Startup time for Address Book window is 1 second slower than 4.6 Here are our automated results: Address Book Startup Time 4.6: 7.347000 s 5.0: 8.414000 s System and Build: Win95, 166 MHz, 48mg RAM, LAN Build Number: 1999062808 Methodology: Measures the time it took to open up the Address Book window. No other application is running except the QA Partner client and Agent [used for automation]. Description Address Book is started by picking up the Address Book menu item.. Timer is started as soon as the menu item is selected and wait till Address Book window is loaded completely. Timer is stopped at this moment and the elapsed time is reported as the Address Book start up time. Pls let me know if you need more details. The difference between 4.6 and 5.0 is not that great so I've categorized the severity as normal.

Paul or Candice, we don't have any disk I/O around opening an AB yet, right? If that's true, I bet we get a bunch slower when we're actually reading the AB DB when we open it.

Is there a difference between the first time you open and the second time? Most of that difference could be due to the time it takes to load dlls. This time should be reduced a lot when we start re-basing dlls correctly (see bug # 7249) - note this is a windows only thing.

I'll find that out for you.

There is about a 1.2 second time difference with opening the address book two consecutive times in the same apprunner session. This is on the same machine which these performance timings are on. This is with today's release build.

Bringing up XUL windows in general is very slow. I my machine the compose window comes up even slower than the address book. We need to work on the overall performance of these dialogs. I guess we should talk to the XP group and possibly file a bug on their performance as well.

Have not had a chance to talk to the other groups about the speed.

What is the current status of the speed here. Seamonkey is much faster today than it was when this bug was written. Is this still slower than 4.x?

Will check next week when performance tests are run again.

Moving to post PR1.

I think this needs to be retested again. I'll ask Suresh to do so.

Suresh - can you post your latest numbers here and retest as needed? Thanks.

Here are some numbers collected using 1999-12-20-09-M13 windows commercial build. Bindu's automation script just loads the addressbook window using 'mozilla -addressbook'. NOTE: Slight change in the system tested. The system I tested is Win95, 166 MHz, 64 MB RAM. (Originally it states 48 MB RAM). Method 1: Using 'mozilla -addressbook'. Intial launch: 16.791 seconds Re-launch: 15.643, 15.769, 15.768 seconds(3 runs). Method 2: Using 'Task | addressbook' after loading the 3-pane mail window. (This is done manually using a stop watch. So there can be little room for error). Initial launch: 3.28 seconds. Re-launch: 1.75, 1.81, 1.81 seconds. (3 runs)

Bulk add of "perf" to new keyword field. This will replace the [PERF] we were using in the Status Summary field.

Bumping to M15

Can we time this again? I would be interested in the current comparison.

Suresh - when you have a moment, can you get results for Paul? We don't show this in our performance data. Thanks.

Using 2000-02-07-09-M14 windows commercial build. Win95, 133 Mhz, 64 MB RAM. 1. Startup time for Address Book, using Task | Messenger 5.74, 5.39, 5.22 seconds. (3 runs) 2. Startup time for Address Book, using mozilla -addressbook 17.08, 16.30, 16.54 seconds.

Mass moving to M17

Mass move mailnews bugs to Putterman. Ouch.

marking WONTFIX.

oops. that was supposed to get a future, not a wontfix.

reassigning to chuang.

Anything need to be done with this bug?

I'd like to get this tested and worksforme/verified, but I need some additional information. Is this just testing a blank addressbook? If so, I can get this done today.

This is fairly old bug! I think testing an address book with entries would be good. What is the size of your address book? I think Esther has an address book of 800 entries.

Only ~100 entries. For testing this, do you remember the criteria (addressbook size) for the 1st time we performance tested this? Otherwise, there wouldn't be a consistency, resulting in an inaccurate metric.

I cannot recall. suresh?

Would running this on a P133 with 64 megs count? With an empty address book? (It's as close to the original filing/data that I think we can get.)

Update, running a P133 with 64 megs of RAM, with no collected/personal address books on Windows 98 took 8.41 seconds to load using Tasks | Address Book.

*** Bug 89066 has been marked as a duplicate of this bug. ***

reassigning to cavin

*** This bug has been marked as a duplicate of 117863 ***

Gag, I'm sorry. Loading ! = startup. Sigh.

taking, as it's nab perf and I'm in there now. donner, is 1 second correct? what are you numbers on the test machine?

I've uploaded the latest metrics to the url above (http://www.mozilla.org/mailnews/performance/abookperf.html), but 4.7x shows us at .57 seconds to launch from the Communicator menu (similar to our implementation of the Tasks menu). Outlook Express takes roughly 0.60 seconds. The latest build of 6.x (2002-02-17-08) takes 3 sec 28 mil. I think we may be closer on raw launch times (each component's executable, rather than from a menu in our existing apps). I'll report back with those figures, too.

I thought you could run the 4.7x client's address book as a standalone app under Win32, but I'm mistaken. I also can't get data at the moment on our perf machine, because it has other profiles for perf (page load, etc) that I can't wipe out. Anyways, their stand-alone app (OE 6) wab.exe, takes 0.67 seconds to launch.

I'm marking this as wontfix as after discussion on IRC, there isn't too much concern over this at the moment. If it starts looking bad we can always reopen.

Verified (at least for now).