Re: [JPP-Devel] starter script

2013-10-02 Thread edgar . soldin
On 01.10.2013 22:09, Michaël Michaud wrote:
 Hi Ede,
 
 A few days ago, I ran OpenJUMP on a windows server with 16G RAM,
 and I noticed that it gave nearly 16G to the JVM (I first thought it gave
 even more as it showed 17 000 000 000 bytes).
 In the script you wrote :
 use 100% of ram as default limit (1.124 is a factor to make java64 
 really use that much)
 Don't understand 1.124, but it seems you really give the maximum 
 possible memory
 to the process.

for some reason java does not interpret values literally when assigned via 
-Xmx, but always some percentage less. 

 Can you explain a bit more. Giving 100% ram seems excessive. Moreover, I 
 thing that
 the ram allocated with -xmx excludes the one used by the system to run 
 the jvm itself.

-Xmx is a soft limit telling jre never to use more than that. you can actually 
set -Xmx to bigger values than ram (incl. page file size) w/o a problem. when 
there is no memory to reserve anymore jre will throw heap space error anyway.

rationale is that desktop software should all memory available.

..ede

--
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from 
the latest Intel processors and coprocessors. See abstracts and register 
http://pubads.g.doubleclick.net/gampad/clk?id=60134791iu=/4140/ostg.clktrk
___
Jump-pilot-devel mailing list
Jump-pilot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel


[JPP-Devel] Default Type for bigint

2013-10-02 Thread Eric Lemesre
Hi,

in com.vividsolutions.jump.datastore.jdbc.ValueConverterFactory

Currently the bigint in database table is transcribed into a String object
type.
But the default type of aggregation is also bigint.
for example:

SELECT code_departement, count (*) as nb
   FROM anyTable
   GROUP BY 1

nb has bigint!

I think we should change the type attribute has AttributType.Object to
AttributType.INTERGER
What is your opinion on the subject?

regards
Eric
--
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from 
the latest Intel processors and coprocessors. See abstracts and register 
http://pubads.g.doubleclick.net/gampad/clk?id=60134791iu=/4140/ostg.clktrk___
Jump-pilot-devel mailing list
Jump-pilot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel


Re: [JPP-Devel] Default Type for bigint

2013-10-02 Thread Michaël Michaud

Hi Eric,

Nice to see you still use OpenJUMP ;-)
Currently, I'm not in favour of this change and have a few reasons :

I'm working on a new postgis driver where I need to manage database 
identifiers.

Often, database identifiers use bigint (from sequence). bigint are naturally
converted to java long and I choose to encapsulate it into an Object (it 
has been

converted to a String until september 2013).

Casting a bigint to an int does not seem safe enough to me. For a count
operation, it is ok as we rarely manage more than 1M features in OJ, but 
for

identifiers, it would be really unsafe.

Moreover, it is quite easy to cast count() to int explicitely with for 
example

SELECT code, count (*)::integer as nb FROM anyTable GROUP BY 1

I have also tested to cast AttributeType.OBJECT to AttributeType.INTEGER
afterwards in the schema editor, and it seems to work well.

Other opinions ?

Michaël


Hi,

in com.vividsolutions.jump.datastore.jdbc.ValueConverterFactory

Currently the bigint in database table is transcribed into a String 
object type.

But the default type of aggregation is also bigint.
for example:

SELECT code_departement, count (*) as nb
FROM anyTable
GROUP BY 1

nb has bigint!

I think we should change the type attribute has AttributType.Object to 
AttributType.INTERGER

What is your opinion on the subject?

regards
Eric




--
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register 
http://pubads.g.doubleclick.net/gampad/clk?id=60134791iu=/4140/ostg.clktrk


___
Jump-pilot-devel mailing list
Jump-pilot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel


--
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from 
the latest Intel processors and coprocessors. See abstracts and register 
http://pubads.g.doubleclick.net/gampad/clk?id=60134791iu=/4140/ostg.clktrk___
Jump-pilot-devel mailing list
Jump-pilot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel


Re: [JPP-Devel] Default Type for bigint

2013-10-02 Thread Eric Lemesre
Hi Michaël

Yes i always uses OpenJump. And I tried to do accept as GIS tools by my
society (a big europeen insurance society).

If I resume the situation: handle identifier( sequence) by integers is not
sure ( I don't understand why, but I trust you ) but for others values is
preferable because requests are less long to write.

PostgreSQL jdbc implementation can't retrive key or primary key information
about table? Handle sequence with object and other bigint with integer.

An other subject can I help you on the new postgis driver? Do you have a
road map?

Kinds regards
Eric Lemesre (de mon Androïde)
Le 2 oct. 2013 22:24, Michaël Michaud michael.mich...@free.fr a écrit :

  Hi Eric,

 Nice to see you still use OpenJUMP ;-)
 Currently, I'm not in favour of this change and have a few reasons :

 I'm working on a new postgis driver where I need to manage database
 identifiers.
 Often, database identifiers use bigint (from sequence). bigint are
 naturally
 converted to java long and I choose to encapsulate it into an Object (it
 has been
 converted to a String until september 2013).

 Casting a bigint to an int does not seem safe enough to me. For a count
 operation, it is ok as we rarely manage more than 1M features in OJ, but
 for
 identifiers, it would be really unsafe.

 Moreover, it is quite easy to cast count() to int explicitely with for
 example
 SELECT code, count (*)::integer as nb FROM anyTable GROUP BY 1

 I have also tested to cast AttributeType.OBJECT to AttributeType.INTEGER
 afterwards in the schema editor, and it seems to work well.

 Other opinions ?

 Michaël

   Hi,

  in com.vividsolutions.jump.datastore.jdbc.ValueConverterFactory

 Currently the bigint in database table is transcribed into a String object
 type.
 But the default type of aggregation is also bigint.
 for example:

 SELECT code_departement, count (*) as nb
FROM anyTable
GROUP BY 1

 nb has bigint!

 I think we should change the type attribute has AttributType.Object to
 AttributType.INTERGER
 What is your opinion on the subject?

  regards
  Eric




 --
 October Webinars: Code for Performance
 Free Intel webinars can help you accelerate application performance.
 Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
 the latest Intel processors and coprocessors. See abstracts and register 
 http://pubads.g.doubleclick.net/gampad/clk?id=60134791iu=/4140/ostg.clktrk



 ___
 Jump-pilot-devel mailing 
 listJump-pilot-devel@lists.sourceforge.nethttps://lists.sourceforge.net/lists/listinfo/jump-pilot-devel




 --
 October Webinars: Code for Performance
 Free Intel webinars can help you accelerate application performance.
 Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
 from
 the latest Intel processors and coprocessors. See abstracts and register 
 http://pubads.g.doubleclick.net/gampad/clk?id=60134791iu=/4140/ostg.clktrk
 ___
 Jump-pilot-devel mailing list
 Jump-pilot-devel@lists.sourceforge.net
 https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel


--
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from 
the latest Intel processors and coprocessors. See abstracts and register 
http://pubads.g.doubleclick.net/gampad/clk?id=60134791iu=/4140/ostg.clktrk___
Jump-pilot-devel mailing list
Jump-pilot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel