s are not so complex.
The only way we've been able to achieve stability is to use Sun JDK 1.1.7v3
with green threads, which clearly defeats the purpose of using SMP systems.
Caveat emptor.
Matt Welsh, UC Berkeley
--
To UNSU
a very interesting space to work in right now. Demanding
I/O applications place a lot of new demands on Java (and operating systems
as well). I would be interested in having more discussion with people on
this list about their experiences!
Cheers-
Matt Welsh, [EMAIL PROTECTED]
-
remove
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Cornell proposed a similar mechanism called "jbufs"
which is based on an extension to the Microsoft Marmot Java compiler.
Basically the same idea.
Cheers -
Matt
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a su
Christopher Smith <[EMAIL PROTECTED]> writes:
> > Matt Welsh wrote,
> > I guess that in principle it ought to be possible to tweak
> > JVMs to special-case a priviledged class of byte[]s to allow
> > them to be pinned for an extended interval without completely
&
Java BOF
sometime during the week, perhaps just meet up at a bar for drinks.
I would like to be able to put names to faces!
Matt
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Also, if you really want this to work, you need to inform the user-level
thread scheduler whenever a kernel thread might block - such as during a
page fault. While something like scheduler activations addresses this, it's
not easy to build.
Matt W
s to Java). It is slated for JDK version 1.4.
This is the earliest it can be done, since 1.3 is already out!
Matt Welsh, UC Berkeley
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
core to our model.
Anyway, I'd be interested to hear any comments about this if you
get a chance to check it out.
Matt Welsh, UC Berkeley
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
ly.
I thought that Sun's RMI did fairly aggressive thread and socket sharing.
NinjaRMI didn't do this because it complicated the design and made it
hard to do pluggable transport layers.
Matt Welsh
http://www.cs.berkeley.edu/~mdw
---
ur standard C library
happens to be installed), and call them directly.
The good news is that it seems to work. Note that when native threads are used,
the JDK does not attempt any funny business, and there are no problems.
Matt Welsh, UC Berkeley
--
lect() and nonblocking I/O internally, they don't seem
to scale -- exactly why, I don't know -- probably because of things like
scheduler overhead.
Matt Welsh
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
wit
Juergen Kreileder <[EMAIL PROTECTED]> writes:
> >>>>> "Matt" == Matt Welsh <[EMAIL PROTECTED]> writes:
>
> Matt> Turns out this is not as easy as it could be -- because even
> Matt> though you call system calls like read(), write
re doing in a 2-tier implementation is
somehow exposing those events to the user in a very coarse-grained and
expensive way (e.g., using signals). In the end your user-level scheduler
does all of the same work and knows about the same things as the kernel
scheduler; why not just optimize kernel threads and be
ares about, regardless of whether programmers "should be" implementing
things that way.
Matt Welsh
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Java. We (the expert group for JSR 51) is currently doing (b).
I am proposing that (a) is useful as well.
Matt Welsh
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
roblems with it. The language spec
is very preliminary and leaves out a lot of details, such as what threads
and locks actually do, what the memory model is, etc. As we all learned
from Java, these seemingly minor details end up being very important down
the line...
Matt Welsh, [EMAIL PROTECTED]
http://www.cs.berkeley.edu/~mdw
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
of the JCP executive
committee so, if the JCP works, some of your changes could easily be
pushed into reality.
Matt Peterson
Caldera Systems, Inc
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
eveloper not to the VM,
threads library, or the OS. The above mentioned article is arguing that
Java threads do not allow the developer enough flexibility to write
something really good and that changes need to be made. I think I
agree.
Matt Peterson
Caldera Systems, Inc
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
t almost ineffective for discussions about Java on Linux.
If we can throttle those discussions I think the list would clear up and
be more effective for its original purpose.
Thanks.
Matt Welsh, UC Berkeley
--
To UNSUBSCR
for people not to send
replies to these "generic" questions to the entire list; we don't need
to see those responses since they aren't relevant.
Matt Welsh
--
To UNSUBSCRIBE, email to [EMAIL PROTECTE
Jesse Erdmann wrote:
>
> Hello all,
>
> I'm trying to port a Java server from Windows to Linux. Both need to
> be interoperable and the clients (Win*/Linux on x86) need to be able to
> connect to either.
>
> On the Windows platform, the developers were able to bind a
> Multicas
John,
Thanks for sending the results of the Volano benchmark to the list. It
was interesting to see the performance comparisons. However, I was
disappointed to see that SMP hardware was not used in the test. Is
there a chance that we will see published numbers that show results on
SMP hardware
John,
You can run "/usr/bin/strings core" to get some info from the core file.
I have been doing Java development daily on a Mandrake 8.2 system for
several months. Glibc version is 2.2.4, but I continuously move to the
latest versions of the Blackdown JVM (currently 1.4.1 beta) so it's been
I would be more than happy to help with builds of the Blackdown JVM.
What do we have to do to get CVS access?
Narendra Sankar wrote:
Hi Everyone
Since I discovered Jedit, I have been looking into jvm performance,
specifically on linux as that is my platform of choice. I love jedit and it
has a
If any memory has paged to swap, garbage collection through the swap
will be excruciatingly slow. Is another app consuming memory, forcing
the server to use the swap partition heavily?
Dave A King wrote:
I'm trying to troubleshoot some really bizarre behavior in a recent build of
our applicatio
ze if 512MB.
Where is all this extra memory coming from? 1GB real memory and 700+MB swap
in use. Does the jvm itself leak memory, even with a fixed heap size?
-d-
-----Original Message-
From: Matt Avery [mailto:mavery@;einnovation.com]
Sent: Friday, November 08, 2002 9:51 AM
To: Dav
Some of the best thread programming advice I've ever read was to never
use Thread.yield() or mess with thread priorities or ThreadGroups in
Java. In my opinion, it's a mistake in the Java API that we even have
the *idea* of Thread.yield() and priorities for the very reason that we
*like* to pr
Muhammad,
Other objects requiring a Local will pick up a default locale if it
isn't supplied. On Unix systems you can set the default Locale that
Java will use by setting some environment variables. Type "locale" at
he command line to see the value of these variables.
Otherwise I have had co
I have had some very strange problems running Java apps with KDE. The
Oracle Java installer comes to mind.
I always run Gnome and developed a fairly large Swing app on my Linux
box ( Mandrake 8.2 then 9.0, Eclipse 2.0 then 2.1 ). We did not
experience any focus problems. The other developers
These new worms are harvesting e-mail off of Windows users machine and
rewriting the "from" field. I use Mozilla on a Linux box and I've
already received several messages from people stating I sent them an
e-mail with a virus. I'm pretty sure they didn't come from my machine! ;)
Ricardo Limon
One comment -- [EMAIL PROTECTED] is a technical mailing list.
Gerald Bauer wrote:
Hello,
I just got an email from a Sun lawyer advising me to
rename the Java Republic news blog to the Republic for
discussion of Java(tm) technology or to the Republic
regarding the Java(tm) platform.
I'm not maki
Thank you!
I know it is time consuming to run all these benchmarks and objectively
publish the results. I desperately wanted to know how the JDK 5.0
compared to other VMs as well, but I simply didn't have the time to run
all these benchmarks myself. Thanks again!
[EMAIL PROTECTED] wrote:
I ha
y of java dev. in my time.
thanks in advance,
matt law.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
101 - 134 of 134 matches
Mail list logo
