| |
| |
| |
|
Page: 1 2
Comments:
<0> Hi behdad.
<1> hi plam
<2> hola
<3> ugh
<3> what happened to the gitweb stylesheet?
<4> krh: slashes in the url
<4> the css include needs to be made non-relative
<3> hm, but why did it suddenly break?
<4> your browser might have been cacheing it before
<3> we should do a fd.o stylesheet for gitweb, btw
<4> and might cache it again if you visit the root page
<4> i had a custom css on git.ndesk.org but unfortunately gitweb's html isn't well marked up
<4> so i had to fix gitweb, and then gitweb got refactored and my changes broke
<4> i might submit patches to fix that unless someone beats me to it. it should output h1,h2 etc. tags for headings and the blob | shortlog should become ul li
<5> question: The new X1600's will they work with the radeon driver in Xorg? I have someone who wants to buy a laptop and try out the Xgl/aiglx stuf
<6> i don't think the xorg radeon driver does 3d on them
<7> the xorg radeon driver doesn't even do 2D on them
<5> not good :/
<7> tell ATI
<0> keithp: that's a lot of changes.
<8> alanc_: just got myself an X300se...
<8> alanc_: besides that should read s/ATI/AMD
<7> that should work - my X300 and X700 both do - it's the X1_00 that ATI has held out on
<7> sx-wks: only if you're planning ahead - the AMD takeover isn't supposed to happen until "Q4 2006"
<9> plam: heh. indeed it was
<9> plam: the goal was to switch the caches over to using relative offsets for all pointers; eliminates the need to track which cache something is in
<8> keithp: would that be worth it for a next generation car computer ? http://linitx.com/product_info.php?cPath=12_101&products_id=1103
<0> That's what those intptr_t guys do?
<9> plam: yeah, offsets everywhere
<9> plam: one cool side effect is that the lang sets are just 'normal' FcCharSets now
<9> although initializing them was 'interesting'
<0> Initializing langsets has always been 'interesting'.
<9> yeah, but now it's all relocation free and needs no special code in the library
<9> Once I realized how that would work, the changes made a lot more sense
<0> And so now you find the address of something else by taking its difference from you?
<9> Difference from some defined base address
<9> Which base address is used needs to be documented...
<6> sx-wks: sounds like a bit much for a carpc
<8> dobey: heh
<8> dobey: I have, erh... issues with the current EPIA MII12000 and cairo...
<6> sx-wks: hrmm
<9> plam: also, eliminating the bank makes the data structures smaller, which is a nice bonus
<0> Yes. I do remember considering something like that, but don't remember why I didn't use it originally.
<8> like, on the intel based dell laptop (Pentium M 1.4G usually only at 350, I can sustain proper functionnality at 1920x1200, whereas on the epia, I had to disable 1/2 of the stuff at 800x600
<8> so I'm wondering...
<8> 1) is the epia pure crap
<6> the epia isn't crap
<8> 2) is my Xorg missing something obvious
<6> but i don't know what you're doing
<6> or how well the drivers you have work with the card on it
<6> the unichrome cards should be pretty nice, but your laptop probably has a better radeon on it?
<8> mebbe.
<8> dobey: the laptop has integrated intel graphics
<6> sx-wks: ok. the epia has via unichrome
<10> probably all down to the cpu, and the memory bandwidth
<10> cpu = C3 nehemiah at 1G
<10> tiny cache, no out of order execution, etc
<6> perhaps
<10> then there's the fact that there's only 266Mhz ddr to play with
<6> i know all the people running winxp on them have no issues with doing funky 3d things at 800x480 or whtever though
<8> true
<0> keithp: in fact I had a vague memory of trying that but bailing because it didn't work at the time. Might just be a hallucination.
<8> libv: I could try removing cairo from the equation and see how that goes
<6> sx-wks: you might just be lacking useful acceleration on the unichrome, and a compositing manager? it's hard to say really
<8> dobey: well the app is running alone fullscreen...
<8> so the compositing manager would be irrelevant, I gues
<8> +s
<6> perhaps
<6> but the accel isn't
<8> right
<6> the compositor might not be insignificant either, depending on how it does things i guess
<8> dobey: that's the running code: http://www.navsys.org/cvs/pynavsys/navsys.py?rev=1.13&view=auto
<9> plam: probably the asbestos in the heating tunnels
<0> keithp: I hope there's not some gotcha, though.
<6> sx-wks: ugh. python :)
<8> hey
<8> dobey: ah, and it's querrying a postgres db for the data :D
<6> ...
<6> might just be slow because the cpu/memory is slower and you're doing things like using postgres
<6> heh
<8> mebbe
<10> no centaur made this decade will ever match a pentium m
<8> libv: heh
<10> C7 versus C3 nehemiah is mostly the bump in frequency that's noticable
<8> lol
<8> so not really worth it.
<0> 533MHz, though, is slow.
<9> plam: appears to work in my testing; need to fix fc-cat though.
<0> (I use a Via Samuel 2)
<10> plam: my condolences
<6> 533mhz is slow when you're running software built for a 2GHz box on it... :)
<0> libv: The point is that it has no moving parts.
<0> keithp: Yeah, it's probably at least mostly correct. I ***ume that most of it was search-and-replace?
<9> not as much as I would have liked..
<10> plam: true, the new edens are up to 1.2Ghz these days iirc
<8> and the C7 are said to go all the way up to 2G
<0> keithp: Yeah, I was guessing that the 7am commit was due to an all-nighter.
<10> but they are no longer that acceptable with p***ive cooling only, i think i saw some complaints
<9> 4am my time :-)
<0> libv: I might get one before leaving the country.
<10> plam: so you buy one and then you run?
<0> libv: Graduate, actually. But sure.
<10> find a deep well, drop it in, then emigrate :p
<0> That would solve the cooling problem.
<8> libv: that would induce pollution with heavy metals
<10> one less horrible box on the planet :p
<6> running in Montreal will solve the cooling problem :)
<10> sx-wks: rohs compliant :p
<9> plam: my laptop is p***ively cooled :-)
<8> keithp: it's an OLPC ?
<0> dobey: It was warmer in Montreal yesterday than in Boston.
<9> sx-wks: nope, panasonic R4. 1.3GHz pentium M.
<6> plam: wait 2 months
<8> libv: so, guess I'll either reduce the complexity of my software for the moment, then switch to that commell mobo...
<6> keithp: uses the lap-heatsink cooling method?
<9> dobey: yeah, it can get a bit warm when running flat-out
<8> libv: and use the epia for something else, like controlling my toaster :D
<0> keithp: Hopefully in 1.5 wks I'll have time to look at some of the open bugs. There are a lot.
<0> sx-wks: Can't you use it as your toaster?
<10> sx-wks: careful now, you might stress the nehemiah too much :p
<9> plam: once the code has stablized, I'll start looking at the open bugs too
<8> plam: doesn't generate nearly enough heat
<8> libv: lol
<0> sx-wks: Eat smaller toasts?
<0> keithp: Yeah. Some of them should be redundant now.
<6> plam: the epia mII has a fan
<8> plam: that would be cat-sized bites at most
<10> sx-wks: buy the DVI-01 module
<0> dobey: It's really hard to get a fanless computer now. Sigh.
<8> libv: and ?
<10> the voltage stabiliser becomes scorching hot
<8> libv: hahahaha
<6> plam: it's not that hard. it's just annoying, because you end up with a heatsink bigger than anything else in it
<10> why it needs to go from 5V -> 3.3V seperately, i don't know, as that's supposed to be provided
<9> plam: some of the media machines are fanless, but they've generally got m***ive heat sinks
<8> blame via :D
<10> oh, i do, all the time :p
<8> lol
<6> the heat sink on the nano-itx boards is ridiculous
<10> dobey: and still not entirely sufficient
<8> dobey: the rest too
<0> I don't mind that so much though.
<6> sx-wks: the heat sink on my mII1000 isn't bad
<8> dobey: I was merely conducing the idea that the nano-itx itself was ridiculous
<10> dobey: but it has a cpu fan
<0> Anyway I guess I can worry about that after my thesis defense.
<6> libv: yes, it does
<6> ...
Return to
#freedesktop
or
Go to some related
logs:
install hash_hmac
iotop linux
#mysql
solanadine
#centos
samba for ubuntu64bit
skge ndiswrapper
Premature end of script headers: Setup.cgi blah
#linux
GENTOO xorg for ATI RAGE LT PRO
|
|