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 2004-05-16

00:00:40[IDC]Dragonthere are some mort CRT tidbits which could go into IRAM
00:01:11[IDC]Dragone.g. the mul/div subroutines
00:02:22amiconnGrayscale framework and jpeg update including that committed now.
00:03:47amiconnIf there is space left in IRAM, of course this space could be used for speedup
00:04:20amiconnThe question is: how do you tell the compiler/linker that it should put a library routine in IRAM?
00:04:46[IDC]DragonI think that's easy
00:04:57[IDC]Dragonjust name it in that section or so
00:07:03[IDC]Dragonwe can almost fit the whole libgcc.a
00:07:36amiconnAnyway, routines that don't access other memory and are not called in short sequence won't profit that much from putting them into IRAM
00:07:37[IDC]Dragonwhich is 856 bytes
00:08:26amiconnThe only thing that would avoid is the double page miss (one at call time, one at return)
00:09:00[IDC]Dragonand probably twice the bus speed
00:09:26[IDC]Dragon816 bytes left#
00:09:59amiconnThe sh can execute one instruction (16 bit) per cycle, and read from external ram one word per cycle (16 bit), as long as there is no page miss, since we have no extra wait.
00:11:14amiconnImho, if we want more speed, we should try to put routines into iram that _do_ access other memory.
00:12:03amiconnIf these are not in IRAM, this produces two page misses per memory access (one for the access and one for the next instruction fetch)
00:12:54amiconnI think of strcpy, strcmp..
00:15:33[IDC]Dragonwhow, now _this_ is what I call scrolling!
00:15:50amiconnLast time you talked about IRAM, you mentioned that some additional space is needed for the debug builds. Did you take this into account?
00:16:16[IDC]Dragonno, I didn't :-(
00:16:57[IDC]Dragonreally cool, that scrolling!
00:17:05midkwhat scrolling
00:17:07midkwhat did i miss
00:17:11[IDC]Dragon(just can't get enough of it)
00:17:25[IDC]Dragonimproved JPEG viewer
00:17:30midkhow does it look?
00:17:34midkcan't test until tomorrow
00:17:55[IDC]Dragonoops, mine was too short
00:19:10amiconnBtw: the new 8-pixel subroutine uses 12 registers at once. Pretty close to the available 16...
00:21:44[IDC]Dragonsaw that long streak of shift and rot
00:23:01[IDC]Dragonmattzz could change his mandelbrot then again, to output 8 lines at a time
00:23:16amiconnProbably the more interesting part is the asm block above that - the precalculation of all 8 patterns, with random shift.
00:23:55amiconnHmm, the problem with the mandelbort is that it calculates pixel by pixel.
00:24:40[IDC]Dragonwe had it 8-line-wise once already, remember?
00:24:48amiconnIf he wanted to take advantage of the speedup, he would have to use a little bitmap (8 pixels high), render into that, and then draw that.
00:26:46[IDC]Dragongoodnight folks
00:26:52 Quit [IDC]Dragon ()
00:27:07amiconnOoops, too late
00:32:44elinenbeamiconn: what type of speedup can you expect with the new grayscale routines?
00:34:21amiconnelinenbe: gray_fillrect() and gray_drawgraymap are 2..3 times as fast (for larger rectangled and images).
00:34:44amiconnScrolling left/right is also significantly faster, scrolling up/down by 8 pixels a bit faster.
00:34:47 Join edx{jogging} [0] (
00:35:11midkmsut talk to linusn about that repeat rate function.
00:35:13 Nick edx{jogging} is now known as edx{code} (
00:35:24elinenbeI'll compile the latest and tell you if you deserve congradulations!
00:35:41midkoh amiconn btw i like the new grayscale demo
00:35:57amiconnA bit of speedup is expected for other functions too.
00:36:11midkhmm sort of a dumb question... is there no way to remove the horizontal lines that run across the lcd?
00:36:24elinenbeamiconn: how long until we see anti-aliased grayscale fonts and a grayscale interface :)
00:37:09amiconnelinenbe: I don't thinkl there will be a grayscale interface, but font support is scheduled next (not anti-aliased yet)
00:37:35elinenbeamiconn: sounds great!
00:37:50elinenbeI think it is fabulous....
00:38:31amiconnmidk: Unfortunately not. These lines are caused by interference of the grayscale frame rate with the internal scan rate of the lcd, which isn't absolutely constant and cannot be measured by a program.
00:39:05elinenbeis the site down?
00:39:15elinenbeboo! can anyone else access it?
00:39:26amiconnWorks perfectly here...
00:39:45elinenbereally? can't access the cvs or the web page from here.
00:40:30amiconnJust clicking my way through ViewCVS...
00:41:04elinenbei just use cygwin...
00:41:24elinenbesomething is up with my provider...
00:41:39elinenbeamiconn: not to sound rude or anything, but what is the benefit of grayscale fonts?
00:41:41amiconnGrr, there are TAB chars. Again!
00:41:50elinenbeor is it just completing the grayscale framework?
00:43:01amiconnThis is just completing the grayscale framework, and not about real grayscale fonts but showing the regular rockbox fonts within the grayscale are, with selectable foreground & background values.
00:43:15amiconn*grayscale area
00:43:30 Quit Nibbler (Read error: 54 (Connection reset by peer))
00:43:40 Join Nibbler [0] (
00:44:22elinenbewell, can you place the regular font in any shade then?
00:45:23amiconnAnd you can select if the background is transparent or has another shade (just like for 1-bit bitmaps now)
00:45:27elinenbeit would be nice to implement that into the WPS
00:45:37elinenbealong with a "wallpaper" (now I am getting silly here)
00:45:48elinenbewhat is the IP you get for
00:47:06elinenbesame here... ah... I'll just wait until later.
00:48:04elinenbeamiconn: you really have put some great work into rockbox... thanks so much!
00:48:22elinenbeamiconn: keep it up! it's exciting to see what you think of.
00:48:23amiconnStrange. There was another user approx. 2 hours ago who also couldn't access
00:48:42elinenbemaybe there is a problem from the states?
00:48:50amiconnBut for Jörg as well as for me it worked the whole time.
00:49:19amiconnMaybe, since we both are in ol' europe here.
00:50:54amiconnelinenbe: Tried a tracert?
00:51:26 Quit edx (Read error: 110 (Connection timed out))
00:55:30elinenbe$ tracert
00:55:31elinenbeTracing route to []
00:55:31elinenbeover a maximum of 30 hops:
00:55:31DBUGEnqueued KICK elinenbe
00:55:31elinenbe1 101 ms 90 ms 120 ms
00:55:31elinenbe2 120 ms 141 ms * []
00:55:31***Alert Mode level 1
00:55:31elinenbe3 * 221 ms 220 ms []
00:55:33elinenbe4 371 ms 320 ms 251 ms []
00:55:35elinenbe5 * * 20 ms []
00:55:36midkDA DA DA DA
00:55:37elinenbe6 * 50 ms 20 ms []
00:55:38midkDA DA
00:55:39elinenbe7 20 ms 30 ms 30 ms []
00:55:48elinenbe8 50 ms 30 ms 30 ms []
00:55:48elinenbe9 100 ms 120 ms 481 ms []
00:55:48elinenbe10 * * * Request timed out.
00:55:48***Alert Mode level 2
00:55:48elinenbe11 * * * Request timed out.
00:55:49***Alert Mode level 3
00:55:49***Alert Mode level 4
00:55:49elinenbe12 * * * Request timed out.
00:55:51***Alert Mode level 5
00:55:51elinenbe13 * * * Request timed out.
00:55:54elinenbeboooo! :(
00:59:06amiconnThe last host belongs to "hurricane electric internet services", a Canadian firm. "lon" probably means London
00:59:36amiconnOoops, Californian..
01:00:28amiconnAnyway, I'm off.
01:00:41amiconnnite folks
01:01:21 Part amiconn
01:05:53***Alert Mode OFF
01:19:00 Quit Nibbler (Read error: 54 (Connection reset by peer))
01:24:42 Quit edx{code} ()
01:46:42***Saving seen data "./dancer.seen"
01:57:06 Quit _aLF ("Leaving")
01:59:32 Quit top_bloke (Read error: 60 (Operation timed out))
02:05:29 Quit midk (Read error: 104 (Connection reset by peer))
02:05:47 Join midk [0] (
02:34:30 Join [1]c0utta [0] (
02:57:35 Quit c0utta (Read error: 110 (Connection timed out))
02:57:38 Nick [1]c0utta is now known as c0utta (
03:04:03 Join Nibbler [0] (
03:11:02 Join Guest12 [0] (
03:11:35Guest12Is the main site down or is it gone?
03:12:57Guest12How long has it been down and anyone know when it will be back up?
03:13:05midknot sure, and not sure
03:13:44Guest12Oh, thanks for the info
03:14:29midkheh k
03:14:31Guest12Lets hope it is not down to long and there is no problems
03:14:39midknot too long... but not sure why
03:14:43midkhappened before
03:15:22Guest12Well I just got my archos jukebox not to long ago, already have rock loaded, very nice. :)
03:15:40midksounds good, enjoy
03:16:08 Quit Guest12 ("Leaving")
03:44:45 Quit Nibbler (Read error: 54 (Connection reset by peer))
03:46:44***Saving seen data "./dancer.seen"
03:49:58 Quit AciD (Read error: 54 (Connection reset by peer))
05:16:14 Join Nibbler [0] (
05:20:27midkholy monkys.
05:46:46***Saving seen data "./dancer.seen"
06:30:20 Quit Nibbler (Read error: 104 (Connection reset by peer))
07:27:31 Join Nibbler [0] (nibbler@
07:46:48***Saving seen data "./dancer.seen"
09:46:52***No seen item changed, no save performed.
09:58:25 Quit adi|home (Client Quit)
10:02:30 Quit midk ("yo yo yo cya later YO YO YO wasa wasa!")
10:04:01 Join midk [0] (
10:04:23 Join [1]c0utta [0] (
10:08:17 Join adi|home [0] (~adi|
10:17:43 Quit c0utta (Read error: 110 (Connection timed out))
10:17:47 Nick [1]c0utta is now known as c0utta (
10:48:17 Quit midk ("yo yo yo cya later YO YO YO wasa wasa!")
11:15:14 Quit Nibbler (Read error: 54 (Connection reset by peer))
11:26:53 Join Nibbler [0] (
11:42:40 Join cjnr11 [0] (
11:42:44 Part cjnr11
11:46:56***Saving seen data "./dancer.seen"
11:55:10 Quit Nibbler (Read error: 104 (Connection reset by peer))
12:02:46 Join gon165 [0] (
12:03:17 Quit gon165 (Client Quit)
12:03:41 Join Gon165 [0] (
12:03:57DBUGEnqueued KICK Gon165
12:04:08Gon165is anybody here?
12:04:20 Quit Gon165 (Client Quit)
12:04:24 Join gon165 [0] (
12:04:39gon165i need some help
12:15:44 Quit gon165 ("Euh .........Rien")
12:27:46 Join gon165 [0] (Gon165@
12:27:56gon165still nobody?
12:37:51gon165so if somebody know what is a ATA error: -11 ====>>
12:37:54 Quit gon165 ("Euh .........Rien")
12:39:52 Join Nibbler [0] (
12:40:11 Quit Nibbler (Client Quit)
12:40:15 Join Nibbler [0] (
12:40:18 Quit Nibbler (Remote closed the connection)
12:40:49 Join Nibbler [0] (
13:09:51 Join AciD [0] (~acid@
13:46:57***Saving seen data "./dancer.seen"
14:23:17 Quit Nibbler (Read error: 104 (Connection reset by peer))
15:46:54 Join Nibbler [0] (
15:47:01***Saving seen data "./dancer.seen"
16:07:02 Join matsl [0] (
16:17:35 Join edx [0] (
16:27:53 Join _bleh [0] (Joosa@
16:28:20 Part _bleh
16:40:41 Quit elinenbe (Read error: 104 (Connection reset by peer))
16:40:47 Join elinenbe [0] (
17:33:51 Quit elinenbe (" HydraIRC -> <- *I* use it, so it must be good!")
17:47:05***Saving seen data "./dancer.seen"
18:06:56 Quit matsl (Remote closed the connection)
18:50:41 Join matsl [0] (
19:17:52 Join diddystar5 [0] (
19:17:52 Quit Nibbler (Read error: 104 (Connection reset by peer))
19:18:50diddystar5how are you all?
19:28:38 Quit diddystar5 ("Leaving")
19:29:02 Join diddystar5 [0] (
19:42:57 Join top_bloke [0] (
19:46:05 Quit top_bloke (Client Quit)
19:46:50 Join top_bloke [0] (
19:46:57diddystar5hey toppy
19:47:09***Saving seen data "./dancer.seen"
19:53:32 Join scott666_ [0] (~scott666@
19:55:30diddystar5yo scott666_
20:42:43 Join whiskieman [0] (
20:43:11whiskiemanhey guys
20:43:17whiskiemanwell done for ya project ;)
20:43:39whiskiemani was reading around ur site to see if there was a plan to release rockbox for the multimedi juckebox
20:43:54whiskiemanunfortunately it doesn't look like it right P
20:43:55diddystar5will not happen
20:44:00whiskiemani kow
20:44:03whiskiemani know
20:44:11whiskiemanread the request
20:44:20diddystar5there is a firmwaree being made for the av3xx units
20:44:22whiskiemanwaht about other projects?
20:44:42whiskiemanwhat bout mm projects?? know any?
20:45:53diddystar5i doubt there will be one, since the MM are going away fast
20:46:02diddystar5and are being replaced by the av
20:46:11diddystar5but who knows
20:46:15diddystar5it COULD happen
20:49:10whiskiemanwish so
20:49:39whiskiemantks for the info neway ;)
20:51:14diddystar5sorry about your dissapointment
21:00:41 Join Nibbler [0] (
21:03:19 Quit whiskieman ("—I-n-v-i-s-i-o-n— 2.0 Build 2520")
21:10:39 Quit diddystar5 ("Leaving")
21:12:30 Quit AciD (Read error: 54 (Connection reset by peer))
21:30:22 Join diddystar5 [0] (
21:32:17 Join lImbus [0] (
21:47:12***Saving seen data "./dancer.seen"
22:19:17 Quit edx ()
22:19:17 Quit Nibbler (Read error: 54 (Connection reset by peer))
22:27:02 Quit lImbus (" HydraIRC -> <- *I* use it, so it must be good!")
22:38:30 Quit diddystar5 ("Leaving")
22:42:38 Join diddystar5 [0] (
22:44:44 Join amiconn [0] (
22:52:24diddystar5yo amiconn
22:52:32amiconnhi diddystar5
23:02:39 Quit top_bloke ("The mind is a terrible thing to taste. Wasted 1 hour 47 minutes and 31 seconds online.")
23:12:37 Join AciD [0] (
23:13:12 Part scott666_
23:26:56 Join midk [0] (
23:27:55midkhi dids
23:29:36 Quit matsl (Remote closed the connection)
23:47:14***Saving seen data "./dancer.seen"
23:55:27 Join Nibbler [0] (

Previous day | Next day