 | |
performance issues writing udf
comp.databases.ibm-db2 comp.databases.ibm-db2
 |
| | Thread Tools | Display Modes |
#1
 
| |||
| |||
|
performance issues writing udf -
08-09-2010
, 12:00 AM
#2
| |||
| |||
|
|
hello, i try to implement some methods in udf for db2 (no specific platform). now i am unsure if it would be better using c or java as programming language. is there a benchmark for getting information about the performance using these languages or a comparison? thanks and regards |
you could bring down DB2.
Java is the worst (always fenced).
SQL is in-between.
#3
| |||
| |||
|
|
"datenwort" <sebastian.saemi... (AT) sowarent (DOT) de> wrote in message news:87615d9a-ea60-48d7-b122-b3f8fe415b08 (AT) d17g2000yqb (DOT) googlegroups.com... hello, i try to implement some methods in udf for db2 (no specific platform). now i am unsure if it would be better using c or java as programming language. is there a benchmark for getting information about the performance using these languages or a comparison? thanks and regards C Not Fenced is the best (but you must make sure they are thread safe our you could bring down DB2. Java is the worst (always fenced). SQL is in-between. |
#4
| |||
| |||
|
|
hello, thanks for answering. why is java the worst case in developing user defined functions? is there a benchmark or just own experience? |
reasons. The information was previously provided by IBM architects on this
newsgroup, although it was specifically referring to Stored Procedures. If
you go to Google groups you may be able find the old posts on this subject.
#5
| |||
| |||
|
|
thanks for answering. why is java the worst case in developing user defined functions? is there a benchmark or just own experience? |
--
Helmut K. C. Tessarek
DB2 Performance and Development
/*
Thou shalt not follow the NULL pointer for chaos and madness
await thee at its end.
*/
#6
| |||
| |||
|
|
"datenwort"<sebastian.saemisch (AT) sowarent (DOT) de> wrote in message news:a3eab89a-ba2e-4e45-bad6-433920717e9a (AT) f6g2000yqa (DOT) googlegroups.com... hello, thanks for answering. why is java the worst case in developing user defined functions? is there a benchmark or just own experience? Partly because Java always runs as fenced mode, and partly for other reasons. The information was previously provided by IBM architects on this newsgroup, although it was specifically referring to Stored Procedures. If you go to Google groups you may be able find the old posts on this subject. I'm actually presently preparing a deck called "versus" where I |
SQL UDF actually are faster than C-UDF when they are just encapsulating
simple expression evaluation. That is because even an unfenced UDF
undergoes overhead. And of course a SQL UDF DB2's optimizer can understand.
--
Serge Rielau
SQL Architect DB2 for LUW
IBM Toronto Lab
#7
| |||
| |||
|
|
I'm actually presently preparing a deck called "versus" where I investigate such questions. SQL UDF actually are faster than C-UDF when they are just encapsulating simple expression evaluation. That is because even an unfenced UDF undergoes overhead. And of course a SQL UDF DB2's optimizer can understand. |
Will you post the details somewhere or send me a copy?
#8
| |||
| |||
|
|
I'm actually presently preparing a deck called "versus" where I investigate such questions. SQL UDF actually are faster than C-UDF when they are just encapsulating simple expression evaluation. That is because even an unfenced UDF undergoes overhead. And of course a SQL UDF DB2's optimizer can understand. Will you post the details somewhere or send me a copy? It will be presented at DB2 Aktuell in Germany in mid September and then |
Once it's completed I have no qualms distributing it to interested parties.
Cheers
Serge
--
Serge Rielau
SQL Architect DB2 for LUW
IBM Toronto Lab
|
« Previous Thread
|
Next Thread »
| Thread Tools | |
| Display Modes | |
Linear Mode
| |
Powered by vBulletin Version 3.5.3
Copyright ©2000 - 2012, Jelsoft Enterprises Ltd.
Copyright ©2000 - 2012, Jelsoft Enterprises Ltd.