Re: Speed issues

I have problems getting apachebench to work correctly on my linux
machines; it fails about 90% of the time to complete its tests.  So I
generated these results with siege:

siege -r 10 -c 10 -u http://localhost/xml.gif # served by skunkweb via mod_skunkweb

Transactions:                     100 hits
Availability:                 100.00 %
Elapsed time:                   1.54 secs
Data transferred:              42900 bytes
Response time:                  0.05 secs
Transaction rate:              64.94 trans/sec
Throughput:                 27857.14 bytes/sec
Concurrency:                    3.03
Successful transactions:         100
Failed transactions:               0


siege -r 10 -c 10 -u http://localhost/images/xml.gif # served by apache

Transactions:                     100 hits
Availability:                 100.00 %
Elapsed time:                   0.93 secs
Data transferred:              42900 bytes
Response time:                  0.07 secs
Transaction rate:             107.53 trans/sec
Throughput:                 46129.03 bytes/sec
Concurrency:                    7.14
Successful transactions:         100
Failed transactions:               0

This isn't directly comparable to your results, but the performance
difference is not as dramatically different as you are seeing.
xml.gif is 429 bytes.  So here's the same test with a file ten times
larger:

  for i in `seq 10`; do cat xml.gif >> foo.gif; done
  cp xml.gif images

# skunk
ransactions:                     100 hits
Availability:                 100.00 %
Elapsed time:                   1.42 secs
Data transferred:             429000 bytes
Response time:                  0.05 secs
Transaction rate:              70.42 trans/sec
Throughput:                302112.69 bytes/sec
Concurrency:                    3.49
Successful transactions:         100
Failed transactions:               0

# apache
ransactions:                     100 hits
Availability:                 100.00 %
Elapsed time:                   1.48 secs
Data transferred:             429000 bytes
Response time:                  0.04 secs
Transaction rate:              67.57 trans/sec
Throughput:                289864.86 bytes/sec
Concurrency:                    2.72
Successful transactions:         100
Failed transactions:               0

I don't know why skunk did better than apache this time round.
Testing on localhost is a very poor indicator of anything.

As for my cpu:

smulloni@bracknell www.smullyan.org $ cat /proc/cpuinfo 
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 8
model name      : Pentium III (Coppermine)
stepping        : 10
cpu MHz         : 999.721
cache size      : 256 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 2
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
mca cmov pat pse36 mmx fxsr sse
bogomips        : 1992.29


I wouldn't claim that skunkweb is super fast, especially not for
static files -- generally speaking I think it is better to put static
files somewhere where they can be served by apache or boa or something
less heavy.  But an order of magnitude performance difference is
unacceptable, obviously.

What OS/Python version?  What is numProcs set to in sw.conf?

js

On Wed, Jan 28, 2004 at 07:51:38AM -0500, Vsevolod (Simon) Ilyushchenko wrote:
> Hi,
> 
> Can anyone provide me with the results of the following test?
> ab -n 100 -c 10 http://your.skunk.site/small.static.file.html
> 
> When I try to open 10 concurrent connections to skunk to retrieve a 
> small file with no Python code (say, an image or a javascript file), the 
> mean/median retrieval times hover around 100-200 ms. The results are the 
> same whether I am going through Apache or not.
> 
> Retrival of the same file on the same machine through a Mason website 
> provides times below 10 ms.
> 
> I have turned off debugging and turned in-memory caching on, but it did 
> not have much effect.
> 
> If you kindly run the test, could you please also send me your CPU speed?
> 
> Thanks,
> Simon
> -- 
> 
> Simon (Vsevolod ILyushchenko)   [email protected]
> 				http://www.simonf.com
> 
> The unknown is honoured, the known is neglected -
>                              until all is known.
> 
>               The C? Chulaind myth
> 
> 
> 
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> Skunkweb-list mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/skunkweb-list
>

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQFAGDj2uqamFyFXXLIRAkkCAJwJTLnb9YHpnFFpRWxENXkn0TeD0QCeOwqA
aFnt+Zs/1/yDqfCPAFJ2GXw=
=ULYR
-----END PGP SIGNATURE-----