Each operation has a CPU Cost, and that cost is not related to real time.
the CPU time is based on a single CPU at some clock speed (don't remember
what). Because 12 cores could work on your request at twice the CPU you
speed of the billable cpu, you can run up 24s of CPU in 1s of real time.
And I think the actual numbers can be higher than that, depending on the
operation.

 

Conversely I have operations that consume like 4ms of billable CPU and take
6 seconds to complete.   you have to play with the code and use logging
events to really know how much an operation will cost.

 

 

 

From: [email protected]
[mailto:[email protected]] On Behalf Of rmflow
Sent: Wednesday, February 02, 2011 4:35 AM
To: [email protected]
Subject: [google-appengine] response time vs cpu_ms

 

I'm trying to determine what is CPU Quota and how it is measured. I wrote a
simple app which loads CPU doing some Datastore writes and then looked at
results:


/worker 200 14107ms 367870cpu_ms 359470api_cpu_ms 


The application responded after about 15 seconds, so 14107ms is probably
response time. But then why cpu_ms is so high? It actually consumed 6 mins
of my CPU Quota. How do I compare CPU Quota and actual time spent by
application? If I know that my task will complete in 5 mins, how much CPU
Quota will it cost?

-- 
You received this message because you are subscribed to the Google Groups
"Google App Engine" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/google-appengine?hl=en.

-- 
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/google-appengine?hl=en.

Reply via email to