Re: Interpreting Graph Listener results

[Ryszard Lach <ry...@contium.pl>]

On Fri, Aug 29, 2003 at 12:53:51PM -0400, mstover1@apache.org wrote:
> Yes, you are wrong! :-)  The test is copied to each engine, so a test 
> of 10 threads sent to three remote servers equals 30 threads 
> overall.

Well, "User Manual" sais:

"By running JMeter remotely, you can replicate a test across many low-end
computers and thus simulate a larger load on the server"

They wrote "replicate", not "distribute", so according to this Mike is
right. But my tests show that "distribute" would be more accurate here.
If this is not a misspelling - I still have a problem.

> [To Original Writer]:
> You didn't mention average response time, which would be 
> interesting to know.

With static html pages this is about 500ms.

> Also, throughput measurements are dependent on the system 
> clock.  When distributing across 3 machines, if the 3 machines do 
> not agree on the time, your throughput calculations will not be 
> usable.

I think I have NTP installed and working, but I'll check this again and
repeat my tests.

Interesting is, that "idealized throughput" calculated by Aggregate
Listener looks very reasonable.

Richard.

-- 
"First they ignore you. Then they laugh at you. Then they
fight you. Then you win." - Mohandas Gandhi.