• Post Reply Bookmark Topic Watch Topic
  • New Topic

Comparing the "time spend in method" of two applications  RSS feed

 
Svend Rost
Ranch Hand
Posts: 904
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
Hi,

I've got two versions of the same application. Application A and Application B
should have the same functionality, but application B have been tuned,
and should therefore be faster. The application is kind of a server
which handles request from it's clients. AFAIK, the requests arrive in
the same order so the results from A and B should be comparable.

A and B have been monitored with YourKit for 8 minuts.


The result is, that App. A (the un-optimized) spends 815 secs in threads,
523 secs of those in Thread.run(). App B should have a better performance,
some methods have been optimized and a threadpool have been introduced.
App. B spends 1669 secs in threads, 1310 secs of those in Thread.run().

What does this result tell you? Is it possible to deduct anything from
such a result? AFAIK, A and B should have handled the
same amount of jobs since they ran for the same amount of time.

Thanks in advance.

/Svend Rost
 
William Brogden
Author and all-around good cowpoke
Rancher
Posts: 13078
6
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
It tells me to suspect that whatever was done to "tune" application B didn't work. I suspect that measurements inside application A to determine potential bottlenecks was not done and that "programmer folklore" techniques were applied to application B.

Alternate explanation: The loads for the two tests were not, in fact, equal. When testing server performance you have to pay attention to simulating the load properly.

Bill
 
Svend Rost
Ranch Hand
Posts: 904
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
Thanks for replying.

I was fearing, that this was the conclusion

So, I hope that
a) the load was not equal
b) A and B was not functional equivalent afterall, i.e. B has more
functionality and therefore had to do more work.

I think I will look into ( b ), and hope that the guys who had to handle mergning
the different paths together gave me wrong information so I checked the
wrong application out..

/Svend Rost
 
  • Post Reply Bookmark Topic Watch Topic
  • New Topic
Boost this thread!