RE: heap space and performance
Hallo Edson Carlos Erickss, some results for the Blackdown JDK 1.2.2 RC 3, native threads, sunwjit: It does at least 80.000 loops without problems (I canceled it afterwards); rt.free() is constantly 1048568 Bytes; the effective memory used by the program is constantly 4520 K (+1024 K shared). My system is a up-to-date Debian potato (unstable). ECER>My comparision consist in executing same program using JDK 1.2.2 in a ECER>K6II300/96Mb with WinNT4 using -classic option (I have hotspot ECER>installed). ECER> ECER>loops RAM Virt. Mem. rt.free() ECER>start 7124K5072K 2097144 ECER>150007128K5072K 2097144 ECER>87128K5072K 2097144 ECER>15 6764K 5072K 2097144 <--! ECER>30 6764K 5072K 2097144 MbG, Ekkehard -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
RE: heap space and performance
I've a crash in my machine at 89 loops (not due to JDK, but to S.O. WinNT - the machine has freeze when I've opened a StarCalc sheet, and I've to powerdown.). My conclusion is that JDK is satisfactory to my needs, since he have no changes (Last report show RAM=3600K, Virtual=5072, free=2097144). I think that this type of test is very important, because this show some really conditions with really users. Thanks for all. Edson Richter -- From: Ekkehard Kraemer Sent: quarta-feira, 5 de janeiro de 2000 22:16 To: [EMAIL PROTECTED] Subject:RE: heap space and performance Hallo Edson Carlos Erickss, some results for the Blackdown JDK 1.2.2 RC 3, native threads, sunwjit: It does at least 80.000 loops without problems (I canceled it afterwards); rt.free() is constantly 1048568 Bytes; the effective memory used by the program is constantly 4520 K (+1024 K shared). My system is a up-to-date Debian potato (unstable). ECER>My comparision consist in executing same program using JDK 1.2.2 in a ECER>K6II300/96Mb with WinNT4 using -classic option (I have hotspot ECER>installed). ECER> ECER>loops RAM Virt. Mem. rt.free() ECER>start 7124K5072K 2097144 ECER>150007128K5072K 2097144 ECER>87128K5072K 2097144 ECER>15 6764K 5072K 2097144 <--! ECER>30 6764K 5072K 2097144 MbG, Ekkehard -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Sun JDK 1.2.2 RC2 for Linux
Hi, Has anyone noticed the RC2 of the Sun JDK 1.2.2 for Linux (Except the xdelta "If you don't know what I'm talking about, don't ask" guy)? It has native thread support and mentioned the Blackdown team and individual names in the Credits section, but did not mention Inprise. What's the story? -- Weiqi Gao [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: heap space and performace (what is going on??)
you are TOTALLY incorrect in your assumptions about where i am using roxen. i don't even have any idea what hell you are toaking about. real servers? encoding? whatever man. At 03:13 AM 1/6/00 -0600, Chris Abbey wrote: > >my nose is itching. I might be smelling what I think I smell, >in which case I'm sick of it. I might be imagining things... > > >At 21:47 1/5/00 -0800, Michael E. Moores wrote: >[...] >>i work in one of the biggest linux shops in the world. >[...] >>millions of hits, download at 50+Mbits/second, do lots of complex dynamic >>content; and have uptime of months..months..even years. >>many of these servers live totally unattended in remote locations >>around the world. > >from these and your addie I'll assume you work for real media, >and the boxen in question are the real servers you folks put in >radio stations, etc... > >based on that observation I'll take a couple more disjoint pieces >of your post and go on... > >>but if i use a java environment, it must be relatively free of leaks. >>so i am looking for advise on where to go with blackdown today. >[...] >>we have not had to spend much cash to get >>what we need on the web side. >[...] >>what is the consensus about when the blackdown jdk >>will be ready for what i need to do? i would rather >>use it; money not being a major issue. > >might I suggest you folks put some of that money where your mouths >are then. (no offense) If you want it sooner then higher a couple of >the blackdown team, or find your selves some very good engineers and >pay them to work with blackdown 40 hrs a week. (although you're going >to have to pay a LOT to get talent as good as the blackdown community >has. ;) > >You don't think the nazi used the free decoder rings in the cereal >boxes do you? No. They went out and bought/stole/developed ENIGMA. >How much have you paid for that JVM you're using? Nothing? well... >you get what you pay for That's not a slam against blackdown, >I use it daily even though I have access to the IBM VMs; but then >I'm NOT running a server that I expect to stay up for MONTHs on >end handling gigs a minute and hundreds of dollars worth of >transactions. Speaking of that months on end, bussiness critical >item... have you read items 5 and 6 of the SUN license? I swear >that back in the 102 days it included a warning that java technology >should not be used in any situation with life and death consequences, >or something to that effect. Now that they've changed the marketing >push from toys to servers I can't find that piece anymore. > > cabbey at home dot net <*> http://members.home.net/cabbey > I want a binary interface to the brain! >Today's opto-mechanical digital interfaces are just too slow! > > >-- >To UNSUBSCRIBE, email to [EMAIL PROTECTED] >with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
RE: heap space and performance
that is a valuable (non-political.. hee hee) data point ekkehard. so you must be using glibc2.1.2? i am using mandrake 6.1, which uses glibc2.1.1, so i have also been using JDK 1.2 RC2 to be compatible. i wonder if glibc is causing some of the problesm? i am also seeing some intermittent thread problems when spawning multiple threads which test socket communication. --michael At 12:16 AM 1/6/00 MEZ, Ekkehard Kraemer wrote: >Hallo Edson Carlos Erickss, > >some results for the Blackdown JDK 1.2.2 RC 3, native threads, sunwjit: > >It does at least 80.000 loops without problems (I canceled it afterwards); >rt.free() is constantly 1048568 Bytes; the effective memory used by the >program is constantly 4520 K (+1024 K shared). > >My system is a up-to-date Debian potato (unstable). > >ECER>My comparision consist in executing same program using JDK 1.2.2 in a >ECER>K6II300/96Mb with WinNT4 using -classic option (I have hotspot >ECER>installed). >ECER> >ECER>loops RAM Virt. Mem. rt.free() >ECER>start 7124K5072K 2097144 >ECER>150007128K5072K 2097144 >ECER>87128K5072K 2097144 >ECER>15 6764K 5072K 2097144 <--! >ECER>30 6764K 5072K 2097144 > >MbG, Ekkehard > > >-- >To UNSUBSCRIBE, email to [EMAIL PROTECTED] >with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
RE: heap space and performace (what is going on??)
> > i will have to take a look at TowerJ. > www.deja.com uses TowerJ, but I don't know on which platform. see http://www.towerj.com/pressroom/towerif.shtml for all the TowerJ hoopla rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
RE: heap space and performace (what is going on??)
i'm not sure if i need that kind of performance, but it looks pretty nice and at 1000 bucks a machine it might be effective. i'm still messing around with all the messy glibc stuff; down the road when the term "RC" no longer applies, i would rather say i am doing credit card charges with blackdown. At 10:27 AM 1/6/00 -0600, Wright Robert B Civ 96 CG/SCWDE wrote: > > > > i will have to take a look at TowerJ. > > > >www.deja.com uses TowerJ, but I don't know on which platform. >see http://www.towerj.com/pressroom/towerif.shtml for all the TowerJ hoopla > >rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: heap space and performance
> Michael E Moores writes: Michael> that is a valuable (non-political.. hee hee) Michael> data point ekkehard. Michael> so you must be using glibc2.1.2? Michael> i am using mandrake 6.1, which uses glibc2.1.1, Michael> so i have also been using JDK 1.2 RC2 to be compatible. Michael> i wonder if glibc is causing some of the problesm? Yep, the native threads VM in 1.2-pre-v2 triggered a bug in the LinuxThreads library. The consequence is that GC doesn't work reliable. Michael> i am also seeing some intermittent thread problems when Michael> spawning multiple threads which test socket Michael> communication. The native threads VM has been re-written for 1.2.2-RC1 so most of these problems should be solved in the 1.2.2 RC series. 1.2.2-RC4 (should be available in a couple of days) will fix a few remaining problems like interruption of the main thread and the EBADF messages. Juergen -- Juergen Kreileder, Blackdown Java-Linux Team http://www.blackdown.org/java-linux.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
RE: heap space and performance
that is good news.
i did alot of the same kind of testing.
i did "new Integer[100]"
and i never ran out of heap.
again, i want to look at what libraries your jdk is depending on.
i may have to point my cheap finger at glibc.
i have been using RC2/glibc1.2.1 cause my 3rd party
JNI interface crashes with a glibc1.2.2 dist.
i have a slackware 7.0 box i'll try the code on.
that uses glibc1.2.2
all the glibc stuff is a tough problem. 1.2.1 and 1.2.2
are semaphore incompatible! that makes backward compatibility
an issue. and i am not a library wizard (yet).
have you done any testing with threading?
how about joining on a set of five threads and placing this in
the loop?
At 11:49 PM 1/5/00 +0100, Ekkehard Kraemer wrote:
>Hello Michael,
>
>MEM>i think the jvm/jdk has a big leak with one or more of the
>MEM>classes used.
>
>I'm running the Blackdown JDK (1.2-RC3, with sunwjit and native threads) here
>with very good results. I have only one (non-trivial) application running, and
>it doesn't show your problem. It uses sockets (although only two permanent
>connections) with a steady throughput of around 1 kb/s and is quite graphics
>intensive at the same time.
>
>I have built a memory counter into it (displaying the free memory every 1/10
>s), and the counter is constantly decreasing from around 3600 kb free to 2800
>kb free - then the GC kicks in and the available memory goes back to 3600.
>This whole process of spending and realocating 800 kb of memory takes between
>1 and 2 seconds.
>
>The values displayed by "top" are absolutely constant. I can run the
>application for hours (with the memory usage as described above) and the
>"SIZE" value will not increase at all (as long as I don't tell the program to
>do anything to increase its amount of "real" working data, of course).
>
>Granted, I have done quite a few optimizations in the application to reduce
>the usage of "new", but obviously it recycles quite a bit of memory (800 K per
>1 or 2 seconds) without any (visible) leak.
>
>On the other hand, a small test programm like
>
>class test
>{
> public static void main(String[] args)
> {
> while (true) { new Integer(0); }
> }
>}
>
>indeed crashes with a segmentation fault after about 600.000 iterations (1.3
>seconds) on my system (regardless of any -Xmx setting), so there IS obviously
>a problem -sigh-. But it doesn't seem to be related to a memory leak:
>-verbose:gc shows that after the second GC, every GC frees exactly 23523
>objects/188.864 bytes and finishes with exactly "665320/838856" free. And at
>the time the crash occurs, the application size is only 6 MB.
>
>MbG, Ekkehard
>
>
>--
>To UNSUBSCRIBE, email to [EMAIL PROTECTED]
>with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: heap space and performance
ah, back to the same old problem. maybe i should get a phd in compilers/linkers. i suppose my best bet is to start by compiling everything on the box with the same library path; including the jdk. so when i dive into compiling the jdk when RC4 releases, is the latest source available to do it? thanks, --michael At 06:09 PM 1/6/00 +0100, Juergen Kreileder wrote: > > Michael E Moores writes: > > Michael> that is a valuable (non-political.. hee hee) > Michael> data point ekkehard. > > Michael> so you must be using glibc2.1.2? > > Michael> i am using mandrake 6.1, which uses glibc2.1.1, > Michael> so i have also been using JDK 1.2 RC2 to be compatible. > Michael> i wonder if glibc is causing some of the problesm? > >Yep, the native threads VM in 1.2-pre-v2 triggered a bug in the >LinuxThreads library. The consequence is that GC doesn't work >reliable. > > Michael> i am also seeing some intermittent thread problems when > Michael> spawning multiple threads which test socket > Michael> communication. > >The native threads VM has been re-written for 1.2.2-RC1 so most of >these problems should be solved in the 1.2.2 RC series. 1.2.2-RC4 >(should be available in a couple of days) will fix a few remaining >problems like interruption of the main thread and the EBADF messages. > > > Juergen > >-- >Juergen Kreileder, Blackdown Java-Linux Team >http://www.blackdown.org/java-linux.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: heap space and performance
> Michael E Moores writes: Michael> ah, back to the same old problem. maybe i should get a phd in Michael> compilers/linkers. Michael> i suppose my best bet is to start by compiling everything Michael> on the box with the same library path; including the jdk. You should update your glibc first, the LinuxThreads library is still work in progress. E.g. the original glibc-2.1.2 release didn't pass sigcontext to user handlers, this was fixed in glibc-2.1.2 CVS tree as of 1999/10/24 (this is the version Debian potato uses). glibc-2.1.3 will fix another bunch of thread problems. Michael> so when i dive into compiling the jdk when RC4 releases, Michael> is the latest source available to do it? I'll release diffs for RC4. The diffs will against the SCSL 1.2.2 source code + parts of the SCSL 1.3beta source code (JPDA). Juergen -- Juergen Kreileder, Blackdown Java-Linux Team http://www.blackdown.org/java-linux.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
RE: heap space and performance
Hallo Michael, MEM>that is good news. MEM>i did alot of the same kind of testing. MEM>i did "new Integer[100]" MEM>and i never ran out of heap. I did THAT one, too. I didn't run out of heap, too. But when using "new Integer(0)" instead, it crashes - mind you, it doesn't run out of heap, though. MEM>again, i want to look at what libraries your jdk is depending on. i MEM>may MEM>have to point my cheap finger at glibc. As I said, I'm using Debian potato, that is glibc 2.1.2, AFAIK. MbG, Ekkehard -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: heap space and performance
On Thu, 06 Jan 2000, Juergen Kreileder wrote: > You should update your glibc first, the LinuxThreads library is still > work in progress. E.g. the original glibc-2.1.2 release didn't pass > sigcontext to user handlers, this was fixed in glibc-2.1.2 CVS tree as > of 1999/10/24 (this is the version Debian potato uses). glibc-2.1.3 > will fix another bunch of thread problems. > Juergen, In your opinion, is there a combination of glibc, jdk, and gdb stable enough to debug my native library? I am getting spurious "Program received signal ?, Unknown signal." in my ddd session's gdb window shortly after the formation of the three jvm threads. That is, I am not getting to the meat of my problem. The debug session becomes unusable before I make my first call to the native lib. A more detailed account of what I have been trying is found at the end of this message. It is a post I have made to the SuSE suse-linux-e thread. I am pretty new to linux devlopment and I know that some of my problems are no doubt self inflicted, but it strikes me that you must face this tool compatibility thing every day trying to put out the new jdk builds. What combination of lib, tool, jdk and os versions would you recommend? Thanks in advance, Richard Johnson ** A more detailed account of my problem: I am trying to debug a dynamic native library I have ported from windows that is used with java 1.2. Unfortunately, I am experiencing "Program received signal ?, Unknown signal" while trying to create one of the threads in the java vm. Searching with google and a search string of "unknown signal gdb linux threads" produces a number of posts that describe this problem exactly. For example, http://www.uwsg.indiana.edu/hypermail/linux/kernel/9909.0/0497.html is a post of this problem and solution by a SuSE 6.2 user posted to the linux-kernel list in early September 1999. It suggests rolling back to gdb-4.17.0.11. There IS a version of 4.17.0.12 on my 6.2 distribution and unfortunately, when i use it, I still get this "unknown signal ?" problem when I use it. It does appear that the linux thread patches have been made to this version, as I see output in the ddd's gdb window announcing the new threads as they are created: Continuing. [New Thread 2906] [New Thread 2905] [New Thread 2907] Unfrontunately for me, this is followed by: Program received signal ?, Unknown signal. [Switching to Thread 2907] Program received signal ?, Unknown signal. 0x4011cb6e in __sigsuspend (set=0xbf7ffc74) at ../sysdeps/unix/sysv/linux/sigsuspend.c:48 ../sysdeps/unix/sysv/linux/sigsuspend.c:48: No such file or directory. Trying to continue with a "signal 0" or continue just hangs the ddd/gdb session. I am going to try the .11 version. I think I have it on the 6.1 cdrom. But I guess the question is, shouldn't the .12 version on the 6.2 cd work? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Font Problems
Juergen Kreileder wrote:
The problem is that the URW fonts are available with two different
> foundries: urw and abode. The two foundries use different family
> names for the fonts too. The reason why we currently use the urw
> foundry is that the abode foundry uses ambiguous family names, this
> may result in using the symbol fonts from {75,100}dpi instead of the
> symbol font from urw, this gives a bad font display.
In the font set that I downloaded from the gimp site, the family names are
the same between the "adobe" and "urw" foundary. In particular, in both
cases, it is "Zapf Dingbats" and "Symbol". So perhaps, the names were
changed as part of making the Debian distribution.
> Also, it's advisable to make the fonts available through the Xserver
> directly (i.e. do not use a font server for the urw fonts). This
> allows the JDK to get direct access to the font files.
So are you saying that in my XF86Config file, I should have the actual
directory for the URW fonts in addition to the line which specifies the font
server (and remove them from the configuration file for the font server)? Or
would it be better to put all of the JDK fonts together in the
"jre/lib/fonts" directory? Is this directory always checked first? How does
the JDK behave differently when you make the fonts available through the font
server? I do not see the "dingbats" errors anymore.
Thanks for your input and also the various font files.
Cynthia Jeness
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
New site for doing Java on Linux jlinux.org
Hi there. I have collected various bits of info on running Java on Linux and decided to put it all together on a site. It's still very early on, so there isn't a ton of information, but I'd love to hear feedback from people. It can be found at: http://www.jlinux.org/ Please note that I'm aware that the site has zero cosmetic value. ;-) I'll address that sooner or later, but right now I just wanted to get the info up. --Chris -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: New site for doing Java on Linux jlinux.org
Very good idea. However, how to separate Java from Java/Linux from Linux? Java/Linux from PC? Are you going to have different pages for those topics? Jacob Nikom Christopher Smith wrote: > > Hi there. I have collected various bits of info on running Java on > Linux and decided to put it all together on a site. It's still very > early on, so there isn't a ton of information, but I'd love to hear > feedback from people. It can be found at: > > http://www.jlinux.org/ > > Please note that I'm aware that the site has zero cosmetic value. ;-) > I'll address that sooner or later, but right now I just wanted to get > the info up. > > --Chris > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: New site for doing Java on Linux jlinux.org
On Thu, Jan 06, 2000 at 07:59:55PM -0500, Jacob Nikom wrote: > Very good idea. However, how to separate Java from Java/Linux from > Linux? Java/Linux from PC? Are you going to have different pages > for those topics? I'm not sure what you mean. Do you mean to break the information up into sections according to these categories: a) Java info b) Java & Linux info c) PC info If so, my intention was to specifically point Linux developers in the direction of tools/info that would help them develop Java. If those tools/info help other developers for Java or other developers for Linux, so be it, but lots of other sites provide that info as well, and they undoubtedly have more detailed information. I've done 2 presentations now about doing Java on Linux, and I field questions about it all the time, so I figured this was a specific problem that people have a lot of questions about. Does that answer your question, or did you have something else in mind? --Chris PGP signature
Wanted: Testers for Tritonus MIDI I/O support
Hi, ever wanted to use your MIDI stuff using Java on Linux? Now, there is the beginning of a chance: Tritonus, an open-source implementation of the JavaSound API for Linux, now includes experimental support for MIDI I/O. I've done several tricks but can't finally test it because I have no MIDI gear. If you're interested in trying it, please mail me. Requirements are: - a soundcard supported by ALSA (http://www.alsa-project.org) - Linux 2.2.x kernel (tested with 2.2.10, every recent version should work) - cvs version of the ALSA drivers (tested with cvs snapshot 2000/01/04, be sure to build --with-sequencer=yes) - jdk1.2.x for GNU/Linux i386 (tested with blackdown 1.2.2 RC3 [http://www.blackdown.org], others should work) - the usual development tools Matthias -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
