Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.java.programmer > #3787 > unrolled thread

JNA performance

Back to article view | Back to comp.lang.java.programmer

Contents

  JNA performance "Qu0ll" <Qu0llSixFour@gmail.com> - 2011-05-08 15:35 +1000
    Re: JNA performance Knute Johnson <nospam@knutejohnson.com> - 2011-05-07 22:48 -0700
      Re: JNA performance "Qu0ll" <Qu0llSixFour@gmail.com> - 2011-05-08 16:03 +1000
        Re: JNA performance Knute Johnson <nospam@knutejohnson.com> - 2011-05-08 09:54 -0700
      Re: JNA performance Steve Sobol <sjsobol@JustThe.net> - 2011-05-08 13:26 -0700
        Re: JNA performance "Qu0ll" <Qu0llSixFour@gmail.com> - 2011-05-09 08:44 +1000
          Re: JNA performance Jeff Higgins <jeff@invalid.invalid> - 2011-05-08 21:17 -0400
          Re: JNA performance Jeff Higgins <jeff@invalid.invalid> - 2011-05-08 21:20 -0400
            Re: JNA performance Jeff Higgins <jeff@invalid.invalid> - 2011-05-08 21:24 -0400
              Re: JNA performance "Qu0ll" <Qu0llSixFour@gmail.com> - 2011-05-09 13:39 +1000
                Re: JNA performance Steve Sobol <sjsobol@JustThe.net> - 2011-05-08 21:54 -0700
                  Re: JNA performance "Qu0ll" <Qu0llSixFour@gmail.com> - 2011-05-09 15:47 +1000
                    Re: JNA performance Steve Sobol <sjsobol@JustThe.net> - 2011-05-08 23:26 -0700
                      Re: JNA performance "Qu0ll" <Qu0llSixFour@gmail.com> - 2011-05-09 17:07 +1000
                  Re: JNA performance Steve Sobol <sjsobol@JustThe.net> - 2011-05-10 12:44 -0700
          Re: JNA performance Tom Anderson <twic@urchin.earth.li> - 2011-05-09 18:21 +0100
            Re: JNA performance Steve Sobol <sjsobol@JustThe.net> - 2011-05-09 13:05 -0700
              Re: JNA performance Tom Anderson <twic@urchin.earth.li> - 2011-05-09 21:09 +0100
                Re: JNA performance Steve Sobol <sjsobol@JustThe.net> - 2011-05-09 17:58 -0700
    Re: JNA performance Roedy Green <see_website@mindprod.com.invalid> - 2011-05-07 23:25 -0700
      Re: JNA performance "Qu0ll" <Qu0llSixFour@gmail.com> - 2011-05-08 16:32 +1000
      Re: JNA performance Steve Sobol <sjsobol@JustThe.net> - 2011-05-08 13:27 -0700
    Re: JNA performance Lew <noone@lewscanon.com> - 2011-05-08 07:58 -0400
      Re: JNA performance Roedy Green <see_website@mindprod.com.invalid> - 2011-05-08 19:17 -0700
        Re: JNA performance Lew <noone@lewscanon.com> - 2011-05-09 08:36 -0400
          Re: JNA performance Roedy Green <see_website@mindprod.com.invalid> - 2011-05-09 06:34 -0700
            Re: JNA performance Lew <noone@lewscanon.com> - 2011-05-09 09:57 -0400
    O/T: Tag lines, or sigs (Was: JNA performance) Lew <noone@lewscanon.com> - 2011-05-08 08:02 -0400
      Re: O/T: Tag lines, or sigs (Was: JNA performance) Arved Sandstrom <asandstrom3minus1@eastlink.ca> - 2011-05-08 11:24 -0300
        Re: O/T: Tag lines, or sigs (Was: JNA performance) Lew <noone@lewscanon.com> - 2011-05-08 12:35 -0400
          Re: O/T: Tag lines, or sigs (Was: JNA performance) Knute Johnson <nospam@knutejohnson.com> - 2011-05-08 09:54 -0700
        Re: Tag lines, or sigs (Was: JNA performance) "Qu0ll" <Qu0llSixFour@gmail.com> - 2011-05-09 03:37 +1000
        Re: O/T: Tag lines, or sigs (Was: JNA performance) Lawrence D'Oliveiro <ldo@geek-central.gen.new_zealand> - 2011-05-09 15:50 +1200
          Re: O/T: Tag lines, or sigs (Was: JNA performance) Lew <noone@lewscanon.com> - 2011-05-09 08:36 -0400
      Re: O/T: Tag lines, or sigs (Was: JNA performance) "Qu0ll" <Qu0llSixFour@gmail.com> - 2011-05-09 03:37 +1000
        Re: O/T: Tag lines, or sigs (Was: JNA performance) Daniele Futtorovic <da.futt.news@laposte-dot-net.invalid> - 2011-05-08 19:55 +0200
          Re: O/T: Tag lines, or sigs (Was: JNA performance) Martin Gregorie <martin@address-in-sig.invalid> - 2011-05-08 19:30 +0000
            Re: O/T: Tag lines, or sigs (Was: JNA performance) Steve Sobol <sjsobol@JustThe.net> - 2011-05-08 13:28 -0700
    Re: JNA performance Jim Janney <jjanney@shell.xmission.com> - 2011-05-09 08:54 -0600

Page 2 of 2 — ← Prev page 1 [2]

#3794

"Roedy Green"  wrote in message 
news:qmdcs6hrgicldcrr8sa3irkr3n1me4vaon@4ax.com...

<snip>

> I presume you mean JNI not JNA.

No, I mean JNA.

http://jna.java.net

And loving it,

-Qu0ll (Rare, not extinct)
_________________________________________________
Qu0llSixFour@gmail.com
[Replace the "SixFour" with numbers to email me] 

[toc] | [prev] | [next] | [standalone]

#3826

In article <qmdcs6hrgicldcrr8sa3irkr3n1me4vaon@4ax.com>, Roedy Green 
says...
> 
> On Sun, 8 May 2011 15:35:08 +1000, "Qu0ll" <Qu0llSixFour@gmail.com>
> wrote, quoted or indirectly quoted someone who said :
> 
> >I am about to do some Java programming involving native libraries so can 
> >anyone comment on their experiences with the performance of JNA?  I note the 
> >following (somewhat concerning) statement from the JNA website:
> 
> I presume you mean JNI not JNA.

JNA is an alternative to JNI:

http://jna.java.net/

It's much easier to use.

"JNA" is probably *not* a typo.



-- 
Steve Sobol - Programming/WebDev/IT Support
sjsobol@JustThe.net

[toc] | [prev] | [next] | [standalone]

#3806

On 05/08/2011 01:35 AM, Qu0ll wrote:
> I am about to do some Java programming involving native libraries so can
> anyone comment on their experiences with the performance of JNA? I note the
> following (somewhat concerning) statement from the JNA website:
>
> "While some attention is paid to performance, correctness and ease of use take
> priority."
>
> I guess there's little point in using it if it sucks performance-wise even if
> it is "easy to use"! I have never used JNI either but it looks not so "easy to
> use" (i.e. requires some C programming).

Using JNA or JNI for performance is likely to be a fool's errand.  What is the 
performance of your non-JNA approach?  How much better does it need to be? 
Why do you think that calls to non-native libraries are the bottleneck?

Also, where do you get the notion that JNA "sucks performance-wise"?  The JNA 
folks claim that "some attention is paid to performance".  That indicates /a 
priori/ that performance is likely not to "suck".  Therefore any conclusion 
you made that they failed in that attention is based on measurement.  What 
measurements did you perform, what were the results, and what do you need the 
results to be?

-- 
Lew
Honi soit qui mal y pense.
http://upload.wikimedia.org/wikipedia/commons/c/cf/Friz.jpg

[toc] | [prev] | [next] | [standalone]

#3835

On Sun, 08 May 2011 07:58:40 -0400, Lew <noone@lewscanon.com> wrote,
quoted or indirectly quoted someone who said :

>
>Using JNA or JNI for performance is likely to be a fool's errand.  What is the 
>performance of your non-JNA approach?  How much better does it need to be? 
>Why do you think that calls to non-native libraries are the bottleneck?
>
>Also, where do you get the notion that JNA "sucks performance-wise"?  The JNA 
>folks claim that "some attention is paid to performance".  That indicates /a 
>priori/ that performance is likely not to "suck".  Therefore any conclusion 
>you made that they failed in that attention is based on measurement.  What 
>measurements did you perform, what were the results, and what do you need the 
>results to be?

The implementors of JNA will use the same JNI interface you would.
That's the bottleneck.  It is not as though the code is doing
anything, it is just acting as glue to call system functions.  There
is not much you could optimise.

If speed is a problem, caching results may be the way to go, or
running some independent C process that works on your behalf in the
background, or ...

You could get a little speed with a non-WORA interface that was
directly tuned to a particular OS.
-- 
Roedy Green Canadian Mind Products
http://mindprod.com
How long did it take after the car was invented before owners understood 
cars would not work unless you regularly changed the oil and the tires?
We have gone 33 years and still it is rare to uncover a user who 
understands computers don't work without regular backups.

[toc] | [prev] | [next] | [standalone]

#3848

Roedy Green wrote:
> Lew wrote, quoted or indirectly quoted someone who said :
>
>>
>> Using JNA or JNI for performance is likely to be a fool's errand.  What is the
>> performance of your non-JNA approach?  How much better does it need to be?
>> Why do you think that calls to non-native libraries are the bottleneck?
>>
>> Also, where do you get the notion that JNA "sucks performance-wise"?  The JNA
>> folks claim that "some attention is paid to performance".  That indicates /a
>> priori/ that performance is likely not to "suck".  Therefore any conclusion
>> you made that they failed in that attention is based on measurement.  What
>> measurements did you perform, what were the results, and what do you need the
>> results to be?
>
> The implementors of JNA will use the same JNI interface you would.
> That's the bottleneck.  It is not as though the code is doing

So you have access to his performance test results to know where his 
bottleneck is?  Amazing.

I question whether the OP has a performance problem, not where it is.

-- 
Lew
Honi soit qui mal y pense.
http://upload.wikimedia.org/wikipedia/commons/c/cf/Friz.jpg

[toc] | [prev] | [next] | [standalone]

#3850

On Mon, 09 May 2011 08:36:00 -0400, Lew <noone@lewscanon.com> wrote,
quoted or indirectly quoted someone who said :

>So you have access to his performance test results to know where his 
>bottleneck is?  Amazing.

In one sense yes.  All JNI programs that are pure glue have the same
problem. There is nothing else too them but the glue.
-- 
Roedy Green Canadian Mind Products
http://mindprod.com
How long did it take after the car was invented before owners understood 
cars would not work unless you regularly changed the oil and the tires?
We have gone 33 years and still it is rare to uncover a user who 
understands computers don't work without regular backups.

[toc] | [prev] | [next] | [standalone]

#3852

On 05/09/2011 09:34 AM, Roedy Green wrote:
> On Mon, 09 May 2011 08:36:00 -0400, Lew<noone@lewscanon.com>  wrote,
> quoted or indirectly quoted someone who said :
>
>> So you have access to his performance test results to know where his
>> bottleneck is?  Amazing.
>
> In one sense yes.  All JNI programs that are pure glue have the same
> problem. There is nothing else too them but the glue.

You raise a good point, several really.

While it is true that "premature optimization is the root of all evil", as 
Knuth famously pronounced, there is optimization that is not premature.

Roedy's advice is sound because he addresses a well-known source of 
bottlenecks generally, the communication across the boundary between the JVM 
and native code.  It's entirely appropriate to consider the impact of such a 
predictable architectural constraint.

OTOH, appropriate responses to this constraint are very specific to the 
particulars of the application.  If the Java program initiates all contact 
with native logic or data, for example, it might have a lot more control over 
the frequency of contact and the granularity of activity per call.  If contact 
is rare, or only flows from the JVM to the native world without need for 
reply, then the JNI impact might be nil.  No "optimization" needed.

For use cases involving tight communication between native logic and Java 
logic, the pendulum swings the other way.  You'll need to be hyper-aware of 
the issues around the link, not only for performance but for resource disposal 
and many other matters.  Thinking about performance during your architecture 
moments is a good thing to do in this context.

For a great many applications, you should consider performance measurement. 
This does not replace the kind of architectural planning or up-front wisdom 
that Roedy suggests.  It supplements your overall quality strategy.  Profiles 
and baseline throughput measurements help you understand which theoretical 
bottlenecks actually present practical problems, and contrariwise.  They also 
help you decide which ones of the real bottlenecks are worth fixing.  Some 
might be outside program control, or already at optimum albeit disappointing 
performance.  Some might be a little slower than you really could do, but not 
worth the expense or effort to improve slightly.

So do follow Roedy's advice - which I'll paraphrase as "consider the impact of 
the slow communication gate between JVM and native code" - and welcome the 
knowledge he gave, that that gate is a slow one.

-- 
Lew
Honi soit qui mal y pense.
http://upload.wikimedia.org/wikipedia/commons/c/cf/Friz.jpg

[toc] | [prev] | [next] | [standalone]

#3808 — O/T: Tag lines, or sigs (Was: JNA performance)

Qu0ll wrote:
> And loving it,
>
> -Qu0ll (Rare, not extinct)
> _________________________________________________
> Qu0llSixFour@gmail.com
> [Replace the "SixFour" with numbers to email me]

The convention for sigs, or tag lines, is to set them off with "dash dash 
space", that is, "-- ", on its own line, and for the "cute" (in one's own 
opinion) comments to follow that line.

Most good news readers, and a few lamer ones such as what I use, understand 
this convention and adjust the visual display of messages accordingly.

It also sets off stupid remarks like, "Honi soit qui mal y pense" or "And 
loving it" so that they don't appear to be part of the actual point.

-- 
Lew
Honi soit qui mal y pense.
http://upload.wikimedia.org/wikipedia/commons/c/cf/Friz.jpg

[toc] | [prev] | [next] | [standalone]

#3814 — Re: O/T: Tag lines, or sigs (Was: JNA performance)

On 11-05-08 09:02 AM, Lew wrote:
> Qu0ll wrote:
>> And loving it,
>>
>> -Qu0ll (Rare, not extinct)
>> _________________________________________________
>> Qu0llSixFour@gmail.com
>> [Replace the "SixFour" with numbers to email me]
> 
> The convention for sigs, or tag lines, is to set them off with "dash
> dash space", that is, "-- ", on its own line, and for the "cute" (in
> one's own opinion) comments to follow that line.
> 
> Most good news readers, and a few lamer ones such as what I use,
> understand this convention and adjust the visual display of messages
> accordingly.
> 
> It also sets off stupid remarks like, "Honi soit qui mal y pense" or
> "And loving it" so that they don't appear to be part of the actual point.
> 
Qu0ll's phrase always has me wondering whether he's paraphrasing
McDonald's recent branding campaign, or he's old enough to have enjoyed
Get Smart. :-)

AHS

[toc] | [prev] | [next] | [standalone]

#3815 — Re: O/T: Tag lines, or sigs (Was: JNA performance)

Arved Sandstrom wrote:
> Lew wrote:
>> Qu0ll wrote:
>>> And loving it,
>>>
>>> -Qu0ll (Rare, not extinct)
>>> _________________________________________________
>>> Qu0llSixFour@gmail.com
>>> [Replace the "SixFour" with numbers to email me]

>> The convention for sigs, or tag lines, is to set them off with "dash
>> dash space", that is, "-- ", on its own line, and for the "cute" (in
>> one's own opinion) comments to follow that line.
>>
>> Most good news readers, and a few lamer ones such as what I use,
>> understand this convention and adjust the visual display of messages
>> accordingly.
>>
>> It also sets off stupid remarks like, "Honi soit qui mal y pense" or
>> "And loving it" so that they don't appear to be part of the actual point.

> Qu0ll's phrase always has me wondering whether he's paraphrasing
> McDonald's recent branding campaign, or he's old enough to have enjoyed
> Get Smart. :-)

Maybe he's just young enough to have missed it by that much.

Would you believe he's even younger than that and he found his parents' old 
collection of paperback /Get Smart/ novelizations?

No?  Would you believe that he's an historian of classic American television, 
and that he has sixteen researchers digging up citations as we speak?

How about four researchers and an overeager intern?

One long-running Google query and far too much coffee?

-- 
Lew
Sorry about that, chief.

[toc] | [prev] | [next] | [standalone]

#3817 — Re: O/T: Tag lines, or sigs (Was: JNA performance)

On 05/08/2011 09:35 AM, Lew wrote:
> Arved Sandstrom wrote:
>> Lew wrote:
>>> Qu0ll wrote:
>>>> And loving it,
>>>>
>>>> -Qu0ll (Rare, not extinct)
>>>> _________________________________________________
>>>> Qu0llSixFour@gmail.com
>>>> [Replace the "SixFour" with numbers to email me]
>
>>> The convention for sigs, or tag lines, is to set them off with "dash
>>> dash space", that is, "-- ", on its own line, and for the "cute" (in
>>> one's own opinion) comments to follow that line.
>>>
>>> Most good news readers, and a few lamer ones such as what I use,
>>> understand this convention and adjust the visual display of messages
>>> accordingly.
>>>
>>> It also sets off stupid remarks like, "Honi soit qui mal y pense" or
>>> "And loving it" so that they don't appear to be part of the actual
>>> point.
>
>> Qu0ll's phrase always has me wondering whether he's paraphrasing
>> McDonald's recent branding campaign, or he's old enough to have enjoyed
>> Get Smart. :-)
>
> Maybe he's just young enough to have missed it by that much.
>
> Would you believe he's even younger than that and he found his parents'
> old collection of paperback /Get Smart/ novelizations?
>
> No? Would you believe that he's an historian of classic American
> television, and that he has sixteen researchers digging up citations as
> we speak?
>
> How about four researchers and an overeager intern?
>
> One long-running Google query and far too much coffee?
>

:-) - because I'm old enough!

-- 

Knute Johnson
s/knute/nospam/

[toc] | [prev] | [next] | [standalone]

#3821 — Re: Tag lines, or sigs (Was: JNA performance)

"Arved Sandstrom"  wrote in message 
news:vIxxp.52814$sS4.36045@newsfe11.iad...

> Qu0ll's phrase always has me wondering whether he's paraphrasing
> McDonald's recent branding campaign, or he's old enough to have enjoyed
> Get Smart. :-)

Let's just say it has nothing to do with Macca's :-)

-- 
And loving it,

-Qu0ll (Rare, not extinct)
_________________________________________________
Qu0llSixFour@gmail.com
[Replace the "SixFour" with numbers to email me] 

[toc] | [prev] | [next] | [standalone]

#3838 — Re: O/T: Tag lines, or sigs (Was: JNA performance)

In message <vIxxp.52814$sS4.36045@newsfe11.iad>, Arved Sandstrom wrote:

> Qu0ll's phrase always has me wondering whether he's paraphrasing
> McDonald's recent branding campaign, or he's old enough to have enjoyed
> Get Smart. :-)

Of course! The old confuse-them-with-the-obscure-catchphrase trick! :)

[toc] | [prev] | [next] | [standalone]

#3849 — Re: O/T: Tag lines, or sigs (Was: JNA performance)

Lawrence D'Oliveiro wrote:
> Arved Sandstrom wrote:
>> Qu0ll's phrase always has me wondering whether he's paraphrasing
>> McDonald's recent branding campaign, or he's old enough to have enjoyed
>> Get Smart. :-)
>
> Of course! The old confuse-them-with-the-obscure-catchphrase trick! :)

That was actually rather clever.  Good show.

-- 
Lew
Honi soit qui mal y pense.
http://upload.wikimedia.org/wikipedia/commons/c/cf/Friz.jpg

[toc] | [prev] | [next] | [standalone]

#3820 — Re: O/T: Tag lines, or sigs (Was: JNA performance)

"Lew"  wrote in message news:iq60p4$9i5$2@news.albasani.net...

> The convention for sigs, or tag lines, is to set them off with "dash dash 
> space", that is, "-- ", on its own line, and for the "cute" (in one's own 
> opinion) comments to follow that line.
>
> Most good news readers, and a few lamer ones such as what I use, 
> understand this convention and adjust the visual display of messages 
> accordingly.

Most "good" news readers yes, but not this one.  I upgraded to the latest 
version of Windows Live Mail and it is the sorriest piece of utter crap I 
have ever seen.  It used to insert said dashes but then again it used to put 
a '>' in front of quoted lines (I now have to insert them manually) and do a 
bunch of other things which have been removed.

Time for a different news reader it seems.  Any suggestions?

-- 
And loving it,

-Qu0ll (Rare, not extinct)
_________________________________________________
Qu0llSixFour@gmail.com
[Replace the "SixFour" with numbers to email me] 

[toc] | [prev] | [next] | [standalone]

#3823 — Re: O/T: Tag lines, or sigs (Was: JNA performance)

On 08/05/2011 19:37, Qu0ll allegedly wrote:
> Time for a different news reader it seems. Any suggestions?

Thunderbird is decent. It lacks a bit on sigs and message filters, and 
killfiling is tedious. But it's decent.

-- 
DF.
An escaped convict once said to me:
"Alcatraz is the place to be"

[toc] | [prev] | [next] | [standalone]

#3824 — Re: O/T: Tag lines, or sigs (Was: JNA performance)

On Sun, 08 May 2011 19:55:30 +0200, Daniele Futtorovic wrote:

> On 08/05/2011 19:37, Qu0ll allegedly wrote:
>> Time for a different news reader it seems. Any suggestions?
> 
> Thunderbird is decent. It lacks a bit on sigs and message filters, and
> killfiling is tedious. But it's decent.
>
When I used Windows I preferred Forte's Agent to any other reader I'd 
used. 


-- 
martin@   | Martin Gregorie
gregorie. | Essex, UK
org       |

[toc] | [prev] | [next] | [standalone]

#3827 — Re: O/T: Tag lines, or sigs (Was: JNA performance)

In article <iq6r1h$gi9$1@localhost.localdomain>, Martin Gregorie says...
> 
> On Sun, 08 May 2011 19:55:30 +0200, Daniele Futtorovic wrote:
> 
> > On 08/05/2011 19:37, Qu0ll allegedly wrote:
> >> Time for a different news reader it seems. Any suggestions?
> > 
> > Thunderbird is decent. It lacks a bit on sigs and message filters, and
> > killfiling is tedious. But it's decent.
> >
> When I used Windows I preferred Forte's Agent to any other reader I'd 
> used. 

I use MicroPlanet Gravity and "I'm loving it." ;)

http://mpgravity.sf.net


-- 
Steve Sobol - Programming/WebDev/IT Support
sjsobol@JustThe.net

[toc] | [prev] | [next] | [standalone]

#3857

"Qu0ll" <Qu0llSixFour@gmail.com> writes:

> I am about to do some Java programming involving native libraries so
> can anyone comment on their experiences with the performance of JNA?
> I note the following (somewhat concerning) statement from the JNA
> website:
>
> "While some attention is paid to performance, correctness and ease of
> use take priority."
>
> I guess there's little point in using it if it sucks performance-wise
> even if it is "easy to use"!  I have never used JNI either but it
> looks not so "easy to use" (i.e. requires some C programming).

I haven't tried JNA -- it either didn't exist or I didn't know about it
when I needed it -- but it looks attractive.  I would start with JNA and
then decide what to do if performance turns out to be a problem.

JNI depends very much on the size of your project.  If you're only
calling a few functions it isn't too bad, but there's a lot of
boilerplate code needed, which is easy to get wrong and hard to read.
For larger projects SWIG does a decent job of generating code
automatically.

-- 
Jim Janney

[toc] | [prev] | [standalone]

Page 2 of 2 — ← Prev page 1 [2]

Back to top | Article view | comp.lang.java.programmer

csiph-web