It's all about the answers!

Ask a question

What is the performance improvement of using lscm?


diogo cruz (21512) | asked Aug 25 '15, 9:41 a.m.
edited Aug 26 '15, 3:25 a.m. by Krzysztof Kaźmierczyk (7.4k34694)
I don´t see any performance updagres by usage of lscm command instead of scm.

For scm I was able to use the -debug option, which shows me service trips statistics like: 
Interface/method                  Calls  Time
  IRepositoryRemoteService.......     4  48ms
    fetchOrRefreshItems               4  48ms
  IScmQueryService...............     2  53ms
    findBaselineSets                  1  41ms
    findComponents...............     1  12ms
  IScmRichClientRestService           3  35ms
    getVersionableChildren.......     1  16ms
    getVersionables                   1  15ms
    postGetVersionableIdentifiers     1   4ms
-- Total time in service calls: 136ms


The total time needed to execute the command is about 10 seconds, which leads me to think that 8.8 seconds is being used for application bootstrap.

My questions:
- Why the -debug option is not available also for lscm ? Any alternatives to reach the same result?
- How do I make sure that there is indeed a daemon which is being reused to reduce bootstrap? With 10 seconds response time I really doubt it´s being used.


Thank you.


Comments
Lily Wang commented Aug 26 '15, 12:48 a.m.

 I can not answer the two questions you asked. But if the lscm is slower than scm command in your environment, you can try to add an environment variable "LSCM_USE_JAVA_FEC=xyz", and then run lscm command in the command line prompt again.

This workaround is mentioned in https://jazz.net/forum/questions/164481/lscm-scm-command-line-is-very-slow-solved and works for me.

One answer



permanent link
Donald Nong (14.4k313) | answered Aug 26 '15, 3:09 a.m.
I believe the benefit of using "lscm" is to eliminate the startup time of Eclipse, and that's it. The time to execute the service calls on the server side should be the same regardless scm or lscm is used. You can run the "scm version" command to see how long it takes to execute (no server side services involved), and this is roughly the time you can save.

I believe the "-debug" option only reports the "service time" on server side only, and does not take into account the processing time on the client side.

I guess the reason not having "-debug" available to lscm is that the parameter is passed to the Eclipse and is active during the whole time, which is not so good for lscm (every lscm command will end up with the -debug option).

You can use the "list daemon" command to show to existing daemon(s). If you see the same port, you can be sure it's the same daemon.

Your answer


Register or to post your answer.