httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
Subject ap_r patch performance comparison
Date Fri, 19 Jan 2001 05:32:00 GMT

I ran all tests once, and because the screen on my laptop flickered 
(batter flickers occasionally) while running the first rbb2 test, I 
decided to re-run that server just to see if it made a difference.  When
I saw more than double the performance, I re-ran Greg's patch and the 1.3
server.  I do not really consider the unpatched 2.0 server or my original
patch to be interesting enough to run them again.  I absolutely can not
figure out what happened.  Every condition I can see is the same in both
runs.  To ensure that this was not a fluke, I did re-run the tests a third
time for both of the 2.0 patches, and the results are just like the 2nd run,
so they aren't included here.

The tests themselves were run on my computer.  I started the server, and ran
ab on the same machine.  After each test, I stopped the server, and moved the
error log out of the way (That's another topic).  The ab command was always
the same:
	./ab -n 10000 -c 200 http://localhost:8080/ 

And, the results:   (strace's to be generated ASAP)

	Run		1			2

1.3			115.96 req/sec		115.96 req/sec
			768.49 kb/sec		771.45 bb/sec

2.0 (unpatched)		35 req/sec		Not Done
			164.87 kb/sec

2.0 (greg)		90.16 req/sec		227.69 req/sec
			424.72 kb/sec		1083.38 kb/sec

2.0 (rbb)		89.71 req/sec
			420.65 kb/sec

2.0 (rbb2*)		93.00 req/sec		234.09 req/sec
			436.06 kb/sec		1106.45 kb/sec

*  The differences between rbb and rbb2 are relatively simple.  I wanted to 
compare apples to apples, so I modified my patch to use AP_MIN_TO_WRITE as 
size of the bucket buffer, and I used malloc with heap buckets instead of 
palloc with pool buckets.

The difference between Greg's patch and mine is minimal when it comes to
the actual performance.  I want to examine the straces a bit, results will
be posted, but it really looks like this is going to come down to a matter
of personally preference.  Do we want a solution that works for all bucket
programs, or do we want one that allows programmers to use both ap_r and
direct bucket code.  More coming in the next half hour or so.  I am only
going to run strace's for the two patches, because we all know what the
current 2.0 looks like.


Ryan Bloom               
406 29th St.
San Francisco, CA 94131

View raw message