You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@directory.apache.org by Emmanuel Lécharny <el...@gmail.com> on 2015/06/19 11:41:47 UTC

[Studio] startup performances

Hi guys,

I just profiled Studio this morning, and I see nothing in our code that
causes the slowdown we can perceive with 2.0.0-M9 compared to the
previous version. I suspect LUNA to be more costly than Kepler or JUNO...

Anyone has tested Studio with Mars ?

Re: [Studio] startup performances

Posted by Stefan Seelmann <ma...@stefan-seelmann.de>.
On 06/20/2015 08:33 AM, Jeff MAURY wrote:
> You should check it they are running with the same JVM

Yes, I run all tests with Oracle 1.8.0_45 on Linux 4.0.4.


> On Fri, Jun 19, 2015 at 10:00 PM, Stefan Seelmann <ma...@stefan-seelmann.de>
> wrote:
> 
>> On 06/19/2015 11:41 AM, Emmanuel Lécharny wrote:
>>> Hi guys,
>>>
>>> I just profiled Studio this morning, and I see nothing in our code that
>>> causes the slowdown we can perceive with 2.0.0-M9 compared to the
>>> previous version. I suspect LUNA to be more costly than Kepler or JUNO...
>>>
>>> Anyone has tested Studio with Mars ?
>>>
>>
>> I downloaded Java versions of Juno (3.8) Juno (4.2), Luna, Mars (RC3). I
>> started them with "time ./eclipse" and closed immediately when the
>> windows appeared. Results:
>>
>> Juno 3.8: 10.63s user 0.32s system 193% cpu 5.643 total
>> Juno 4.2: 25.20s user 0.49s system 289% cpu 8.873 total
>> Luna 4.4: 26.96s user 0.54s system 296% cpu 9.288 total
>> Mars 4.5: 33.26s user 0.61s system 308% cpu 10.978 total
>>
>> You see there is a big differences between 3.x and 4.x. Also within the
>> 4.x series the startup time increases slightly from version to version.
>> What's interesting that startup utilizes multiple cores.
>>
>> For Juno there exist two variants:
>> * 3.8 based on old Eclipse 3 code
>> * 4.2 based on new Eclipse 4 code
>>
>> Studio 2.0.0-M8 is based on the Juno 3.8 which doesn't include all the
>> e4 stuff (EMF based UI, CSS styling, dependency injection) and Eclipse 3
>> compatibility layer.
>>
>> I think we have to live with it if we want to use newer Eclipse
>> versions. There are some "tuning" tips in the web like adding
>> -Xverify:none but that didn't change anything for me.
>>
>> Kind Regards,
>> Stefan
>>
>>
> 
>

Re: [Studio] startup performances

Posted by Jeff MAURY <je...@jeffmaury.com>.
You should check it they are running with the same JVM

Jeff

On Fri, Jun 19, 2015 at 10:00 PM, Stefan Seelmann <ma...@stefan-seelmann.de>
wrote:

> On 06/19/2015 11:41 AM, Emmanuel Lécharny wrote:
> > Hi guys,
> >
> > I just profiled Studio this morning, and I see nothing in our code that
> > causes the slowdown we can perceive with 2.0.0-M9 compared to the
> > previous version. I suspect LUNA to be more costly than Kepler or JUNO...
> >
> > Anyone has tested Studio with Mars ?
> >
>
> I downloaded Java versions of Juno (3.8) Juno (4.2), Luna, Mars (RC3). I
> started them with "time ./eclipse" and closed immediately when the
> windows appeared. Results:
>
> Juno 3.8: 10.63s user 0.32s system 193% cpu 5.643 total
> Juno 4.2: 25.20s user 0.49s system 289% cpu 8.873 total
> Luna 4.4: 26.96s user 0.54s system 296% cpu 9.288 total
> Mars 4.5: 33.26s user 0.61s system 308% cpu 10.978 total
>
> You see there is a big differences between 3.x and 4.x. Also within the
> 4.x series the startup time increases slightly from version to version.
> What's interesting that startup utilizes multiple cores.
>
> For Juno there exist two variants:
> * 3.8 based on old Eclipse 3 code
> * 4.2 based on new Eclipse 4 code
>
> Studio 2.0.0-M8 is based on the Juno 3.8 which doesn't include all the
> e4 stuff (EMF based UI, CSS styling, dependency injection) and Eclipse 3
> compatibility layer.
>
> I think we have to live with it if we want to use newer Eclipse
> versions. There are some "tuning" tips in the web like adding
> -Xverify:none but that didn't change anything for me.
>
> Kind Regards,
> Stefan
>
>


-- 
Jeff MAURY


"Legacy code" often differs from its suggested alternative by actually
working and scaling.
 - Bjarne Stroustrup

http://www.jeffmaury.com
http://riadiscuss.jeffmaury.com
http://www.twitter.com/jeffmaury

Re: [Studio] startup performances

Posted by Stefan Seelmann <ma...@stefan-seelmann.de>.
On 06/19/2015 11:41 AM, Emmanuel Lécharny wrote:
> Hi guys,
> 
> I just profiled Studio this morning, and I see nothing in our code that
> causes the slowdown we can perceive with 2.0.0-M9 compared to the
> previous version. I suspect LUNA to be more costly than Kepler or JUNO...
> 
> Anyone has tested Studio with Mars ?
> 

I downloaded Java versions of Juno (3.8) Juno (4.2), Luna, Mars (RC3). I
started them with "time ./eclipse" and closed immediately when the
windows appeared. Results:

Juno 3.8: 10.63s user 0.32s system 193% cpu 5.643 total
Juno 4.2: 25.20s user 0.49s system 289% cpu 8.873 total
Luna 4.4: 26.96s user 0.54s system 296% cpu 9.288 total
Mars 4.5: 33.26s user 0.61s system 308% cpu 10.978 total

You see there is a big differences between 3.x and 4.x. Also within the
4.x series the startup time increases slightly from version to version.
What's interesting that startup utilizes multiple cores.

For Juno there exist two variants:
* 3.8 based on old Eclipse 3 code
* 4.2 based on new Eclipse 4 code

Studio 2.0.0-M8 is based on the Juno 3.8 which doesn't include all the
e4 stuff (EMF based UI, CSS styling, dependency injection) and Eclipse 3
compatibility layer.

I think we have to live with it if we want to use newer Eclipse
versions. There are some "tuning" tips in the web like adding
-Xverify:none but that didn't change anything for me.

Kind Regards,
Stefan