15:01:35 <skalnik1> #startmeeting Power management for Fedora
15:01:35 <zodbot> Meeting started Thu Sep 17 15:01:35 2009 UTC.  The chair is skalnik1. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:35 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:01:58 <skalnik1> Hello and welcome everybody
15:02:18 <skalnik1> I'd like to start one line recap of last week
15:02:34 <marcela> hello
15:02:40 <skalnik1> #topic Last week recap
15:03:34 <skalnik1> I was as a pm team member  on irc alone
15:04:28 <skalnik1> Phil Knirsch was on vacation the same this week
15:04:57 <skalnik1> The other people were on Development Conference here in Brno
15:06:08 <skalnik1> I informed about PM activity on the conference and  I promised to provide info about 'watt hunting' competition
15:06:22 <skalnik1> My promise is on the way :-)
15:07:03 <skalnik1> it should be available on our SIG sites next week and I'll put the link here next week
15:07:14 <skalnik1> well, what today
15:07:46 <skalnik1> #topic Test and results
15:08:43 <skalnik1> jvcelak did comparison test of scomes results for 32 and 64 bits
15:08:45 <skalnik1> os
15:09:21 <skalnik1> he prepares report of this test
15:09:32 <skalnik1> this will be on SIG too
15:10:02 <skalnik1> I'd like to mention shortly, that the test was done in virtual box
15:10:47 <plautrba> why not in kvm?
15:11:15 <skalnik1> his choice
15:11:38 <skalnik1> scomes results discover more ticks (~5 %) in 64 bit system
15:12:04 <skalnik1> i/o operations was almost the same (a bit effective 64 bit system)
15:12:33 <plautrba> ok. but it we want use scomes for meassuring on virtual system, we will probably use kvm as fedora default
15:12:56 <skalnik1> ok
15:13:35 <skalnik1> i think it will be also interesting information to compare virtbox and kvm
15:13:41 <plautrba> yes
15:14:19 <marcela> how many times did he measure one example?
15:15:15 <skalnik1> i'm not able to answer now but i asked him for writting this info into report
15:16:02 <marcela> ok
15:16:04 <skalnik1> there will be complete description of the test + results
15:16:44 <plautrba> great
15:17:16 <skalnik1> i started to test influence of compiz on system load
15:17:40 <skalnik1> this was only the first attempt and i didn't prepare test enough
15:17:50 <skalnik1> the results was the same
15:17:57 <skalnik1> for comiz on and off
15:18:08 <skalnik1> i tested it with bltk + open office
15:18:43 <skalnik1> the problem is that the features of compiz are not used enough to be seen
15:19:33 <skalnik1> quite more important test is doing jvcelak - live cd for test day
15:19:45 <skalnik1> but we have no results till now
15:20:07 <skalnik1> therefore i'd like to open a discussion how to continue
15:20:26 <marcela> I suppose compiz doesn't consume more energy when you don't move windows etc. which doesn't move with bltk...
15:21:10 <skalnik1> marcela: yes, therefore i wrote i didn't prepare test enough
15:21:39 <marcela> ok
15:21:56 <skalnik1> but i'm not sure if bltk provides enough features to simulate win moving etc.
15:24:05 <plautrba> does it use only keybord shortcuts?
15:25:29 <plautrba> i can move window in kde only by keyboard - alt-F3 <down><down><enter>and move with cursor keys
15:25:29 <skalnik1> probably yes, i didn't see any usage of window handling
15:25:35 <skalnik1> in the scenario
15:26:06 <skalnik1> the bash scripts uses some utility
15:26:23 <skalnik1> i have to investigate it a bit
15:26:26 <marcela> do we have some output with differences from bltk?
15:26:51 <marcela> I'd rather don't create new profile before we have some results from previous measurement
15:27:30 <skalnik1> i have a set of results
15:27:36 <marcela> great
15:28:31 <skalnik1> i used bltk with office workload
15:28:55 <skalnik1> each of situations was tested for 48 min
15:29:28 <skalnik1> the results contains wifi on & off
15:29:50 <skalnik1> wifi test was repeated for various state of battery
15:30:34 <skalnik1> this means test started when battery is fully charged and then the battery capacity is  60%
15:31:06 <skalnik1> there was no difference
15:31:38 <skalnik1> then i tested wifi in two distances between laptop and ap
15:32:33 <skalnik1> signal strenght 97% and ~50%
15:32:40 <skalnik1> no difference
15:33:23 <skalnik1> i plat to prepare test when the wifi will transfer some date but i have to eliminate other system activity (no writting to disk etc.)
15:35:42 <marcela> ok when we have some differences we can prepare some new profiles, maybe.
15:36:03 <skalnik1> yes
15:36:42 <marcela> OT I have note to developer documentation - the man who works on it is ddomingo, so we should write to him these kind of notes
15:40:13 <skalnik1> mracela, really ddomingo?
15:41:52 <marcela> yeah
15:45:53 <marcela> what about Test Day - I've started writing test examples
15:46:02 <marcela> I suppose it will be ready for Test Day
15:46:58 <skalnik1> great, i asked jscotka about inventory of test cases from last test day
15:47:26 <skalnik1> he did it and he should find a time to help us this week
15:48:00 <skalnik1> what about scomes testing?
15:48:50 <marcela> who does the testing with scomes?
15:49:07 <plautrba> we will test if scomes does same results on different systems
15:50:06 <plautrba> tests like measure rpmbuild of package should has almost same number or read/written bytes, syscall and so
15:51:02 <plautrba> i'll prepare some cases with jhutar next week
15:51:46 <skalnik1> ok
15:53:32 <skalnik1> plautrba: jscotka is ready to help
15:55:03 <plautrba> ok
15:55:41 <skalnik1> dpravec prepared test day site then we should move all info there: http://fedoraproject.org/w/index.php?title=Test_Day:2009-10-22
15:56:56 <marcela> it looks good
15:57:33 <marcela> let's summarize what is needed for next week
15:59:47 <skalnik1> 1. jvcelak will publish test report that is important for scomes test case
16:00:57 <skalnik1> 2. ivana promised putting info about competition on SIG
16:01:41 <skalnik1> 3. marcela will continue in test case for test day - is it, ok?
16:02:35 <marcela> aye
16:02:40 <skalnik1> 4. plautrba + jhutar will work on test case for scomes
16:02:50 <skalnik1> jscotka will help us
16:03:14 <skalnik1> 5. i'll take care about test day sites
16:03:41 <skalnik1> 6. then i'll contact last test day attendees
16:04:00 <skalnik1> i suppose this is enough and our time slot is over
16:04:23 <skalnik1> then any remarks?
16:04:37 <marcela> no
16:04:48 <plautrba> no
16:05:03 <skalnik1> so thanks everyone for joining and his activity and next week again with phil
16:05:10 <skalnik1> #endmeeting