Hi,
I've got the impression that Sun's new JDK 1.3.1 does
never load any JIT via the JIT interface (jit.h) ???
In other words:
java -classic -Djava.compiler=any_jit_at_right_place MainClass
will always fail, even if the JIT is okay and located at the right place.
Is that a known bug?
Hi,
> > java -classic -Djava.compiler=any_jit_at_right_place MainClass
> >
> > will always fail, even if the JIT is okay and located at the right
>
> Sun doesn't register the native implementation for
> java.lang.Compiler.initialize() in their 1.3.1.
> Our 1.3.1 does.
Yes, thanks,
Hi,
I've just released another version of the jLouiss java tracer.
At all this tool helps understanding the execution paths by
logging start and return of Java methods together with their parameters
and result values.
Is is designed as an add-on to the bd Linux JDKs 1.2 and 1.3.
The tracer ou
Hi,
> >I.e., I'm wondering if a gc can discard a compiled method and
> >cause it to run interpreted again the next time it is called.
> >Is this possible?
>
> I imagine it is possible - I'm not aware of anything in the Java specs
> that define behaviour of JITs. I can't guess how likely it is, i
You wrote:
> I had an inquiry/ feature request about how to do this for my JavaUnix project.
> Looks pretty easy on LINUX. Damn if I know how to do on Solaris tho.
> You said some directories in /proc//* are in accessible, ok just try it.
>
> I could also default to `/usr/sbin/ps -afe' or `/usr
Hi,
the subject says it all:
there's a new release of TYA JIT (tya17v2.tgz) available.
http://sax.sax.de/~adlibit/index.html
http://sax.sax.de/~adlibit/tya17v2.tgz
(If you have pointing links to TYA at your site please
update this, because older TYA home page @gonzales.cyberus.ca
is down
Hi,
the subject says it all,
[EMAIL PROTECTED] informed me that he has packaged
a TYA1.6 rpm-archive ready for download:
http://www.spsselib.hiedu.cz/~ivop/ftp/tya/
Tnx to Ivo,
Cheers,
Albrecht
--
To UNSUBSCRIBE, email to [
Hi,
while announcing my new JIT compiler TYA 1.6 release
I wrote about some mirror sites, and now Jens M Andreasen
informed me that tya16.tgz can now be downloaded from:
http://www.dsv.su.se/~jens-and/tya/tya16.tgz
For next days also Ivo Panacek has annonced to build an
rpm archive.
You'll g
Hi,
some news from TYA:
1) Jens M Andreasen helped to make TYA available in Europe
(TYA0.1, 1.0, 1.4v2 and current TYA1.5 and future versions)
http://www.dsv.su.se/~jens-and/tya/
2) Next time we will have another site in .cz
Thanks Ivo you can fetch newest *.rpm archives!
Hi,
the new TYA JIT compiler version 1.5 is ready for
download at URL:
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya15.tgz
(size is 139157)
The problem @gonzalez site is solved due the great work
of Neal Sanche, so please let me say a big Thank You to him!!
And now some TYA release notes:
*
Hi,
Neal Sanche wrote me about a problem we have on tya's home site
ftp://gonzalez.cyberus.ca/pub/Linux/java/
Currently we can not access. If all is okay I'll mail
the information to you.
It seems we are out of luck, because just now I've packaged
the new release of TYA version 1.5 coming wit
Hi,
I've just found an ugly bug in TYA.
It is related to an optimzation in division by 2,
unfortunately introduced in TYA 1.4 (TYA 1.3x was still okay).
If you detect some TYA caused problems, apply this patch:
--
--- tyarecode.c~
> > Nick Lawson writes:
>
> Nick> This sounds like a HUGE bug !
> Not yet. We can't fix it because we don't have the source code for
> sunwjit, but I'm sure we get a fix from Sun soon.
You could consider using TYA JIT. TYA doesn't have the mentioned bug.
(Maybe others, but there's a
> I forgot to mention that the machines are running the x86 JIT. Our next
> step is to find a way to run the Solaris versions of JDK1.2 and see what
> those numbers look like.
TYA jit 1.4 on a plain P200 jdk1.2
takes 100 sec for myprog_static
and 110 sec for myprog.
(But invocation is the m
Hi,
> Having a problem using 'tya' (JIT-complier) with 'Java3D'.
> 'tya' works fine in all other applications, i.e non-Java3D-apps.
Does ``non-Java3D-apps'' include other big apps like java2Demo, Swingset,
some IDEs like NetBeans2 etc ?. If yes, we have a good basic.
Unfortunately I have not ye
Hi,
>
> The following program causes the Linux JDK 1.1.7v3 to crash with a
> SIGFPE. This is because the x86 causes an arithmetic exception when
> you divide 0x8000 by -1; the JVM spec, however, says that the result
> of this should be 0x8000 with NO exception thrown. (The fix is to catch
> > TYA surely isn't for the bit bucket. It is always nice to have an
> > option. Could you imagine that Sun's JIT has quite a few bugs TYA
>
> Other reasons developing TYA are:
>
> -There are options beside of Linux: TYA is widely used
Oops, have forgotten something:
there's anothter option
> Looks like the SUNWjit beats TYA by a large margin (and, without putting
> down the tremendous effort that TYA represents, I more-or-less expected
If you need real Java speed you should quickly switch to Windows
or buy a RS 6000 computer + software ;)
Cheers
Albrecht
Hi,
> Generic JDK variables, Java 1.1 and Java 1.2:
..
> JAVA_COMPILER If set to "tya", then "libtya.so" will be used as
> a JIT compiler if it is installed correctly. If
> set to "none", no JIT will be used.
IMHO should one use a _case_ _sensitive
> > > Is TYA headed for the great, big, bit bucket in the sky? :-(
> >
> > I very much doubt it. The JVM can run with multiple JITs, and TYA can
> > run with multiple JVMs.
>
> TYA surely isn't for the bit bucket. It is always nice to have an
> option. Could you imagine that Sun's JIT has quite
Hi,
I've just released for you another version (1.3v2)
of TYA JIT compiler for the Linux and FreeBSD platforms.
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya13v2.tgz
size is 130253 byte.
Even archive size is only some 100 byte bigger than
the previous release we have once again some progress,
Hi,
> hold on - this post was intended to be just a statement of fact,
> not a complaint. I use blackdown and am very grateful for its existence.
> I've timed tya across releases, and it's gotten steadily faster over the
> last 6 months. The fact that it already keeps up with the sun-supplied
Dimitris wrote:
> In my opinion, even such a release would make sense, since it could point out even
> more problems (and thus expedite the route to final release), and even unveil some
> jdk problems.
so here it is:
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya13.tgz
size is 129612 byte.
> > At any rate, who could answer questions concerning the linux port of sunwjit?
>
> > PS: Hopefully tya1.3 will be out real soon, so that we don't have to depend on
> > sunwjit...
>
> As mentioned, there is an early TYA1.3 running here but I have some
> heavy problems to get EXCEPTIONS_BY_SI
> Dimitris Vyzovitis wrote:
> At any rate, who could answer questions concerning the linux port of sunwjit?
> PS: Hopefully tya1.3 will be out real soon, so that we don't have to depend on
> sunwjit...
OK, you won't to be depend on sunwjit, but don't expect to be faster
than sunwjit. (Hope yo
Hi everybody,
after hours I ve found a bug in TYA's opcode processing
causing some heavy headache here
The bugfix will appear in TYA's next release I am working on:
to run both with jdk1.1.x and jdk1.2 (will be TYA v1.3).
The bug produces floating point unit overflows in certain
situations
> Hi,
>
> Has anybody run any benchmarks comparing jdk117 with tya vs jdk1.2 with
> sunwjit?
> My runs so far (console applications with timing), indicate that
> ***jdk1.1.7+tya1.2 is at least two times faster than jdk1.2-pre1!!!***
> This behavior is for computationally intensive procedures, I d
Hi,
I've heard that it is possible to get
JDK sources in a much relaxed way than earlier.
I know, this is no lawyer's list, but perhaps some
of you can write about license compatibility to GPL?
For example I've written the TYA JIT under clean room
conditions, but looking on JDK sources would
p
Hi,
just released the JIT compiler TYA1.2-version4 on
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya12v4.tgz
(size is 125031 byte)
Some release notes: most important are
Kazuyuki Shudo's fix for better interfacing with 2.2. kernels.
Thanks! And we have some minor speedup in CM3 benchmark, but n
Hi
> > Can the sunwjit be used in combination with the latest JDK1.1.x on Linux?
>
> No, the JIT interface in Java 2 is very different from that in JDK1.1.
Is there any information about (public!) available ?
Cheers
Albrecht
Hi,
> Gerald> Also, am I correct in assuming that the Linux JDK1.2 port
> Gerald> will not have a JIT? Does anyone have any performance
> 1.2 for i386 includes sunwjit. The port isn't finished yet and so it
> doesn't make much sense to show benchmark results now.
Does that mean tha the
Hi,
for all of who have had problems using my mailed diff file:
now you can fetch the complete TYA1.2v3 from URL
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya12v3.tgz
And to be quite complete I've added the missing old TYA 0.1
(originally dated Sep 26 1997 !) into the same directory ;-)
This
Hi,
> > I did _not_ receive any configuration file patches
> > since release of TYA1.2.
>
> I could not get TYA to compile out of the box, as Nate suggested. This was
> becuase I had to set #defined FreeBSD in configure.in, as you suggest in your
> I am not a configure guru
``Me too''
> a
> What does "TYA: no space in inline coding" mean?
>
>
> Tnks!
> Amancio
That's a quite harmless msg: TYA pre-alloced a certain amount
of memory, but due of _inlining_ the complied method exhausted
this memory and so we need to realloc() for some more memory.
Cheers
Albrecht
<[EMAIL PROTECTED]>
+ -tya.c
+prevent that java_lang_Compiler_start(...)
+runs more than once (i.e. at startup only) [due to GC!]
+
+Sun Jan 17 20:19:15 1999 Albrecht Kleine <[EMAIL PROTECTED]>
+ -tya.c
+added bugfix RI-string end address [was introduced in
Hi,
I've just updated the TYA JIT compiler archive.
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya12v2.tgz
size is 123860 byte
This release contains bugfixes and some minor speedup in
invocation, useful in programs like Hanoi towers etc.
Just a note to the freeBSD people:
I did _not_ receive a
> Hi!
>
> What was that #define that were supposed to be commented out to get TYA 1.2 to
> work with Swing, even though it tries to disable JIT during init?
#define IGNORE_DISABLE at first lines of tya.c.
Unfortunately I've forgotten to switch this by default before
releasing.
Cheers,
Albrecht
Hi,
the subject says it all:
TYA1.2 is out now!
For the first time we have also experimental
support for the FreeBSD ports of JDK1.1
Also there are bugfixes, for example a problem
running TogetherJ is reported to be fixed.
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya12.tgz
(size is 125053 b
> Hi,
> Out of interest, did you have any problems getting Together to work with
> TYA? With TYAs default configuration, Together crashed (on either JDK, with
> any thread model) whenever I clicked on any object in the diagram view. I
Next TYA release will STOP running you into this problem :)
T
Hi,
The _latest_ TYA JIT compiler release you find on TYA's ftp home:
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya11v4.tgz
Some notes: This release introduces a new JNI approach.
If you have any JNI-trouble report this please and meanwhile
disable JNINATIVENEW.
If you are runing an newer 2.1.x
> As the linux patch to JDK shows us, linux JDK
> distinguishes whether the platform supports signal
> handler with three arguments or only one argument, with
> the macro SA_SIGINFO.
> JIT compilers want to utilize second and third arguments
> of signal handler for checking access to null and
> d
Hi,
I have rewritten TYA's native code invocation part,
mostly related to newer JNI interface.
For testing purposes I am searching for some
``real world'' applications using JNI compiled
under Linux OS.
I am pleased if you give me some URL hints (http,ftp)
to such programs out there on the ne
Hi,
some users wrote about problems to patch the latest
TYA JIT compiler for Linux (i586) release.
So now I've uploaded the complete archive file to URL
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya11v3.tgz
not just ``ready to run'' ,
but ready to ./configure ; make ; make install
Cheers
me)
Cheers,
Albrecht
patch starts here---
diff -u tya1.1v2/ChangeLog tya1.1v3/ChangeLog
--- tya1.1v2/ChangeLog Thu Oct 15 21:33:20 1998
+++ tya1.1v3/ChangeLog Tue Oct 20 19:27:09 1998
@@ -1,4 +1,17 @@
-Thu Oct 15 21:23:39 1998 Albrecht Kleine <[EMAIL PROTECTED
geLog Thu Oct 15 21:33:20 1998
@@ -1,3 +1,40 @@
+Thu Oct 15 21:23:39 1998 Albrecht Kleine <[EMAIL PROTECTED]>
+ -tyarecode.c
+fixed athrow bug: more independence of EXCEPTION_BY_SIGNALS
+by CompTriggerNullPointerException2
+[tnx bugreport to
+ Kazuyu
Hi everybody,
I've got some mails concerning TYA bugs and compilation errors,
many thanks to all!
1.)
There's at least one compilation error (using newer gcc)
You should apply a patch,
Alexander Davydenko <[EMAIL PROTECTED]> sent this to me:
--- tya.c Fri Oct 9 23:27:21 1998
+++ tya1.1/t
Hi,
yesterday it was Richard Jones who convinced me
to release the TYA1.1 code even though it is still
not bug free.
But besides next months I don't have enough time to
write much for TYA, so now you get a current snapshot.
And at last: after exactly _200_ builds now it's good
time to releas
> Does anyone have the URL for this or any JIT that will work with
> JDK1.1.6v4a. Thanks.
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya1.0.tgz
Cheers
Alb.
Hi,
> |
> | I would like to ask do TYA support a JAVA application with native
> | C function call? It seem that I got core dump in this suitation. I would
> | like to confirm it!
>
> the first time. I sent a note to Albrecht Kleine and he advised me to
> do
You wrote:
Sorry for late answer, but I've spent some nice holidays
far away from home.
> > I've tested this small application (jdk116-v2 with a RH 5.1) and
> > I've discovered a memory corruption:
> > when the focus is lost from the TextField I assign 0 to the local
>
> Might be related to j
Hi,
FYI: jikes 0.36 (announced here on this list some weeks ago)
cannot compile next t.java file:
--
import java.awt.*;
import java.awt.event.*;
class t
{
public t()
{
int e = 0;
Button b = new Button();
Hi,
> Matthew Hunter wrote:
> > > > http://www.javaworld.com/javaworld/jw-08-1998/jw-08-volanomark.html
> >
> > have JIT compilers. The Linux JDK doesn't by default, but you can plug
> > TYA in and use it, which should give a significant performance increase,
> > albeit not sufficient to cat
Hi,
Rick van Rein <[EMAIL PROTECTED]>
was so kind to create some RPM archives for the current
TYA release:
http://www.cs.utwente.nl/~vanrein/linux/tya-1.0-1.src.rpm
http://www.cs.utwente.nl/~vanrein/linux/tya-1.0-1.i386.rpm
FYI: we have changed meanwhile the -O compiler switch
Hi,
I've just uploaded another TYA JIT release
to tya home site:
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya1.0.tgz
Once again I have added some important bugfixes:
for details look into README and Changelog files.
Although TYA is not perfect I think we now have
a good basic for further devel
Hi,
next time I'll release another version of
my TYA JIT compiler add on for Linux x86.
TYA 0.8 is _not_ faster TYA 0.7, but once again
I've fixed some bugs. If you have running TYA 0.7 without
any problems, don't expect too much. But if you are running
java programs TYA 0.7 can't run (aborti
Hi,
got some mails by users reporting a broken TYA archive
at dragon site (tya.home.ml.org).
So perhaps you might better access
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya07.tgz
BTW, the original archive size is 103927 byte.
keep swingin'
Albrecht
Hi,
> Could you confirm that JPEG images work for you? It seems
Yes. Compare other mail.
> the wrong external (readImage), whereas the JPEG library contains
> only sun_awt_image_JPEGImageDecoder_readImage. I don't know enough
> about the java innards to know how the function names are
IMHO no
Stefaan,
> > locate for libjava.so, if you find an additional
> > lib in /usr/X11/lib/ I recommend to move away
> > this file (at least for a test). Now jpeg should work.
> I've only got libjava.so in /usr/lib/jdk1.1.6/lib/i686/green_threads.
Yes okay, of course. There is ONE lib.
But IMH
You wrote
> Hi all,
> I've been away from the list for awhile so forgive me if this has been
> together. I also saw mention of a tya07, but didn't find it on the site I
> went to. Which version of tya is recommended at this time, and when
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya07.tgz
> Hi List,
> java.lang.UnsatisfiedLinkError: readImage
> at sun.awt.image.JPEGImageDecoder.produceImage(JPEGImageDecoder.java:137)
> at
> sun.awt.image.InputStreamImageSource.doFetch(InputStreamImageSource.java:265)
> at sun.awt.image.ImageFetcher.fetchloop(ImageFetcher.java:151)
> at sun.awt
Hi,
> After looking at the official TYA web site and only finding versions
> up to tya06 there, I finally found a tya07 at:
>
>ftp://gonzalez.cyberus.ca/pub/Linux/java/tya07.tgz (June 10)
This is just as I've announced last week :)
Albrecht
PS. on http://tya.home.ml.org we'll have a litt
Steve wrote:
> Fixed in v2 -- coming soon.
>
> > Kevin Fitch
> >
> > ps I think that you are doing great wirk w/ this port. 1.1.6v1 combined
> > w/ tya is just as fast as my regular apps!!
^^^
Cool!
>
> Cool! Thanks!
>
> Steve
>
Thanks!
Albrecht ;-)
Hi,
I've just released the seventh public release of
the JIT compiler add on to Linux-x86-port of JDK1.1.6
to tya home site:
ftp://gonzalez.cyberus.ca/pub/Linux/java/tya07.tgz
and mirrored to:
ftp://ftp.unix-ag.uni-kl.de/incoming/tya07.tgz
ftp://ftp.java-linux.org/pub/incoming/tya07.tgz
Hi everybody,
I am about to release another version of TYA JIT compiler
within the next 10 days (TYA 0.7)
If you are interested in some beta testing (as usual),
please reply to this mail to [EMAIL PROTECTED]
Everybody is welcome, but users of
* glibc
* Nikitin port
* egcs compiler
are
64 matches
Mail list logo
