Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2010-12-02

00:00:59 Quit domonoky1 (Read error: Connection reset by peer)
00:05:52 Quit bertrik (Quit: :tiuQ)
00:06:02 Quit Kupop (Ping timeout: 240 seconds)
00:07:45 Quit kugel (Remote host closed the connection)
00:10:16 Quit stripwax (Quit:
00:12:07 Join Zarggg [0] (
00:12:44 Join JdGordon| [0] (
00:12:44 Quit JdGordon| (Changing host)
00:12:44 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
00:17:40 Quit n1s (Quit: Lämnar)
00:23:18 Quit tchan (Read error: Connection reset by peer)
00:24:37***Saving seen data "./dancer.seen"
00:24:43 Quit ReimuHakurei (Read error: Connection reset by peer)
00:24:58 Join ReimuHakurei [0] (~reimu@
00:25:05 Join stripwax [0] (
00:26:01 Quit parafin (Remote host closed the connection)
00:26:14 Join parafin [0] (
00:27:18 Quit robin0800 (Remote host closed the connection)
00:29:06 Quit mt (Ping timeout: 255 seconds)
00:30:57 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
00:38:01 Quit casainho (Remote host closed the connection)
00:48:17 Quit Jerom (Quit: Leaving.)
00:53:41 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
01:04:39 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
01:06:09 Quit BHSPitMonkey (Remote host closed the connection)
01:11:36 Quit Keripo (Quit: Leaving.)
01:12:24 Part toffe82
01:12:36 Quit pamaury (Read error: Connection reset by peer)
01:13:22 Quit evilnick_B (Quit: Page closed)
01:14:18 Quit dfkt (Quit: -= SysReset 2.53=- Sic gorgiamus allos subjectatos nunc.)
01:17:27 Join Judas_PhD [0] (
01:19:46 Quit stripwax (Quit:
01:20:59 Join BlakeJohnson86 [0] (
01:25:19 Quit GeekShadow (Quit: The cake is a lie !)
01:28:15 Quit Guinness (Read error: Connection reset by peer)
01:32:32 Join Keripo [0] (
01:38:55 Join Llorean1 [0] (
01:40:39 Quit Llorean (Ping timeout: 255 seconds)
01:43:23JdGordon|anyone still awake that can do me a logf build?
01:46:52 Join madalu [0] (~user@unaffiliated/madalu)
01:50:48 Join Guinness [0] (
01:54:29JdGordon|Has anyone else seen playback suddenly stop but the UI keeps working untill you try to actually stop playback?
02:00:35 Join Strife1989 [0] (
02:01:10Strife1989JdGordon|: Here and there, but not often enough or consistent enough to file a bug report.
02:01:25 Quit Strife89 (Disconnected by services)
02:01:30 Nick Strife1989 is now known as Strife89 (
02:06:04 Quit madalu (Ping timeout: 240 seconds)
02:24:38***Saving seen data "./dancer.seen"
02:29:23 Join noamsml_ [0] (
02:31:18 Quit noamsml (Ping timeout: 250 seconds)
02:44:50 Quit froggyman (Remote host closed the connection)
02:51:47 Join Keripo1 [0] (
02:53:15 Join eWill [0] (
02:54:39 Quit Keripo (Ping timeout: 272 seconds)
02:54:42eWillI have a c200 v1. I just turned it on, and RB says battery is at 23%. Then for about 1.5 minutes the % is slowly rising till it reaches 50%. Is that normal? Perhaps it's because I'm in a cold room???
02:57:42 Quit UnclePervyJesus (Read error: Connection reset by peer)
02:57:59 Join UnclePervyJesus [0] (~ubernoob@2001:c08:3700:ffff::17:ff75)
02:58:01Strife89eWill: It may well be the cold temperature.
02:58:51 Quit Judas_PhD (Ping timeout: 255 seconds)
03:00:57 Nick noamsml_ is now known as noamsml (
03:04:16 Nick Llorean1 is now known as Llorean (
03:04:29 Quit Llorean (Changing host)
03:04:30 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
03:04:39 Join krabador [0] (
03:05:31 Quit MethoS- (Remote host closed the connection)
03:11:26 Quit eWill (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
03:13:09 Join Judas_PhD [0] (
03:17:43UnclePervyJesushaving a problem starting the speed test, cant find "test codec" in the openwith dialog
03:23:16 Quit Llorean (Quit: Leaving.)
03:24:22 Join ZhangNing [0] (~ZhangNing@
03:29:57 Join Llorean [0] (
03:30:03 Quit Llorean (Changing host)
03:30:03 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
03:31:30 Join JdGordon1 [0] (~jonno@
03:33:18 Quit designate72 (Ping timeout: 245 seconds)
03:34:20 Part ZhangNing
03:34:35 Quit JdGordon| (Ping timeout: 250 seconds)
03:38:37 Join designate72 [0] (
03:49:07 Join Chesteta [0] (
03:49:33Chestetahello,isewill here?
03:50:17Chestetamessed up spacebar, sorry
03:51:25 Join LambdaCalculus37 [0] (
03:51:25 Quit LambdaCalculus37 (Changing host)
03:51:25 Join LambdaCalculus37 [0] (~rmenes@rockbox/staff/LambdaCalculus37)
03:52:55ChestetaI was wondering what luck you have been having with your fuzev2; have you achieved usb transfer from within rockbox?
03:54:56 Quit designate72 (Ping timeout: 265 seconds)
03:54:59 Quit Chesteta (Quit: CGI:IRC (EOF))
03:59:16 Quit Judas_PhD (Ping timeout: 272 seconds)
04:06:23 Quit InsDel (Read error: Connection reset by peer)
04:08:07 Quit krabador (Quit: Sto andando via)
04:11:09 Join Judas_PhD [0] (
04:13:44 Quit JdGordon (Ping timeout: 272 seconds)
04:15:18 Quit TheSeven (Ping timeout: 245 seconds)
04:19:09 Join DerPapst1 [0] (
04:20:34 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:20:47 Quit DerPapst (Ping timeout: 240 seconds)
04:24:41***Saving seen data "./dancer.seen"
04:29:54 Quit LambdaCalculus37 (Quit: Fwump)
04:30:52 Part UnclePervyJesus
04:31:07 Join UnclePervyJesus [0] (~ubernoob@2001:c08:3700:ffff::17:ff75)
04:44:19 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:44:20 Quit amiconn (Disconnected by services)
04:44:20 Quit pixelma (Disconnected by services)
04:44:21 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:44:23 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:44:38 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:50:02 Quit sasquatch (Quit: WeeChat 0.3.2)
04:50:27 Join sasquatch [0] (
05:03:46 Nick JdGordon1 is now known as JdGordon (~jonno@
05:08:21 Quit rodan ()
05:25:27 Quit Horscht (Quit: Verlassend)
05:26:54 Quit anewuser ()
05:27:38 Join zeleftikam [0] (~zeleftika@unaffiliated/zeleftikam)
05:35:16 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
05:40:28 Quit JdGordon (Read error: Connection reset by peer)
05:40:56 Join JdGordon [0] (
05:40:57 Quit JdGordon (Changing host)
05:40:57 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
05:52:45 Join mordocai [0] (~mordocai@
05:55:03 Quit mordocai (Quit: reboot for kernel update)
05:59:36 Join mordocai [0] (~mordocai@
06:03:00mordocaiAm I right in thinking that changing the hotkeys for my sansa clip+ would require a source code edit?
06:05:36JdGordondepends on what you want to change
06:07:04mordocaiI want to change Long Home to Unlock/Lock and change Home + Select to shortcut menu.
06:07:13JdGordonthen yes
06:07:32mordocai(easier to do one handed while it is in my pocket :P)
06:10:35zeleftikamany idea how to Quit an application on an iPod 5g?
06:10:42zeleftikami'm stuck in the Calendar!
06:12:15 Quit DerPapst1 (Quit: Leaving.)
06:13:08mordocaizeleftikam: I don't really know which one is the 5g(don't use apple devices) but might try menu+select (found here:
06:13:26zeleftikamah! that worked! thanks so much mordocai!
06:13:35zeleftikami swear i tried every other combination :P
06:13:49mordocaizeleftikam: Of course :P. Always works that way lol
06:14:10mordocaizeleftikam: In any case, the applicatin section of the manual for your device should have all the hotkeys for the apps
06:15:58mordocaizeleftikam: Sorry, the actual section in table of contents is "Plugins". Applications is a subsection
06:16:56mordocaiBy the way, to the developers, I'm very impressed with the comprehensive documentation that is provided for all the ports. Very good quality IMHO.
06:17:06zeleftikamyeah i was surprised to even hear of any
06:18:24mordocaispeaking of that, I just found a typo in the clip+ manual.
06:18:30*mordocai goes to find out how to fix manual typos
06:20:12JdGordonfile a bug usually.. or just tell someone here
06:20:21JdGordonwhats the typo?
06:21:40mordocaiOh, I was going to go fix it myself lol. The keymap for the brickmania game plugin is only half showing up. What the keys do is showing up, but not which keys do them.
06:22:10mordocaiLooks like there are a few like that, actually
06:22:12JdGordonthats more than a typo
06:22:37mordocaiYeah, more like someone just hasn't taken the time to add them all to the docs yet.
06:24:42***Saving seen data "./dancer.seen"
06:25:40mordocaiWell, in my free time i'll go ahead and start adding the missing hotkeys to the manual. Would it be better to submit one patch per a plugin, or one patch when i'm done updating a section (like all games, or all applications, etc.)
06:35:37 Join UNHchabo [0] (
06:38:36UNHchabohey all, I have a Sansa E200v1, and it started locking up the other day after I updated the firmware. I tried a firmware downgrade, and other updates, but that didn't fix the issue.
06:39:19UNHchabothe device still works with the default firmware
06:40:23UNHchaboI turn the device on, and the music starts playing, but if I try to do anything except navigate through the menu, it locks up
06:40:58UNHchaboif I don't touch it after powering it on, then it locks up 2 seconds from the end of the track
06:42:52UNHchaboif I delete that song, then the same thing will happen with the next track on the playlist
06:42:56mordocaiUNHchabo: Well, I am new to rockbox so take the suggestion in that context, but have you tried deleting the .rockbox folder and reinstalling the firmware? Might be some kind of config issue.
06:43:28UNHchaboalright, I can try that, and get back
06:43:45mordocairockbox firmware*
06:46:36 Join Strife1989 [0] (
06:48:25 Quit Strife89 (Disconnected by services)
06:48:28 Nick Strife1989 is now known as Strif8e9 (
06:48:33 Nick Strif8e9 is now known as Strife89 (
06:50:50 Quit Loto (Quit: Loto)
06:51:14 Quit factor (Read error: Operation timed out)
07:06:00 Join ReimuHakurei_ [0] (~reimu@
07:07:35 Quit ReimuHakurei (Read error: Connection reset by peer)
07:21:34 Join Loto [0] (
07:21:34 Quit Loto (Changing host)
07:21:34 Join Loto [0] (~nfs@xbmc/user/Loto)
07:29:47 Join factor [0] (
07:32:27UNHchabomordocai, that worked, I just wish I didn't have to delete the directory ;)
07:32:29UNHchabothanks though
07:32:57mordocaiUNHchabo: Yeah, there may have been an less invasive way... but that was the first thing i thought of. Glad it is working again!
07:34:01 Quit factor (Read error: Operation timed out)
07:37:38 Quit jepler (Read error: Operation timed out)
07:38:04 Join hebz0rl [0] (
07:42:11 Join mystica555_ [0] (
07:45:03 Join factor [0] (
07:52:16 Join ender` [0] (
07:53:36 Join LinusN [0] (~linus@rockbox/developer/LinusN)
08:03:13 Quit JdGordon (Quit: leaving)
08:11:20 Join JdGord [0] (
08:24:43***Saving seen data "./dancer.seen"
08:36:10 Part UNHchabo
08:44:13 Quit JdGord (Quit: Bye)
08:45:10 Join xxcv [0] (
09:03:01 Join stripwax [0] (
09:03:12 Quit stripwax (Client Quit)
09:04:42 Quit iq (Ping timeout: 250 seconds)
09:13:32AlexPmordocai: More than one plugin in a patch is fine by me, but I'd prefer any e.g. white space changes to be separate. It makes it difficult to review otherwise
09:15:40 Join Rob2222 [0] (
09:18:25 Join einhirn [0] (
09:19:00 Quit chattr (Ping timeout: 250 seconds)
09:19:11 Quit Rob2223 (Ping timeout: 265 seconds)
09:20:31 Join chattr [0] (
09:20:43mordocaiAlexP: Yeah, there shouldn't be any whitespace or anything. I'll just be adding in the correct \opt directives to get the correct keys to show up in the tables.
09:20:53AlexPsounds good!
09:23:49 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
09:24:47 Join jepler [0] (~jepler@emc/developer/pdpc.professional.jepler)
09:26:20 Quit Strife89 (Ping timeout: 255 seconds)
09:26:46 Join Strife89 [0] (
09:29:42*mordocai just realized that the little icon that he couldn't figure out was saying that the SD card was being used...
09:34:00 Join factor_ [0] (
09:34:27 Quit factor (Read error: Connection reset by peer)
09:35:22 Quit factor_ (Remote host closed the connection)
09:35:36AlexPor internal memory
09:35:36 Join factor [0] (
09:35:45AlexP"disk" access basically :)
09:36:12mordocaiOh, okay. I just noticed it was staying on during database update. Before it would just flash occaisionally, so I was confused lol
09:36:30 Quit chattr (Ping timeout: 245 seconds)
09:36:45Zagorsoap: link fixed
09:43:06 Join efyx [0] (
09:43:16 Join swilde [0] (
09:48:03 Join petur [0] (d408b802@rockbox/developer/petur)
09:53:18 Join MethoS- [0] (~clemens@
09:53:19 Join amee2k [0] (
09:58:08 Quit MethoS- (Remote host closed the connection)
09:58:25 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
09:58:57 Quit zu (Ping timeout: 240 seconds)
09:59:09 Join _jhMikeS_ [0] (~jethead71@
09:59:16 Quit _jhMikeS_ (Changing host)
09:59:17 Join _jhMikeS_ [0] (~jethead71@rockbox/developer/jhMikeS)
09:59:17 Quit jhMikeS (Disconnected by services)
09:59:21 Join Kitr88 [0] (
10:02:08 Quit Kitar|st (Ping timeout: 240 seconds)
10:03:39mordocaiBy the way, using the current build(i was using 3.7.1 before) i'm still getting the error shown here: when adding songs to a dynamic playlist that already has songs in it. Still appears to be happening intermittently, I can't seem to reproduce it too reliably. Any ideas on how to debug this?
10:03:57mordocaiThat's with the sansa clip+
10:04:28 Quit Kitr88 (Ping timeout: 276 seconds)
10:04:29 Quit TheSeven (Ping timeout: 240 seconds)
10:05:04Stummii belive that i readed something like this a few days ago in the bugtracker
10:06:40mordocaiOkay, i'll look through the bugs manually real quick. (I tried a search, but failed)
10:06:47Stummiah yes, #11655
10:08:41 Join Kitar|st [0] (Kitarist@
10:09:25mordocaiYep, that it is! Thanks. I'll have to comment to say that it is doing it on my clip+ with or without a cable plugged in :P
10:09:29mordocaiis it*
10:09:43 Join zu [0] (
10:11:20 Join robin0800 [0] (
10:13:23Stummimordocai, yes, good idea
10:14:32Stummihm, i think it would make sense to set priority for #11655 to high, or?
10:15:13Stummierm, medium
10:15:14pixelmaso far, the priority hasn't been payed much attention
10:15:17ZagorStummi: the priority field isn't really used
10:15:25Stummihm, ok :)
10:15:55mordocaiStummi: I'd agree at least medium if priority field was being actively used though :P. Kind've annoying to not be able to quickly make a nice playlist.
10:16:54Stummiyes. This is a realy serve bug for rockbox imho
10:24:46***Saving seen data "./dancer.seen"
10:40:25 Quit xxcv ()
10:53:05 Quit liar (Ping timeout: 255 seconds)
10:53:35 Join liar [0] (
11:05:19 Quit factor (Ping timeout: 245 seconds)
11:07:21 Join xxcv [0] (
11:10:41 Join pamaury [0] (
11:10:41 Quit pamaury (Changing host)
11:10:41 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:12:39 Join factor [0] (
11:17:22pixelmacan someone ban this wanxin511 - deleting his spam currently but am not sure if I'm allowed to ban (it might be a one time thing though)
11:17:27pixelmain the forums
11:20:00 Quit user890104 (Ping timeout: 272 seconds)
11:24:35 Join swilde` [0] (
11:33:39 Join iq [0] (~iq@unaffiliated/iq)
11:34:40 Quit Judas_PhD (Quit: This is a quitting message)
11:42:51 Quit ReimuHakurei_ (Read error: Connection reset by peer)
12:18:49Torne;u=27931 also needs banning
12:18:53TorneAlexP: poke
12:20:56 Quit xxcv ()
12:24:48***Saving seen data "./dancer.seen"
12:33:13 Join InsDel [0] (
12:40:30 Quit efyx (Remote host closed the connection)
12:41:07 Join efyx [0] (
12:48:02 Join panni_ [0] (
12:48:58 Join casainho [0] (
12:55:37 Quit InsDel (Read error: Connection reset by peer)
13:05:16 Quit antil33t (Read error: Connection reset by peer)
13:05:27 Join antil33t [0] (
13:08:06 Join mt [0] (~mt@
13:13:09 Join chattr [0] (
13:13:29 Join xxcv [0] (
13:15:37 Quit BlakeJohnson86 (Ping timeout: 250 seconds)
13:16:24 Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:16:45 Quit antil33t (Read error: Connection reset by peer)
13:16:54 Join antil33t [0] (
13:20:40AlexPTorne: Done, ta
13:21:08 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
13:23:35 Join BlakeJohnson86 [0] (
13:24:29 Quit chattr (Remote host closed the connection)
13:24:49 Quit casainho (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101027124101])
13:30:47 Join chattr [0] (
13:32:23 Quit bluebrother (Disconnected by services)
13:32:25 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
13:32:52 Join JdGord [0] (
13:36:20 Quit BlakeJohnson86 (Ping timeout: 260 seconds)
13:47:26 Quit Rondom (Ping timeout: 276 seconds)
13:51:52 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:51:54 Quit JdGord (Read error: Connection reset by peer)
13:52:37 Quit sinthetek (Read error: Operation timed out)
14:02:23 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
14:02:57mc2739can a forum admin ban a spammer? only7youz
14:03:46Stummimc2739, ask AlexP
14:03:58Stummihm, there seems to be many spammers today
14:07:52gevaertsor Bagder, soap or GodEater
14:07:58gevaerts(I think)
14:08:41 Join Rondom [0] (
14:17:51 Quit factor (Ping timeout: 255 seconds)
14:24:50***Saving seen data "./dancer.seen"
14:45:46 Quit robin0800 (Remote host closed the connection)
14:53:53 Join factor [0] (
14:58:33 Nick Strife89 is now known as Desk-Strife89 (
15:00:30 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:08:44 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:11:14 Quit xxcv ()
15:20:12 Join user890104 [0] (
15:24:01 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
15:28:07 Join sinthetek [0] (
15:28:07 Quit sinthetek (Changing host)
15:28:07 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
15:30:53 Quit z35 (Ping timeout: 265 seconds)
15:31:43 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
15:32:31 Join anewuser [0] (anewuser@unaffiliated/anewuser)
15:32:45 Join daggett [0] (
15:32:59daggetthi all
15:33:49daggettcan anyone tell me if the support for RTC mod on IHP120 is compiled in the standard stable version rockbox ?
15:34:36daggettI just finished the mod and now wants to use it, but I can't find any menu item to set it or display it
15:34:40 Quit swilde (Quit: ERC Version 5.3 (IRC client for Emacs))
15:35:13gevaertsit's not
15:35:23 Quit swilde` (Quit: ERC Version 5.3 (IRC client for Emacs))
15:35:38gevaertsIn general we assume that if you can manage to do a hardware mod, you can manage to compile too
15:39:34 Join MethoS- [0] (~clemens@
15:41:16 Join bmbl [0] (
15:41:16 Quit bmbl (Changing host)
15:41:16 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
15:43:26daggettgevaerts: ok, thanks, so, I will try to compile it then
15:58:44 Join DerPapst [0] (
16:12:27 Quit kugel (Ping timeout: 265 seconds)
16:20:29 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
16:24:52***Saving seen data "./dancer.seen"
16:28:36 Join mattymo [0] (~mmosesoh@nat/redhat/x-juqdfmlnakujqcng)
16:35:50 Part Zagor
16:36:50 Quit petur (Ping timeout: 265 seconds)
16:43:00 Join toffe82 [0] (
16:46:37daggettok, I compiled and installed the new RTC-enabled rockbox but clock and alarmclock plugins say : "incompatible version"
16:47:08 Part LinusN
16:48:01gevaertsSounds like you didn't install properly
16:48:19evilnick_BThat's a sign of the plugins being from the previous installation
16:59:27 Quit mortalscan (Remote host closed the connection)
16:59:43 Join mortalscan [0] (~mortalsca@
17:11:33 Join n1s [0] (
17:11:33 Quit n1s (Changing host)
17:11:33 Join n1s [0] (~n1s@rockbox/developer/n1s)
17:18:03daggettgevaerts: evilnick: ok, so maybe I need the fullzip version then
17:21:14 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.
17:21:50saratogaAlexP: can you ban UGG603898
17:22:12gevaertsDo those users come from different IPs?
17:23:14saratogalots of audio sites getting slammed with spam lately
17:23:48 Join ReimuHakurei [0] (~reimu@
17:24:11AlexPgevaerts: Sadly yes
17:24:17daggettthanks a lot, it was my file-roller that was incorrectly configured to NOT erase existing files : damn GUI !
17:26:05AlexPgevaerts: completely different subnets too
17:29:52gevaertsAlexP: I seem to always getting the exact same captcha on the registration page
17:30:27AlexPgevaerts: You'll have to talk to scorche about that, I can't get to that bit
17:31:06gevaertsI know, but it might explain the problem
17:31:14LloreanIt's been that way for years, I think
17:31:29AlexPyeah, if it always the same it doen't help fight spammers
17:32:07gevaertsGood for support though
17:35:11 Quit simonrvn (Read error: Connection reset by peer)
17:45:16 Join simonrvn [0] (
17:50:19 Join TheLemonMan [0] (
17:56:40 Join Buschel [0] (
18:03:46 Quit liar (Ping timeout: 255 seconds)
18:04:55 Quit ReimuHakurei (Ping timeout: 245 seconds)
18:11:08 Quit TheLemonMan (Quit: free(me))
18:11:43 Quit riotz (Remote host closed the connection)
18:12:19 Join liar [0] (
18:13:55 Quit Buschel (Ping timeout: 264 seconds)
18:15:48 Join Judas_PhD [0] (
18:17:20 Join casainho [0] (
18:21:09 Join TheLemonMan [0] (
18:24:53***Saving seen data "./dancer.seen"
18:25:20 Join kugel [0] (~kugel@rockbox/developer/kugel)
18:28:21 Quit TheSeven (Ping timeout: 240 seconds)
18:30:55 Join Buschel [0] (
18:32:08kugelBuschel: are your cache align patches in svn now?
18:35:18Buschelyes, you can use MEM_ALIGN_ATTR now for aligning
18:35:50Buschelseveral codecs use it as well, and there is something to test for you :) ->
18:36:31Buschelthe patch changes a52, aac-he and wmapro (iram + aligning)
18:36:40kugelI'm looking at test codec, and I'm curious why my phone is much faster in some codecs (and the results are older) than the beast
18:37:10kugelI'll commit enable-asm for android, then give your patch a test
18:38:16Buschelfine :)
18:44:54kugelboth are arm11, but my phone should be slowed down by OS preemption and java layer, but it's still a lot faster in most codecs
18:50:51Buschelmaybe ram/iram is faster, you could take a look at it with test_mem plugin
19:03:01preglowanyone ever look at the ffmpeg aac decoder properly?
19:03:14 Join ReimuHakurei [0] (~reimu@
19:04:44 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
19:06:24 Quit factor (Read error: Connection reset by peer)
19:06:35 Join factor [0] (
19:07:33Buschelkugel: 8MB/s? that is quite few. boostes or unboosted?
19:07:52kugelRaaA doesn't control boosting
19:08:21kugelI assume that's boosted, as busy looping puts load on the cpu and the OS boosts
19:12:04kugelBuschel: ah no wait, it gives strange numbers
19:12:36kugelrd/wr is 200-266MB/s, st is 400MB/s and cp is 400-800MB/s
19:12:59kugelwhen I touch the screen all are 800MB/s
19:14:05kugel(8MB/s was what I remembered wrongly from running it the last time)
19:18:55kugelBuschel: the asm in test_mem is wrong I think
19:21:07 Quit Keripo1 (Ping timeout: 264 seconds)
19:22:34 Quit MethoS- (Remote host closed the connection)
19:24:13 Join bertrik [0] (
19:24:20 Quit bertrik (Changing host)
19:24:20 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:27:27Buschelkugel: it works for all other targets and gives reasonables numbers
19:28:00kugelit substracts 8 from the address each loop, but reads/writes 16 bytes
19:28:25kugelthe >>21 in the memcpy test is also wrong I think
19:28:28Buschellet me check
19:32:20Buschelthe loops are correct from what I see. the loops do not decrease adress, but the number of registers written or read.
19:32:39Buschelit does the same as the c-implementation below
19:33:54Buschel>>21 is also correct as memcpy is called for half buffer size (see line 68-73)
19:35:22kugelbut the loop count is doubled
19:35:49kugelthe C implementation does 4 rd/wr per loop, the asm does 8
19:35:59kugel(hence my confusion about 16 bytes)
19:36:39Buschelkugel: correct. you confused with your memcpy statement (need to check again)
19:37:44 Quit benedikt93 (Quit: Bye ;))
19:38:15kugelI find it confusing that buf_size is not in bytes unit
19:40:59 Join JesusFreak316 [0] (
19:41:09 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
19:41:38gevaertsAlexP, soap, Bagder, GodEater: more banning needed (hair06ij)
19:41:51AlexPffs, what is going on today?
19:43:30Buschelkugel: you might be right about >>21.
19:45:32*gevaerts wishes he knew
19:46:11 Join ReimuHakurei_ [0] (~reimu@
19:47:24 Join Keripo [0] (
19:47:38 Nick YPSY is now known as Ypsy (
19:47:45 Quit ReimuHakurei (Read error: Connection reset by peer)
19:49:32 Quit sinthetek (Ping timeout: 250 seconds)
19:49:35 Join stripwax [0] (
19:50:27Buschelkugel: yes, it is wrong. thank god it only impacts the printf :)
19:50:59kugelI have a patch for test_mem for fixing a few things, I will test and commit
19:51:09Buscheljust do it :)
19:53:11kugelI'll need a higher loop count for reasonable numbers on my phone
19:53:51gevaertshm, would a system to warn us when a newly registered forum user posts more than e.g. three times in ten minues be hard to build?
19:54:49Buschelit's a shame the S5L has such a slow ram... PP5022 (at same clock) has 1.5x faster iram read, 3x faster iram write, 1.1x faster dram read and 50% faster dram write
19:55:09Buschel(using the sped up DRAm setting for S5L)
19:55:50 Nick tmzt_g2root is now known as tmzt (~tmzt@
19:56:13 Quit stripwax (Quit:
19:56:58 Quit factor (Read error: Connection reset by peer)
19:57:09 Join factor [0] (
19:58:01 Quit factor (Read error: Connection reset by peer)
19:58:41 Join factor [0] (
20:07:47 Join jgarvey [0] (
20:19:12 Join InsDel [0] (
20:20:34kugelBuschel: ok, on my fuze it's 50/114/114/72 now
20:20:43kugelmemcpy shouldn't be faster than read
20:21:36 Quit bmbl (Quit: Verlassend)
20:23:55 Quit casainho (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101027124101])
20:24:31 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
20:24:31Buschelthe read measurement looks strange in relation to the other...
20:24:55***Saving seen data "./dancer.seen"
20:26:28Buschelkugel: was this iram or dram, boosted or unboosted?
20:27:04Buschelat which clock speed for the dram?
20:27:25 Join b0hoon [0] (
20:29:38kugelsomething was wrong with my built I think
20:29:51kugelI couldn't boost or unboost
20:31:30saratogapreglow: i read through it a while ago
20:31:38saratogabut didn't do anything else with it
20:32:24kugelhm, cpu freq scaling seems broken on fuzev2
20:33:22kugelah oops
20:33:32kugelit's not activated on amsv2, right?
20:33:38 Join ReimuHakurei__ [0] (~reimu@
20:35:09 Quit ReimuHakurei_ (Read error: Connection reset by peer)
20:37:49kugelBuschel: I get the same numbers in svn though
20:38:22Buschelkugel: mayb eeffect of caches?
20:42:16n1sBuschel: do you still need testing of that patch?
20:42:31Buschelyes, if you still have some time?
20:45:25 Join Horscht [0] (
20:45:26 Quit Horscht (Changing host)
20:45:26 Join Horscht [0] (~Horschti@xbmc/user/horscht)
20:45:50 Quit Keripo (Quit: Leaving.)
20:47:34Buschelsaratoga: your change to libmad for arm9 saves 1.2 MHz on nano 2g on a 200kbps sample
20:48:08saratogaBuschel: huh thats less then I measured
20:48:30saratogakugel: yeah its off, turning it on is generally stable but some people reported occasional problems with it enabled
20:48:55saratogai think i got just less then 2MHz on my nano2g
20:50:01 Join b0hoon_ [0] (
20:50:29*pamaury finds the imx23 power system quite scary
20:50:43 Quit b0hoon (Ping timeout: 255 seconds)
20:51:18 Join riotz [0] (riot33r@gateway/shell/
20:54:30 Quit elcan (Ping timeout: 265 seconds)
20:57:02 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
20:57:11 Quit dfkt (Read error: Connection reset by peer)
20:58:29 Join T44 [0] (
21:00:44kugelBuschel: my patch will also increase the loop count if its too low to be accurate (<10)
21:01:32n1sBuschel: so same paste as yesterday, beast and coldfire, speed vs svn, aache, a52, wmapro?
21:01:48Buscheln1s: yes
21:01:51 Join Keripo [0] (
21:01:57 Quit Topy (Ping timeout: 245 seconds)
21:02:00Buschelkugel: it will do this dynamically?
21:02:51Buschelkugel: cool
21:04:34 Nick dfkt_ is now known as dfkt (dfkt@unaffiliated/dfkt)
21:05:11n1sBuschel: if you are going to keep optimizing the codecs like this you should ask the fund for a coldfire based target :)
21:06:12*Buschel 's wife would kill him (or at least destroy such a target)
21:07:28CIA-7New commit by kugel (r28720): Some improvements for test_mem: ...
21:09:18 Join LambdaCalculus37 [0] (
21:09:47CIA-7r28720 build result: All green
21:12:07 Join Kupop [0] (
21:14:43 Nick Desk-Strife89 is now known as Strife89 (
21:15:25 Quit LambdaCalculus37 (Quit: go time)
21:15:48kugelBuschel: 440/550/600-1000/600-850 now on my phone
21:17:43Buschelkugel: this should explain why your phone is so much faster. i cannot image that the beast's ram is that fast
21:18:24Torneyeah, the beast is probably pretty old sdram tech, but cellphones are typically DDR2 or whatever these days
21:18:35kugelthe loop transfer 192MB instead of 8MB as previously because of the dynamic loop count increment
21:19:20kugelthe mem* results are still varying a lot though
21:21:20pamauryI have disassemble the first part of the first boot section of the fuze+, it only deals with the power system and perfectly matches the imx233 doc as far as I checked
21:23:14pamauryIt's not clear to me what it does, since the power system of the chip is quite complicated
21:23:35pamauryMy guess is that is tries to find the best setup, depending on the power source
21:24:19bertrikcan I have a look at some other parts?
21:25:02pamauryplease do :) I only had a look at the first generated elf of the ____ section, so there are still lots of things
21:25:14pamauryperhaps try to have a look at another section ? like host or play ?
21:25:24CIA-7New commit by b0hoon (r28721): HDD6330: add missing button for test_codec.
21:25:45bertrikok, I'll start with the smaller of those two
21:25:57bertriksbinfo now creates an .elf, right?
21:26:19pamaurythen you want host :) Yes
21:26:56pamaurythe scheme I use is not perfect but it should do, it creates only elf file for each [load; ..;load; fill; ...; fill; call/jump] block
21:26:58CIA-7r28721 build result: All green
21:27:00Buschelkugel: test_mem does not compile on cygwin for simulation here...
21:27:43kugelBuschel: for simulation?
21:27:52bertrikthe "host" one looks very simple :)
21:28:06bertrikjust LOADs and one JUMP
21:28:37 Quit toffe82 (Read error: Connection reset by peer)
21:28:44kugelI'm not sure if it ever worked on the sim
21:28:49Buschelyes, only checked for simulation yet. it does not like the "for(int i=0," −− it wants to have "i" declared outside the "for"-statement
21:29:10pamaurybertrik: any number of loads/fill and then one call; to match the fact that you "call" an elf file, thus it only has one entry point
21:29:11Buschel"error: 'for' loop initial declaration used outside C99 mode"
21:29:23n1sbut we use c99 now
21:29:40kugelwe "officially" allowed that now
21:29:47n1sBuschel: have you regenerated your config lately?
21:30:17BuschelI thought I had, will do "make reconf" now
21:30:58kugelah, configure doesn't add -std=gnu99 for sim builds
21:31:56 Quit jgarvey (Ping timeout: 264 seconds)
21:32:13Buschelkugel: there is also a warning added -> "warning: unused variable `boost'" in line 191
21:35:28CIA-7New commit by b0hoon (r28722): HDD6330: implement block transfers through the LCD2 bridge in lcd_update_rect(), speedup is huge, fps over 4 times faster.
21:35:57 Join jgarvey [0] (
21:36:11 Quit {phoenix} (Read error: Connection reset by peer)
21:37:13CIA-7r28722 build result: All green
21:38:28Buscheln1s: very, very mixed up.... only the wmapro part of the patch is interesting on beast...
21:40:18kugelBuschel: I have a fix but it breaks build in other places
21:40:25n1sthe measuring uncertainty on the beast is about 1% in my experience though so maybe the numbers are just an effect of that
21:40:40bertrikpamaury, do you have an imx233 datasheet?
21:41:31Buscheln1s: for wmapro the difference is >2%. so, this should be reliable?
21:41:45pamaurybertrik: yes, I think there is a link on the fuze+ wiki page
21:41:49n1syeah, i think so
21:42:55pamauryhum, I might be wrong. I'm working with a imx23 datasheet
21:43:20kugelcalling standard library functions via rb-> messes things up...
21:44:14Buscheln1s: strange... I expected the largest change for a52...
21:45:41bertrikpamaury, I found an imx23 applications processor reference manual, but the title of the doc says imx233
21:45:51pamaurysame for me
21:46:31bertrikIMX32RM Rev 1 11/2009
21:46:48pamaurysame here :)
21:47:04saratogatest_mem could probably use more unwrapping on ARM for read/writes
21:47:22saratogaalso adding an option to do aligned, vs. unaligned, and random vs. sequential
21:48:14kugel-O makes some c lib functions a macro, apparently
21:49:29n1sBuschel: on most (all targets with iram?) targets the codec stack is in iram so making those buffers static shouldn't affect the speed on coldfire really
21:50:28Buscheln1s: wait. the movem-advantage when aligning was only valid for DRAM, not IRAM, right?
21:51:16n1sBuschel: yes
21:51:51 Join Llorean1 [0] (
21:51:52saratogawhy does the read test use a label called ".inner_loop_write" :)
21:51:52Buscheln1s: then this was the fact I forgot.. :/
21:52:04n1sthe icache can give noticable effects from just shuffling things around though :(
21:52:19Buschelsaratoga: because it was obviously copy-pasted ;)
21:52:46Buscheln1s: so, I better not change the codecs for now.
21:53:16 Quit Llorean (Ping timeout: 245 seconds)
21:55:43kugelmaking an inline non-inline fixes link error, making it static inline fixes it as well; just keeping it inline gives the error
21:56:12n1swhat is the link error?
21:56:15 Join sirdaniel [0] (
21:57:14 Quit sirdaniel (Client Quit)
21:57:14 Quit Llorean1 (Quit: Leaving.)
21:59:44kugeln1s: mp3_encoder.c:(.text+0x6238): undefined reference to `myswap32'
21:59:58kugelmyswap32 is the mentioned inline function
22:03:14n1salso i imagine any compiler will produce horrid code from that ad it does the actual byteswap in an incredibly weird way
22:03:57 Join yelped [0] (
22:04:37n1sweird errors that go away like that usually mean there is something subtly wrong somewhere in my experience
22:05:05 Quit Slasheri_ (Ping timeout: 255 seconds)
22:05:10 Join Slasheri [0] (
22:05:10 Quit Slasheri (Changing host)
22:05:10 Join Slasheri [0] (miipekk@rockbox/developer/Slasheri)
22:05:10 Quit yelped (Client Quit)
22:05:18 Join yelped [0] (
22:06:27preglowsaratoga: no thoughts on it?
22:06:48yelpedPamaury: Isn't the Lyre Project usin the imx233 as their development board?
22:07:09pamauryI think so
22:07:32kugeldon't know, they seem to be more busy with switching boards than getting something running
22:08:00pamauryyes, that's part of the problem
22:08:19pamauryBut I think they managed to make rockbox run on it
22:08:45yelpedBtw, I think the imx233 is the board, and the imx23 is the renamed STMP3780 processor.
22:09:08yelpedSo, can we build on their work?
22:09:12pamauryAnyway, the board is different so we can't just reuse their code, we first have to figure out the pins and peripherals
22:09:15kugelpamaury: from what I experienced that's not a very hard thing if you have all the docs and a sane way to get code on it
22:09:30n1skugel: hmm, no this is very strange, the only call that is made to myswap32 is conditional on a var that is set according to a compile time constant but since the function is not static gcc should never delete it, even if it decides it is never called
22:09:54 Join b0hoon [0] (
22:10:10pamaurykugel: indeed, the imx23 doc is pretty complete
22:10:14n1sso this could very well be a compiler bug
22:10:23 Quit b0hoon_ (Ping timeout: 250 seconds)
22:10:40kugelgcc version 4.4.5 (Ubuntu/Linaro 4.4.4-14ubuntu5)
22:11:03kugelthat'd mean a regression from 4.4.4 to 4.4.5 which seems unlikely
22:11:26kugelit's triggered by std=gnu99 if I see it right, which we set in target builds
22:11:49 Join casainho [0] (
22:13:46casainhoI am starting to port Rockbox for a target which uses ARM i.MX233
22:14:06casainhothere is already imx31 on Rockbox tree
22:14:11casainhoshuld I merge both?
22:14:40yelpedPerfect timing...
22:15:09 Quit factor (Remote host closed the connection)
22:15:15Torneif they are similar devices, yes
22:15:16Torneotehrwise no
22:15:25pamauryI fear imx23 and imx31 are completely different
22:15:48n1syes, the imx23 is a rebranded sigmatel, right?
22:15:51kugelBuschel: does work for you?
22:16:32pamaurycasainho: which target ?
22:16:52 Join factor [0] (
22:17:26 Join stoffel [0] (
22:18:55Buschelkugel: I just added -std=gnu99 for simbuild and it worked (except the warning reagrding "boost", of course)
22:19:44casainhopamaury: this one:
22:20:17pamauryah, that's the lyre project
22:20:18casainhon1s: yes!
22:20:34CIA-7New commit by kugel (r28723): Change the gcc options for sdl builds to allow for gnu99 features, it needs some fixes in other places. Fixes test_mem compilation failure on cygwin.
22:20:42TheLemonManpamaury: you might need this when disassemblying raw blobs :)
22:20:48Buschelkugel: compiliing your patch now
22:21:01rasherThat would point heavily at "for god's sake don't merge"
22:21:27 Join Luca_S [0] (
22:21:40Buschelkugel: you were faster, sync'ing now
22:22:06pamauryTheLemonMan: so far, I've managed to disassemble things without any any problem, but thx
22:22:50CIA-7r28723 build result: 488 errors, 3 warnings (kugel committed)
22:23:21TheLemonManheh that script is handy as fuck
22:23:42kugelstrange, I checked fuzev2
22:24:59***Saving seen data "./dancer.seen"
22:25:01pamauryTheLemonMan: what is the use ?
22:25:35pamaurywith sbinfo, I build elf files from an sbfile, so I have the entry points, ida finds everything else alone
22:26:08TheLemonManit forgot to analyze some areas for me, mostly when interworking is used
22:29:10CIA-7New commit by kugel (r28724): Fix reds, inclusion of C files into plugins is tricky.
22:31:36CIA-7r28724 build result: 0 errors, 3 warnings (kugel committed)
22:31:42CIA-7New commit by kugel (r28725): Fix player sim yellow
22:33:34CIA-7r28725 build result: All green
22:35:01n1shmm, the use of (UN)LIKELY affects inlining in newer gcc versions
22:35:23n1s"warning: inlining failed in call to ‘oggpack_look’: call is unlikely and code size would grow"
22:35:57Buschelkugel: works for me as well
22:36:47Buschelkugel: bad effect in sim -> the loop count increases unstoppable as memset and memcpy will not exceed 2 or 3 ticks.
22:37:18Buschelkugel: do not knwo why... maybe this is not relevant for real targets
22:37:18kugeldoes it crash?
22:37:30Buschelit just takes longer and longer to update
22:37:43kugelbecause the other loops are slower?
22:37:44Buscheland it can be regularly aborted
22:39:29kugelit won't do several GB/s (will it?), so it should stop getting faster at some point
22:39:38kugelerr, getting higher
22:40:25Buschelthe simulator reach 1600 MB/s here ;o) (very old notebook)
22:40:51kugelI'm getting 18GB/s
22:41:08Buschelas I said, very old notebook :)
22:41:26kugeland memset/cpy is *much* faster then the read/write tests
22:42:00Buscheldoes the tick count increase above 10 on your machine?
22:42:22Buschel(for memset/memcpy)
22:42:25 Quit stoffel (Remote host closed the connection)
22:42:28kugelyes, slowly :)
22:42:56Buschelmaybe you should double the loop_cnt each time :)
22:43:01*Buschel changes code
22:44:22Buschelthere we go
22:44:27Buschel>10 now
22:44:44Buschel~ 8 GB/s now
22:45:40kugel11 now
22:45:53kugeland I have 16/8 GB/s
22:45:56n1shaha, *disabling* branch hints (the UNLIKELY/LIKELY macros) *improves* vorbis decoding performance on cf with gcc 4.5...
22:46:21kugelmakes me think the hint is wrong
22:46:51n1ssince it decides that calls to functions declared "inline" that end up in "unlikely" branches are not worth to inline since the code will grow
22:47:41kugelif inlining makes a difference on speed, then it's not unlikely enough to UNLIKELY() it, I'd say
22:48:18kugelwhat happens if you make the other case LIKELY?
22:48:44n1sthe problem is that the manual hints can say only likely/unlikely and AFAICT unlikely == !likely so if one branch is likely, the other is automatically unlikely
22:51:21n1sthe old gcc didn't take this into account when inlining but could generate code that made taking a conditional branch be the unlikely case which is at least a theoretical win as long as the unlikely branch is executed < 50% of the time but if the inliner decides to make the unlikely branch that is executed 10% of the time 2X slower this will show up in overall speed...
22:51:31kugeldoes __attribute__((always_inline)) help?
22:51:47n1sbut it's a mess
22:52:04n1si'd much rather drop the branch hints
22:52:48pamaurycasainho: did you write any code for the imx23 ?
22:53:47 Nick Ypsy is now known as YPSY (
22:54:28 Quit sinthetek (Remote host closed the connection)
22:54:48n1sah, this is happening on arm builds with gcc 4.4 too...
22:55:10 Join sinthetek [0] (
22:55:10 Quit sinthetek (Changing host)
22:55:10 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
22:56:53casainhopamaury: yes, but just a little, like flash a LED and init and test SDRAM
22:57:30casainhopamaury: my next task it to write code for interrutp and so Rockbox kernel tick
22:58:20 Join MethoS- [0] (~clemens@
22:59:28 Quit Strife89 (Quit: Weird connection?)
22:59:53 Join Strife89 [0] (
23:03:51 Part b0hoon ("GTG. Bye.")
23:06:02kugelBuschel: mpc has gotten slower since my last test, for some reason
23:06:02 Quit yelped (Quit: CGI:IRC (EOF))
23:06:54Buschelmuch slower?
23:07:55kugelnot sure what's wrong
23:08:01Buschelin numbers?
23:08:29kugel2400% vs 3900% realtime
23:08:42kugeleven my tests from october are faster
23:08:56kugelI think not only mpc is affected, ogg is slower too
23:11:06 Quit jgarvey (Quit: Leaving)
23:11:14TheLemonMandoes anyone know where i can find hw docs about the SMTP3700 ?
23:11:40kugelBuschel: ok, perhaps I messed it up, clearing the build directly seems to have fixed it
23:11:53kugel4500% realtime now (my phone is overclocked now)
23:12:45n1svorbis shouldn't have changed much for targets with fast ram
23:12:46Buschelkugel: do you mind if I resort test_mem a bit? this will give better readability for small display (like my nano) and uses the *=2 way to increase the loop count
23:12:53pamauryTheLemonMan: search for imx23
23:12:57kugelnot at all
23:13:22 Join JdGord [0] (
23:13:31TheLemonManit has the same processor ?
23:16:51pamauryit's more or less the same chip, since freescale bought sigmatel
23:18:33TheLemonManin the datasheets there are the references for memory mapped i/o ?
23:18:35 Quit Dreamxtreme (Read error: Connection reset by peer)
23:18:56CIA-7New commit by Buschel (r28726): Changes to test_mem. Improve readability for smaller displays, increase loop count by a factor of 2, if needed,
23:19:10 Join Dreamxtreme [0] (~Dre@
23:19:20pamauryTheLemonMan: there are 1600 pages of them ;)
23:19:49TheLemonManorly? :D link please, i have found just rubbish :\
23:20:30pamaury"imx23rm datasheet" in google
23:20:48CIA-7r28726 build result: All green
23:21:08 Quit hebz0rl (Ping timeout: 245 seconds)
23:22:08 Join {phoenix} [0] (~dirk@xGagC832.WH1.TU-Dresden.De)
23:24:37scorchegevaerts: yeah, i hope to get some time around the end of the month where i can check up on upgrading the forums version which sould hopefully take care of that among other things...
23:25:14Buschelkugel: you reach 4500% for mpc now? at which bitrate?
23:26:28Buschelkugel: the former measurements were done @600MHz, right?
23:27:32 Nick amee2k is now known as amee2corpse (
23:27:32Buschelkugel: 768/45 = 17.1 MHz, still ~2,9 MHz faster than with your last measurements
23:27:32Buschelquite a lot
23:27:33TheLemonManpamaury: this is the holy bible :D thanks a ton
23:28:11kugelBuschel: yep :)
23:29:05Buschelkugel: the alignment stuff seems to work quite well for your target
23:29:30gevaertsscorche: I hope we won't get too many spammers until then
23:30:33Buschelhowever, I need to get some sleep now
23:30:37Buschelsee you!
23:30:50 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
23:32:58 Quit efyx (Quit: Quitte)
23:33:51CIA-7New commit by kugel (r28727): Enable ARM ASM optimzations on android. Speedup in codecs ranges from 20% to 150%. Most codecs now need well below 30MHz.
23:34:21 Join BlakeJohnson86 [0] (
23:34:32 Quit TheLemonMan (Quit: free(me))
23:34:47 Quit Keripo (Quit: Leaving.)
23:36:54CIA-7r28727 build result: 4 errors, 43836 warnings (kugel committed)
23:39:08 Quit Kupop (Ping timeout: 272 seconds)
23:39:54Luca_Swow, impressive warning count :D
23:40:54 Quit Luca_S (Quit: CGI:IRC (EOF))
23:41:01 Join jdgord| [0] (
23:41:02 Quit JdGord (Read error: Connection reset by peer)
23:41:06scorchegevaerts: has there been a recent surge?
23:41:25gevaertsscorche: I think four bans today
23:41:29 Nick amee2corpse is now known as amee2k (
23:41:42*gevaerts vanishes
23:42:34 Quit JdGordon (Ping timeout: 240 seconds)
23:43:43 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
23:43:52kugeltest_mem is a lot faster with the asm
23:44:52 Quit BlakeJohnson86 (Read error: Connection reset by peer)
23:44:54 Join BlakeJohnson861 [0] (
23:45:52 Join JdGord [0] (
23:45:53n1shmm, seems a single call to a function that should be inlined made up most of the difference from removing the branch hints entirely and having them so i'll probably just drop that one
23:45:58 Quit jdgord| (Read error: Connection reset by peer)
23:47:57kugeln1s: are you still targetting gcc 4.4.5 or 4.5 now?
23:49:36n1si've mostly played with 4.5 now since i found out 4.4 has a stupid bug that makes code a lot bigger and slower in many cases which was fixed in 4.5 apparently, I haven't speed tested anything but tremor on 4.5 so far though although i don't expect a large diff to 4.4
23:49:52 Quit bertrik (Quit: :tiuQ)
23:50:17n1s4.4 sometimes decided to access the stack in halfwords instead of words
23:50:36kugelrockbox uses 5% cpu on my phone while playing music
23:50:48kugeln1s: weird
23:50:55n1sif i feel motivated i might run some tests this weekend
23:51:01n1skugel: yes very strange
23:51:04kugeldid you report the bug?
23:51:08saratogapreglow: its pretty nice
23:51:10 Quit kevku (Ping timeout: 272 seconds)
23:51:25saratogalike all the ffmpeg codecs its neatly written
23:51:26n1sno, but since 4.5 doesn't do it i don't think it's worth it
23:51:36 Quit komputes (Remote host closed the connection)
23:52:07kugel4.4 is still supported, and will probably be in use for some years
23:53:07n1ssure, but since m68k is a 3rd tier platform and the bug causes just code bloat and slowdown i don't think thay'll even think of fixing it
23:54:07n1sat least one wrong code bug i reported in 4.3 got fixed eventually but only in 4.4 and was never backported (and my testcase wasn't even included)
23:55:19 Quit evilnick_B (Quit: Page closed)

Previous day | Next day