| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Comments:
<0> pstickne: but it would be nice if i could check it when it is actually done
<0> not really. no, i just don't want the program go on when the threads haven't finished up because it exits the program if it happens :(
<0> i guess what i need is a dynamic sleep
<0> but i can't do it unless i know how to check if the threads are still active
<1> ivan_d, I don't know how to do that. you could used shared data but that sounds like race condition time.
<0> i think i'll just p*** some variables that becomes zero when the last thread has finished
<0> could work..
<1> ivan_d, look in thread::shared, it has a small example
<0> ah. ok. will check it out
<2> fyi: $DB::single kindof does what i want
<2> but it's still not what i had in mind
<3> just grep the pods :)
<4> http://www.astro.umd.edu/~vernaleo/ Can someone help me install this?
<5> what is it?
<4> a photo gallery script
<5> nope
<4> ok
<5> such a polite young man
<2> rmah: lol ;-)
<6> rmah: I dunno if you saw my question the other day, but I wanted to pick your brain further about that hypercube topology you were talking about a couple of weeks ago
<7> What cpan namespace would you suggest for a library that provides an interface to the front led on apple laptops?
<6> rmah: specifically, I'm working on something where that may be fairly applicable, and I can see the conceptual benefit of it, if the underlying network design follows the layout of the hypercube network directly, but I don't see much benefit (aside from avoiding packet storms) in using it against a flat network architecture (like /16 for a cluster of n0,000 machines.
<3> there are a bunch of Mac:: modules
<8> EvilHerbivore: Mac:: or Device::
<2> avar: Guilty::Pleasure
<9> got it, it should be (!defined &$func)
<10> "front led"?
<10> I didn't realize there was an "led" on a laptop.
<7> There's a led on the front of the apple powerbooks and ibooks that can be controled in linux through /dev/adb
<6> my VAIO has a programmable LED, also ..
<11> asdf
<12> hi all. can anyone tell me what im missing when something like: 1155173136 TRAFFIC: unsupported graphics format 'GIF' is displayed?
<11> stupid solaris
<7> Asdf::Aoeu
<10> Are you sure that's an "LED"?
<7> I have no idea what it is
<7> It's a light-thingy that blinks:)
<1> merlyn, Little Entertaining Display ;)
<5> it's an LED
<1> A LED is by definition, a single diode. Not much use for cool information if that's all it is.
<10> might be an OLD
<10> or whatever they call those organic things
<7> who cares anyway?:)
<1> Hehe.
<10> avar - can you get to it under OSX?
<7> merlyn: I don't know, I don't use OSX
<6> rmah: did you happen to see my question above?
<5> merlyn, it might be an OLED or Electroluminescent, but on Macs, it's not. It's an LED.
<5> infi: no
<1> merlyn, or simply a tiny tiny LCD, but I haven't seen whatever this is we're talking about :)
<6> maybe 1 page back, re: hypercube topology
<5> pstickne: it's not a tiny LCD either
<5> it's an LED
<1> rmah, LED makes no sense.
<1> I can say, my HDD activity light is a LED. That makes sense.
<5> pstickne: why do you say that?
<1> rmah, because LED refers to a single diode.
<5> so?
<1> Well, is it only one?
<5> yes
<1> rmah, how uncool :(
<5> heh
<1> rmah, I thought it was a cute little display or something like on the Toshibas.
<5> infi: I didn't see it
<5> pstickne: nope
<1> rmah, what's it used for?
<5> pstickne: LCD's do not emit light
<1> rmah, but the backlight does.
<5> all LCD screens that *seem* to emit light have a cold cathode backlight
<6> rmah: er, fairly long, so I won't repeat it here. jussec
<1> rmah, which is generally considered to be part of the 'display' :(
<5> pstickne: sure, but it's not an LCD
<13> rmah: i want my .125micron LED display
<1> rmah, I have no idea what it even does/look likes.
<5> oh whatever, I'd rather hear about infi's hypercube thing
<1> rmah, I just can't get excited for something I can control with xset.
<14> "infi" at 219.120.193.247 pasted "hypercube topology" (9 lines, 609B) at http://sial.org/pbot/18873
<6> oops.
<13> and it needs a 6m diagonal
<5> infi: ok, what about it?
<5> you want to know the benefits of a hypercube topology?
<1> infi, expand :)
<6> well, just, what would the benefits be aside from avoiding network storms, over a flat architecture
<6> yes
<5> well, each node only has to know about a few of its neighbors, even when the # of nodes is very large
<6> yes
<5> also, the # of hops a message needs to reach any other node in the network has a fixed (small) upper bound. even when the # of nodes is very large
<6> but having a routing table external to the nodes themselves would take care of that in a flat, /16 layout
<6> yes, as opposed to say, 1 hop to reach all machines in a /16, say on a cluster of 10,000 machines, it would take 13 hops to reach another machine, worst case :)
<5> you don't need a routing table in a hypercube layout
<6> right, it's all done via addressing
<5> you can *always* know the "best" route to a destination based on the node's coordinates in the hypercube
<5> hold on, I must get my tortollini and wine
<6> it seems very useful for a) avoiding network congestion, and b) geographically disperse systems, but I still can't see an advantage, if you already have an architecture that is based on a flat /16. If it was done at the network layer, it would potentially have great benefit, but ***uming it's an overlay over tcp/ip or some other network layer, its usefulness seems shot at that point.
<5> that really depends on what the network is used for
<6> I am/was considering this for a cluster control/batch processing scheduling engine, but the network layer having to be rewritten seems like quite a bit of work :>
<5> ok then
<5> consider
<5> if you have a n=9 n-cube
<5> you have 512 nodes
<5> each node will have 16 edges
<5> thus, each node can be connected to a single switch
<6> yes
<5> each node will have 100% bandwidth to adjacent nodes
<5> for many cl***es of problems this is awesome
<5> for others, this is not so great
<5> it all depends on the problem you're trying to solve
<6> yeah, that's what I was thinking, the underlying network topology itself will need to be mapped to reflect the hypercube structure itself
<5> of course
<5> otherwise, it's just an experiment
<5> woops, I mean 9 edges for each node
<6> obviously ;)
<5> infi, another place for n-cube networks is when you have a HUGE number of nodes
<5> say in a P2P system
<6> have you seen the Delaunay triangulation method?
<5> a 32D network would could have as many as 4.3x10^9 nodes
<6> aha, that would be a really good application
<5> i.e. 1 per person on the planet
<5> and each person would only need to know about 32 others
<5> yet you could efficiently route to anyone else just based on their 32bit ID *without* any routing tables!
<6> exactly, right.
<5> no technique is best for all problems but many can often be used for many different applications :-)
<6> yeah, aside from reducing congestion in my case, I don't think there is a good use for it.
<5> ok
<6> for this application anyway. but thanks for bringing it up a couple of weeks ago, it's a really good design to be aware of :^)
<5> and hypercubes made for a really cheezy movie too!
<6> *groan*
<15> GumbyBRAIN: SALAAM MY BROTHER
<16> Is my taste as the answer to my brother today.
<17> Does anyone know if Devel::Cover will correctly report coverage on code exercised in a child process?
<18> APPEARS AS ARMANDO
<18> <CHr> #3XFHF03
<18> Hello allz, how do iz install linuckz on windows PLS? so i can squeeze the pigeon
<19> splendid
<15> sili: your mere presence strikes fear into the heart of evil
<19> just as it should
<15> sili is the terror that flaps in the night
<6> oh nice, I couldn't find old chat logs from MS Comic Chat
<1> ivan_d, http://sial.org/pbot/18875
<15> sili is the clippers that trim your hedges
<1> ivan_d, I have no idea if that is propper but, I think it is :)
<15> OMG SILI!!
<19> zomg
<15> sili: http://www.amazon.com/gp/product/B000FS9MUQ/ref=ase_imdb-adbox/002-7292389-1232808?s=dvd&v=glance&n=130&tagActionCode=imdb-adbox
<15> <3
<19> i watched that show avidly
<6> GumbyBRAIN: Appears as armando
<16> $var='123-32'; split(/-/,$var); if (@_ > 2) {print "- appears more than a string as an ok strategy.
Return to
#perl
or
Go to some related
logs:
php5-extensions-1.0
jython NameError __file__
#ksh
postfix-tls +etch
bwmon rpm
tk3000 bios
#osdev
tzap error while parsing inversion
debchroot
#linux
|
|