Hey Steve, Yea, the elapsed time is usually higher than CPU usage, but I think maybe because I was doing my testing late last night I lucked into really low elapsed time too. I checked my cold start time this morning, and elapsed time was 2867ms with 1010cpu_ms.
Here are some lines from my logs: 03-20 08:30AM 05.838 /donothing 200 2867ms 1010cpu_ms 0kb Mozilla/5.0 03-20 12:15AM 00.832 /donothing 200 2155ms 913cpu_ms 0kb Mozilla/5.0 03-19 11:02PM 28.449 /donothing 200 1119ms 1166cpu_ms 0kb Mozilla/5.0 The elapsed time seems to vary greatly, whereas cpu time seems to be a decent metric to see if an optimization of mine actually had an effect. The page I request for testing is a servlet which returns nothing. Note that if you are requesting a JSP it will increase your cold start times by at least a few hundred milliseconds. For some reason the first access to even the most simple JSP (like <%= 5+2 %> will take quite a while. I'd really love to get my elapsed time down to around the same amount as my CPU time. Does anyone have any insights as to what exactly App Engine is doing during that time? On Mar 20, 7:30 am, Steve Pritchard <[email protected]> wrote: > Update to previous post. > I am measuring elapsed time (what I really care about), not CPU secs. > I could not figure out how a cold start was so fast (1200ms). > Steve > > On Mar 20, 10:26 am, Steve Pritchard <[email protected]> wrote: > > > > > Thanks for the tip. I am in the same boat in that I do not use JDO > > etc. I also do not need GWT stuff. So I deleted: > > > datanucleus-appengine-1.0.5.final.jar > > datanucleus-core-1.1.5.jar > > datanucleus-jpa-1.1.5.jar > > geronimo-jpa_3.0_spec-1.1.1.jar > > geronimo-jta_1.1_spec-1.1.1.jar > > gwt-servlet.jar > > jdo2-api-2.3-eb.jar > > > from my 1.3.1 war\lib and it keeps running. My 'warm' start (using > > MemCache) is around 3200ms. Previously I recall it fluctuated around > > 4000ms. > > Steve > > > On Mar 20, 2:09 am, Spines <[email protected]> wrote: > > > > Well, I figured it out, I just deleted all of the jars I didn't need > > > from my ./build directory. I deleted all of the JDO, JPA, and > > > datanucleus jars. The result of doing this was I got my first cold > > > start that happened under 1200ms :). Previously the CPU time used by > > > a cold start was pretty consistently around 1500ms, and after deleting > > > those jars it is now around 1150ms :). -- You received this message because you are subscribed to the Google Groups "Google App Engine for Java" 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-java?hl=en.
