvs/sablevm-class-library/trunk/ > http://svn.sablevm.org:7467/viewcvs/sablevm-native-library/trunk/ >=20 > If you have a subversion client (version 0.24.2 is *required*), > you can get a copy of the trunk using: >=20 > svn export svn://svn.sablevm.org/sablevm/trunk sablevm > svn export svn://svn.sablevm.org/sablevm-class-library/trunk=20 > sablevm-class-library > svn export svn://svn.sablevm.org/sablevm-native-library/trunk=20 > sablevm-native-library >=20 > Write access to this repository is *not* managed through SourceForge, > so in order to recover write access (if you had write access > to SourceForge's SableVM CVS repository), you should send me an > ssh public key, ideally signed with your GPG key. >=20 > Have fun with the source! >=20 > Etienne > --=20 > Etienne M. Gagnon, Ph.D. http://www.info.uqam.ca/~egagnon/ > SableVM: http://www.sablevm.org/ > SableCC: http://www.sablecc.org/ >=20 >=20 >=20 > ------------------------------------------------------- > This SF.Net email is sponsored by: INetU > Attention Web Developers & Consultants: Become An INetU Hosting Partner= . > Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission= ! > INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php > _______________________________________________ > Sablevm-developer mailing list > Sab...@li... > https://lists.sourceforge.net/lists/listinfo/sablevm-developer --=20 --- David B=E9langer Graduate Student School of Computer Science McGill University Office: MC226 Web page: http://www.cs.mcgill.ca/~dbelan2/ Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt |
|
From: Etienne G. <eti...@uq...> - 2003-06-21 19:25:51
|
Hi all, I have completed the migration of the historical SableVM development repository to a new Subversion server. Subversion is a follow-up to CVS that is much better. (http://subversion.tigris.org/) Read-only access is available for all through the web at: http://svn.sablevm.org:7467/viewcvs/sablevm/trunk/ http://svn.sablevm.org:7467/viewcvs/sablevm-class-library/trunk/ http://svn.sablevm.org:7467/viewcvs/sablevm-native-library/trunk/ If you have a subversion client (version 0.24.2 is *required*), you can get a copy of the trunk using: svn export svn://svn.sablevm.org/sablevm/trunk sablevm svn export svn://svn.sablevm.org/sablevm-class-library/trunk sablevm-class-library svn export svn://svn.sablevm.org/sablevm-native-library/trunk sablevm-native-library Write access to this repository is *not* managed through SourceForge, so in order to recover write access (if you had write access to SourceForge's SableVM CVS repository), you should send me an ssh public key, ideally signed with your GPG key. Have fun with the source! Etienne -- Etienne M. Gagnon, Ph.D. http://www.info.uqam.ca/~egagnon/ SableVM: http://www.sablevm.org/ SableCC: http://www.sablecc.org/ |
|
From: Etienne G. <eti...@uq...> - 2003-06-19 19:44:32
|
Hi David, Great! I need this:-) (It is so convenient to use ant...) FYI, the subversion server is now ready. One of my students, Christian, = is in the process of writing the little Subversion usage guidelines for the Sable[VM/CC] projects. It would be nice if you could send me an ssh public-key so that I grant you write access to the repository. Once this is in place, it will become quite easy for you to tag and share such preliminary bug-fix sablevm versions. Thanks! Etienne David B=E9langer wrote: > In order to run Ant, two fixes are > necessary. The first one in file > sablepath-classes/src/java/lang/reflect/Method.java > was discussed some time ago: > http://sourceforge.net/mailarchive/forum.php?thread_id=3D1857400&forum_= id=3D4154 >=20 > And the fix for java_io_File.c bug that I submitted today. >=20 > I will do patches later. >=20 >=20 > David >=20 > --- >=20 > David B=E9langer > Graduate Student > School of Computer Science > McGill University > Office: MC226 >=20 > Web page: http://www.cs.mcgill.ca/~dbelan2/ > Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt >=20 >=20 >=20 > ------------------------------------------------------- > This SF.Net email is sponsored by: INetU > Attention Web Developers & Consultants: Become An INetU Hosting Partner= =2E > Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission= ! > INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php > _______________________________________________ > Sablevm-developer mailing list > Sab...@li... > https://lists.sourceforge.net/lists/listinfo/sablevm-developer >=20 >=20 --=20 Etienne M. Gagnon, Ph.D. http://www.info.uqam.ca/~egagnon/ SableVM: http://www.sablevm.org/ SableCC: http://www.sablecc.org/ |
|
From: David <db...@cs...> - 2003-06-19 19:16:15
|
Hi, This is to announce that I was able to successfully build SableCC by running Ant using SableVM. The java compiler used to build SableCC was jikes. In order to run Ant, two fixes are necessary. The first one in file sablepath-classes/src/java/lang/reflect/Method.java was discussed some time ago: http://sourceforge.net/mailarchive/forum.php?thread_id=3D1857400&forum_id= =3D4154 And the fix for java_io_File.c bug that I submitted today. I will do patches later. David --- David B=E9langer Graduate Student School of Computer Science McGill University Office: MC226 Web page: http://www.cs.mcgill.ca/~dbelan2/ Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt |
|
From: David <db...@cs...> - 2003-06-09 00:52:30
|
Hi, SableJIT version 0.2 is now available. This is the second development snapshot made available publicly. It is still unstable and is not ready yet for real world applications. Improvement from 0.1: - Fixed several bugs. - Implementation of some float, double and long operations. - Improved code generated on x86. It is available here: http://www.sable.mcgill.ca/~dbelan2/sablevm/files/jit/ About SableJIT -------------- SableJIT is a retargetable just-in-time compiler for SableVM. Two platforms are currently supported: Debian GNU/Linux/x86 and Debian GNU/Linux/ppc. David --- David B=E9langer Graduate Student School of Computer Science McGill University Office: MC226 Web page: http://www.cs.mcgill.ca/~dbelan2/ Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt |
|
From: Etienne G. <eti...@uq...> - 2003-06-03 20:04:39
|
OK. Here it goes.
In the configure.ac file, there is a call to the macro AC_CANONICAL_HOST
which causes the definition of the $host shell environment variable.
You should thus put your test after the call to AC_CANONICAL_HOST
in configure.ac. (I suggest you move this macro cal up the file, just
after the call to AC_DISABLE_STATIC).
Then you would add:
case $host in
*i386*) do whatever you want
*) other systems...
esac
Etienne
David B=E9langer wrote:
> Hi,
>=20
> I have some x86 assembly file that needs to be compiled with libsablevm=
> only on x86. I tried to conditionally compile with #ifdef in the .s
> assembly file but I think the #ifdef are treated as comments.
>=20
> I tried something like this in Makefile.am
>=20
> if IS_X86
> libsablevm_la_SOURCES =3D libsablevm.c jit_x86.s
> else
> libsablevm_la_SOURCES =3D libsablevm.c
> endif
>=20
> and adding:
>=20
> AM_CONDITIONAL(IS_X86, test x$MACHTYPE =3D xi386)
>=20
> in configure.ac.
>=20
> But I am not familiar at all with the autoconf tools.
>=20
>=20
> Any suggestion?
>=20
>=20
>=20
> David
>=20
> ---
>=20
> David B=E9langer
> Graduate Student
> School of Computer Science
> McGill University
> Office: MC226
>=20
> Web page: http://www.cs.mcgill.ca/~dbelan2/
> Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt
>=20
>=20
>=20
> -------------------------------------------------------
> This SF.net email is sponsored by: eBay
> Get office equipment for less on eBay!
> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> _______________________________________________
> Sablevm-developer mailing list
> Sab...@li...
> https://lists.sourceforge.net/lists/listinfo/sablevm-developer
>=20
>=20
--=20
Etienne M. Gagnon, Ph.D. http://www.info.uqam.ca/~egagnon/
SableVM: http://www.sablevm.org/
SableCC: http://www.sablecc.org/
|
|
From: David <db...@cs...> - 2003-06-03 19:26:45
|
Hi, I have some x86 assembly file that needs to be compiled with libsablevm only on x86. I tried to conditionally compile with #ifdef in the .s assembly file but I think the #ifdef are treated as comments. I tried something like this in Makefile.am if IS_X86 libsablevm_la_SOURCES =3D libsablevm.c jit_x86.s else libsablevm_la_SOURCES =3D libsablevm.c endif and adding: AM_CONDITIONAL(IS_X86, test x$MACHTYPE =3D xi386) in configure.ac. But I am not familiar at all with the autoconf tools. Any suggestion? David --- David B=E9langer Graduate Student School of Computer Science McGill University Office: MC226 Web page: http://www.cs.mcgill.ca/~dbelan2/ Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt |
|
From: christian a. <chr...@ho...> - 2003-05-30 13:09:54
|
I am experiencing some difficulties while installing sablevm-1.0.8 on a Sun computer. It seems that the storage size of the timeval structure from the sys/time library isn't accessible from java_lang_VMObject.c. java_lang_VMObject.c:253: storage size of `now' isn't known Does anyone experienced a similar problem before? What did you do to overcome this? Thanks Christian _________________________________________________________________ STOP MORE SPAM with the new MSN 8 and get 2 months FREE* http://join.msn.com/?page=features/junkmail |
|
From: David <db...@cs...> - 2003-05-27 15:48:41
|
On Mon, May 26, 2003 at 07:59:33PM +0200, Grzegorz B. Prokopski wrote: > W li?cie z pon, 26-05-2003, godz. 01:25, David B=E9langer pisze:=20 > > As previously announced by Etienne a snapshot of SableJIT, a > > retargetable just-in-time compiler for SableVM is now available. > That's terrific! I thing that it's *very* important step in improving > general usability of SableVM. I wish I had been doing such an > interesting project for my M.Sc. thesis :-/ >=20 > You're lucky man ;-) Yes, it's getting quite interesting now that the backend is getting stable. Segmentation faults and Illegal Instructions in the compiled code are not fun to debug. Fortunately, I no longer getting these and I can spend more time on the design/functionality/improvement. >=20 > > Results > > ------- > > Little performance measurements have been done so far but for the > > Matrix example (multiplication of 2 matrices) included with the > > JIT distribution, I got: > >=20 > > ppc: G3 700MHz > > x86: AMD 1200MHz > >=20 > > switch direct inlined jit > > ppc 36.78 28.11 14.40 2.96 > > x86 22.05 9.91 7.47 8.30 > > (time in seconds) > FWIK JIT is that kind of software which you can always trim and improve > to get it better. But it's great you have a working release. As for 0.1 > version - it gives quite nice performance improvement (almost 5 times > for PPC). Just want to add that this is without optimization. This gain is obtained by removing dispatches. Also, currently the compiled code do not have checks on loop back edge to suspend threads for GC. >=20 > > More Info > > --------- > > I presented SableJIT to the Sable Research Group several weeks ago, t= he > > slides are available on my website. > > http://www.sable.mcgill.ca/~dbelan2/#research > Due to shortage of time which I lately encounter - I only took a quick > look. But I hope to have more time next week so don't take the pdf down= . Yes, it will there. >=20 > BTW: I am curious what's the estimated amount of work required to port > the backend to new architecture (I am sure you'll put the answer into > your thesis). This is a good question. I cannot really tell now. Once the design is complete, I will try to port it to a new platform to see. Ideally it should be relatively easy to port it to RISC architecture as this is the model used. I would see it as a 2 steps process. A quick port could be obtained by implementing a small set of basic machine instructions: load, store, add, etc. Then, efficiency could be improved by implementing a larger set (instructions with a default implementation in terms of the the basic ones). Also, there could be the possibility of having architecture specific optimizations. Also, it is important to note that the implementation could be done (and tested) incrementally. The JIT can recover from compilation exceptions. For example, if floating points operations or not implemented yet in a first port release, methods containing floating points will simply be interpreted instead. David >=20 > Thanks for great work >=20 > Grzegorz B. Prokopski > --=20 > Grzegorz B. Prokopski <ga...@de...> > Debian http://www.debian.org/ --=20 --- David B=E9langer Graduate Student School of Computer Science McGill University Office: MC226 Web page: http://www.cs.mcgill.ca/~dbelan2/ Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt |
|
From: Grzegorz B. P. <ga...@de...> - 2003-05-26 17:59:40
|
W li=B6cie z pon, 26-05-2003, godz. 01:25, David B=E9langer pisze:=20 > As previously announced by Etienne a snapshot of SableJIT, a > retargetable just-in-time compiler for SableVM is now available. That's terrific! I thing that it's *very* important step in improving general usability of SableVM. I wish I had been doing such an interesting project for my M.Sc. thesis :-/ You're lucky man ;-) > Results > ------- > Little performance measurements have been done so far but for the > Matrix example (multiplication of 2 matrices) included with the > JIT distribution, I got: >=20 > ppc: G3 700MHz > x86: AMD 1200MHz >=20 > switch direct inlined jit > ppc 36.78 28.11 14.40 2.96 > x86 22.05 9.91 7.47 8.30 > (time in seconds) FWIK JIT is that kind of software which you can always trim and improve to get it better. But it's great you have a working release. As for 0.1 version - it gives quite nice performance improvement (almost 5 times for PPC). > More Info > --------- > I presented SableJIT to the Sable Research Group several weeks ago, the > slides are available on my website. > http://www.sable.mcgill.ca/~dbelan2/#research Due to shortage of time which I lately encounter - I only took a quick look. But I hope to have more time next week so don't take the pdf down. BTW: I am curious what's the estimated amount of work required to port the backend to new architecture (I am sure you'll put the answer into your thesis). Thanks for great work Grzegorz B. Prokopski --=20 Grzegorz B. Prokopski <ga...@de...> Debian http://www.debian.org/ |
|
From: David <db...@cs...> - 2003-05-25 23:25:48
|
Hi,
As previously announced by Etienne a snapshot of SableJIT, a
retargetable just-in-time compiler for SableVM is now available.
Current Status
--------------
Note that this is still work in progress and this first release is still
not stable enough to be used by real applications.
Currently two platforms are supported: GNU/Linux/PPC and GNU/Linux/x86.
The performance on the x86 is currently not good but I am currently
working on this problem.
Currently, no intermediate representation is built and no optimizations
are done, the compiler compiles directly bytecode to native instructions.
Installation
------------
Additional requirements:
- Junit (to run test cases)
Installation instructions are found in the README file.
Results
-------
Little performance measurements have been done so far but for the
Matrix example (multiplication of 2 matrices) included with the
JIT distribution, I got:
ppc: G3 700MHz
x86: AMD 1200MHz
switch direct inlined jit
ppc 36.78 28.11 14.40 2.96
x86 22.05 9.91 7.47 8.30
(time in seconds)
The JIT is slower on x86 due to the unecessarily
frequent spill/reload of registers. This is being fixed.
=20
More Info
---------
I presented SableJIT to the Sable Research Group several weeks ago, the
slides are available on my website.
http://www.sable.mcgill.ca/~dbelan2/#research
Have fun!
David
On Sat, May 24, 2003 at 11:52:24AM -0400, Etienne Gagnon wrote:
> Hi all!
>=20
> I am pleased to announce that David Belanger has made public an initial
> snapshot of his work on a retargettable JIT for SableVM, called SableJI=
T.
>=20
> You can download it from:
> http://sourceforge.net/project/showfiles.php?group_id=3D5523&release_id=
=3D161268
>=20
> I will let David write a longer message describing this initial release=
.
>=20
> Thanks David.
>=20
> Etienne
>=20
> --=20
> Etienne M. Gagnon, Ph.D. http://www.info.uqam.ca/~egagnon/
> SableVM: http://www.sablevm.org/
> SableCC: http://www.sablecc.org/
>=20
>=20
>=20
> -------------------------------------------------------
> This SF.net email is sponsored by: ObjectStore.
> If flattening out C++ or Java code to make your application fit in a
> relational database is painful, don't do it! Check out ObjectStore.
> Now part of Progress Software. http://www.objectstore.net/sourceforge
> _______________________________________________
> Sablevm-developer mailing list
> Sab...@li...
> https://lists.sourceforge.net/lists/listinfo/sablevm-developer
--=20
---
David B=E9langer
Graduate Student
School of Computer Science
McGill University
Office: MC226
Web page: http://www.cs.mcgill.ca/~dbelan2/
Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt
|
|
From: Chris P. <chr...@ma...> - 2003-05-25 08:45:58
|
Grzegorz B. Prokopski wrote: >W li=B6cie z nie, 25-05-2003, godz. 03:24, DAI...@pr...= t >pisze:=20 > =20 > >>im sure its been asked if this works on windows. im an=20 >>OO newb and am installing linux, but its not easy. >>how easy/difficult would it be to rewrite/compile a=20 >>windows non-cygwin Sable if it cant work without a=20 >>change? >> =20 >> > >As for SableVM itself - (almost) all it needs for compilation is >POSIX compatible system and C compiler. FWIK on NT line of Windows >(4.0, 2000, XP) you have sth. called "posix subsystem" near their >kernel. I belive it could be used to provide the functionality >SableVM may need. However I never digged this topic further and it >may be much more complicated that it seems. Maybe looking at mingw >and borland C compilers would give some clues how hard would it be? > >Basically what you need for SableVM is: >- POSIX-compatible environment >- C compiler (currently only GCC is capable to compile SableVM, >however if needed - that could be worked around with some assembly) > >Other possibility you mentioned is using cygwin which may be >best option to start a windows port. > >There's also second part of the problem which is GNU Classpath. >FWIK - nobody has ported it to Windows yet. I think getting it to >compile (at least partially, for ex. w/o GTK-peers) for Windows >would be the first needed step. > >Feel free to ask for more informations > Christian Roy attempted to port SableVM to Windows under MinGW and=20 Cygwin, and made some progress. You should contact him and find out=20 some of the problems. He wrote a project report for a course here at=20 McGill on it ... he'd probably let you see it. cr...@cs... --=20 why not CC everything to this list so we can all read it? Cheers, Chris |
|
From: Etienne G. <eti...@uq...> - 2003-05-24 15:52:39
|
Hi all! I am pleased to announce that David Belanger has made public an initial snapshot of his work on a retargettable JIT for SableVM, called SableJIT. You can download it from: http://sourceforge.net/project/showfiles.php?group_id=5523&release_id=161268 I will let David write a longer message describing this initial release. Thanks David. Etienne -- Etienne M. Gagnon, Ph.D. http://www.info.uqam.ca/~egagnon/ SableVM: http://www.sablevm.org/ SableCC: http://www.sablecc.org/ |
|
From: David <db...@cs...> - 2003-05-16 18:34:45
|
Hi, Actually, I implemented some reflection functionality including that method. This error should not occur anymore with recent version of SableVM since Class.getMethod() is no longer a native method. However, my reflection implementation is still incomplete and may still not work. There are still problem when using SableVM with Ant. I will try to fix some of them when I will find some time. David On Fri, May 16, 2003 at 07:27:39PM +0200, Grzegorz B. Prokopski wrote: > tag 170997 upstream > thanks >=20 > Hi! >=20 > Because: > * there's visible lack of movement on the mailing list (altough > I don't doubt you spend you time wisely and effectively altough > probably on other things) and > * because I was doing "big spring cleaning" of my old bugs and > * because this bug looked interesting > I decided to forward you this one: >=20 > http://bugs.debian.org/170997 >=20 > Martin Waitz <ta...@de...> was trying to build Eclipse using > SableVM (he was using free-java-sdk which uses SableVM). >=20 > It surely would be very important app to get running. It's probably > far away currently but you never know when you may fix the last > showstopper bug ;-) >=20 > Cheers, >=20 > Grzegorz B. Prokopski >=20 > PS: I somehow overlooked this bug because I can't remember receiving > it. It certainally would bring my attention which it didn't. > --=20 > Grzegorz B. Prokopski <ga...@de...> > Debian http://www.debian.org/ >=20 >=20 >=20 > ------------------------------------------------------- > Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara > The only event dedicated to issues related to Linux enterprise solution= s > www.enterpriselinuxforum.com >=20 > _______________________________________________ > Sablevm-developer mailing list > Sab...@li... > https://lists.sourceforge.net/lists/listinfo/sablevm-developer --=20 --- David B=E9langer Graduate Student School of Computer Science McGill University Office: MC226 Web page: http://www.cs.mcgill.ca/~dbelan2/ Public key: http://www.cs.mcgill.ca/~dbelan2/public_key.txt |
|
From: Grzegorz B. P. <ga...@de...> - 2003-05-16 17:28:46
|
tag 170997 upstream thanks Hi! Because: * there's visible lack of movement on the mailing list (altough I don't doubt you spend you time wisely and effectively altough probably on other things) and * because I was doing "big spring cleaning" of my old bugs and * because this bug looked interesting I decided to forward you this one: http://bugs.debian.org/170997 Martin Waitz <ta...@de...> was trying to build Eclipse using SableVM (he was using free-java-sdk which uses SableVM). It surely would be very important app to get running. It's probably far away currently but you never know when you may fix the last showstopper bug ;-) Cheers, Grzegorz B. Prokopski PS: I somehow overlooked this bug because I can't remember receiving it. It certainally would bring my attention which it didn't. -- Grzegorz B. Prokopski <ga...@de...> Debian http://www.debian.org/ |
|
From: Grzegorz B. P. <ga...@de...> - 2003-05-13 08:41:46
|
W liście z wto, 13-05-2003, godz. 10:00, Paolino paperino pisze:
> Hi people,
>
> what type of flag or parameter i might pass to the
> configure script to get a switch threaded version of
> the SableVM.
> In my thesis i must modify the SableVM to create a
> thread (in the sense of program) migration mechanism,
> so i need to work only on the fully interpreted
> execution engine!
> Thanks to evrebody could give me a hint
> Andrea Selva
Running ./configure --help shows these lines (among others):
--enable-debugging-features
Add compiler and runtime checks
--disable-signals-for-exceptions
Do not use signals to detect some exceptions
(NullPointerException, ArithmeticException, etc.)
--with-gc=TYPE
Use given garbage collector (none,copying)
--with-obj-layout=TYPE
Use given object layout (bidirectional,traditional)
--with-threading=TYPE
Use given interpreter threading flavor
(inlined,direct,switch)
The last option lets you choose threading flavour - just pass
--with-threading=switch. Also - if you ever wanted to debug
SableVM - you'd better use --disable-signals-for-exceptions
at least or even --enable-debugging-features.
Good luck!
Grzegorz B. Prokopski
--
Grzegorz B. Prokopski <ga...@de...>
Debian http://www.debian.org/
|
|
From: <von...@ya...> - 2003-05-13 08:01:24
|
Hi people, what type of flag or parameter i might pass to the configure script to get a switch threaded version of the SableVM. In my thesis i must modify the SableVM to create a thread (in the sense of program) migration mechanism, so i need to work only on the fully interpreted execution engine! Thanks to evrebody could give me a hint Andrea Selva ______________________________________________________________________ Yahoo! Mail: 6MB di spazio gratuito, 30MB per i tuoi allegati, l'antivirus, il filtro Anti-spam http://it.yahoo.com/mail_it/foot/?http://it.mail.yahoo.com/ |
|
From: <von...@ya...> - 2003-05-13 08:00:21
|
Hi people, what type of flag or parameter i might pass to the configure script to get a switch threaded version of the SableVM. In my thesis i must modify the SableVM to create a thread (in the sense of program) migration mechanism, so i need to work only on the fully interpreted execution engine! Thanks to evrebody could give me a hint Andrea Selva ______________________________________________________________________ Yahoo! Mail: 6MB di spazio gratuito, 30MB per i tuoi allegati, l'antivirus, il filtro Anti-spam http://it.yahoo.com/mail_it/foot/?http://it.mail.yahoo.com/ |
|
From: Grzegorz B. P. <ga...@de...> - 2003-05-08 08:18:04
|
severity 192246 wishlist tags 192246 upstream thanks W li=B6cie z =B6ro, 07-05-2003, godz. 07:38, Adam Heath pisze:=20 > package: libsablevm-classlib1-java > version: 1.0.8-1 >=20 > This package provides the base .class files unpacked. Not only is this u= gly, > but when using jikes to create dependency makefiles(+M), it causes jikes = to > try to write .u files in /usr/share/sablevm/classes/, for all *.class fil= es. I am afraid there's nothing that can be done about it ATM. SableVM requires basic classpath files to be unpacked. Jar file handling is being done using Java classess so jar-handling classes must be loaded as unpacked before any jar is loaded. That's kind of a chicken&egg problem. It may be possible to solve the problem by, for ex., linking to zip library and using zipped basic classpath classes. However nobody had done anything about it so far and it hadn't been seen as critical (in the sense of being able to operate properly) - that's why I am marking this "wishlist" and Cc:ing upstream. Cheers, Grzegorz B. Prokopski --=20 Grzegorz B. Prokopski <ga...@de...> Debian http://www.debian.org/ |
6 messages has been excluded from this view by a project administrator.
