search

roc230 / spymemcached

Automatically exported from code.google.com/p/spymemcached
0 stars 0 forks source link

Escaluation efficiency #145

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago 
What version of the product are you using? On what operating system?
spymemcached2.5 on window platform

Tell me more...
I assessed efficiency by making DefaultConnectionFactory.shouldOptimize() 
return true, found escalation does not linearly decrease cost through reducing 
get access times. See attached test class. Here is the result.

Environment configuration:
Cache client: Intel 2 Duo Cpu.
Cache server: 2 instances, each instance runs on 4 cpu processors.

1). Optimize=true: 243,681,844(ns) 
STAT cmd_get 230

2). Optimize=false: 382,587,348(ns)
STAT cmd_get 2000

So my question is: 2) has about four times accesses than 1), why 1) does not 
run faster 4) times than 2)?

P.s. Test class gets the same object in each request so that escalation can be 
applied. Need somebody to confirm whether this is common issue or specific to 
access the same object in multiple threads.

Original issue reported on code.google.com by smilingai2004@gmail.com on 3 Jul 2010 at 4:54

Attachments:

GoogleCodeExporter commented 9 years ago 
Correct one place: 2) has about 9 times accesses than 1).

Original comment by smilingai2004@gmail.com on 3 Jul 2010 at 4:59