| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Comments:
<0> which is a cool idea because then you can have an SPDT switch on the ground line of the fan that moves from actual ground to 5V for highspeed and lowspeed modes
<1> well, is that superior to just using a resistor?
<0> well, using the resistor will waste power
<1> hmm
<2> it won't waste it
<2> it'll turn it into nice warmth
<1> still curious about this, because i had ***umed they'd be isolated
<1> i suppose id need to see the actual schematic
<3> I have the statement #define STRUCTURE TEMP1. I am writing a file name, target= fopen("TEMP1","w"), where TEMP1 = whatever I define STRUCTURE at runtime.(I.e., if #define STRUCTURE TEMP2, I want the file name to be TEMP2). How can this be accomplished?
<4> jagzz: Google for macro stringification.
<1> # is the stringizing operator.. it requires an extra level of indirection
<1> the easy way is to just #define FILENAME "filename.dat"
<5> the stringizing operator? lol
<1> #define STRINGIZE(X) STRINGIZE_HELPER(#X) #define STRINGIZE_HELPER(X) X
<1> the extra level is necessary because cpp does not re-evaluate after substituting parameters. it will reevaluate when p***ed to the helper.
<3> but in the event the #define must be #define STRUCTURE TEMP1, and i want to concattanate .TXT to the end afterwards, then fopen(#structure ".txt"); should work?
<6> hmm
<6> what's the best modern processor for pure number crunching?
<0> niv_: the terminator T1000.
<6> ;p
<2> opterons
<6> nah, i've been working on cracking this zip file for ages
<7> where is it? distribute the work
<6> it's got source code for some firmware that we were promised when we purchased some framing chips
<6> and now the company is saying we can't have the p***word for it, even though in the readme file it says any customer can have it
<0> heh
<6> they want us to buy a newer model-- apparently they are 'phasing it out'
<0> my uncle's company got stuck in a similar position
<0> the master coder who was working on a product his company bought the rights to quit, and "forgot" to give them the source
<0> so he got to spend 6 months reverse engineering the original C source from the binary he was given
<0> luckily optimizing compilers weren't as smart then so it wasn't as harad
<0> *hard
<8> I'm replacing <a href=""> with <a rel=nofollow href="">
<8> how can I do this only if the <a href=""> does not already have the rel=nofollow in it?
<8> The string: '<a href="1"><a href="2" rel=nofollow>' should become '<a href="1" rel=nofollow><a href="2" rel=nofollow>'
<8> and it should NOT double the nofollow thing: ie. this is undesirable: '<a href="1" rel=nofollow><a href="2" rel=nofollow rel=nofollow>'
<1> s/<a href="">/<a rel=nofollow href="">/
<8> aaron, people's <a href>s will be more complicated.. they aren't under my control
<1> well,
<8> They may do <a target=_top ...> etc.
<1> so you just want the ones that dont have 'rel' parameters?
<1> attributes
<1> im not the man with regex, but something fairly simple should still work
<8> yeah.. and those that do, I'd want to change it to rel=nofollow :)
<8> oh.. I know what..
<8> hmm.. no, I don't..
<8> If I'm doing the replacement, I don't mind <a (anything other than rel=nofollow) ----> <a rel=nofollow
<8> I just need to say that that part must NOT be "rel=nofollow"
<8> if they have a rel in there, I don't mind
<1> <a href="[^"]*" ([^r][^e][^l][^ ]* |[^ ]{0, 2} )*>
<1> well thats horribly wrong
<1> but something like that
<8> yeah.. but I know what you mean..
<8> I hate that way.. :)
<8> I guess it's not too bad though..
<1> will there be more than one per linE?
<8> hell, I can just check these things by hand then.. hmm
<8> yeah, this is a long message
<8> could be 0 length, could be 8k
<1> that complicates things.
<8> while (strstr("<a[whitespace]")) might work
<1> hehe, beats me, then. my inclination with these things is to revert to a language i know better, like C or C++
<1> yeah
<8> yeah.. I might just do it in C.. easy enough
<1> if you want absolute correctness, there still might be issues though
<8> would come out to the same if (ch[0]!='r' || ch[1]!='e' || ch[2]!='l' .... )
<1> with absolute correctness, you'd need an xml or sgml parser...
<1> i think
<8> I could do that too
<8> I do have my vhtml parser right here, handling other things
<8> lemme look into trying that.. it does run-length output of the code.. I could change its routines for the period of time that the message would be displayed
<8> since the same parser is used for the template processing
<8> thanks, good idea.. I considered then dropped it.. but it probably wouldn't be such a bad idea :)
<8> actually, it still requires extra effort and time, which I don't have at the moment (with the amount of spam I want to stop asap)
<8> rel=nofollow tells google not to raise that person's page rank
<8> they're abusing our msgboards to raise their ranks
<1> ah, i didnt know that. thats interesting.
<8> okay.. gonna get back to this.. thanks Aaron..
<8> nod.
<8> sigh.
<9> amd 4800+ am2 dual core, or dual opteron 270 (2.0Ghz dual core)?
<8> it's just that I have a routine I wrote for regex replacements already.. it would have been convenient to use if posix regex's could handle the non-match requirement
<10> For what kind of tasks, Darken-?
<1> Jaggy, can you use two?
<1> eg, equiv of: grep -v, then sed ?
<10> If your typical data set is going to fit better in the Opteron cache, go for it
<9> Maloeran video rendering
<1> well, that probably wouldn't work like that, because its not line-baswed
<8> aaron, hmmm.. I'm not sure I understand. These are messages posted, so it's going to have ... yeah
<9> i'm split here which would be better
<8> it's basically one long string
<8> in fact, it is :)
<9> the opteron mobo's don't support the new DDR2 that the AM2 cpus support
<10> What kind of rendering, path tracing or other photorealistic rendering?
<9> so I'm limited to 400Mhz modules
<9> not sure Maloeran, the guy I'm building it for uses sony vegas 6.0
<9> he has raw HD tapes
<9> he I guess edits, adds effects to, renders
<8> regex("str (exclude me){0}") would be nice .. if I could force the 0 instead of it being an optional thing :)
<10> The cache of Opterons makes a difference for me, much more than the ram frequency
<8> how do you implement this in regex, anyone know?
<10> for ray-tracing that is
<9> yea I don't think this would be ray tracing
<9> more realistic videos
<9> high def quality
<10> Darken-, DDR2 is not what you want for raytracing and similar processing
<10> The latency is too high, even if the bandwidth is good
<9> you say I should not get an amd AM2 for a video processing server?
<10> I say you should probably get DDR ram, yes
<8> the other thing is that my HTML parser is strict and if there's invalid HTML it'll error out..
<10> There's a good reason why AMD stayed with DDR instead of DDR2
<9> so then why release the am2 with ddr2?
<10> Opterons with their large cache, powerful floating point number crunch, low branch misprediction penalty and low-latency DDR memory is perfect for ray-tracing
<2> I have 1GB of ddr2 somewhere
<2> PC4300
<9> but how does ray tracing compare to hd video rendering?
<10> I suppose DDR2 is beginning to be of some value, it depends if you need low latency or high bandwidth
<9> well that I don't know
<9> it's all so new I can't find any benchmarks
<10> Ray-tracing memory access are highly random, bandwidth hardly matters, latency does
<2> RAS CAS BAS TARD
<10> Trust me, I know ray-tracing well :)
<9> ddr2 800Mhz high latency vs 400Mhz low latency
<1> in this case, it depends mainly on whether your algorithm allows you to have long streams of vector operations.
<2> Maloeran si teh fastest raytracerer in ze WORLDS
<9> Maloeran true, but if he's just going to be rendering high def videos
<9> will this need latency or bandwidth?
<1> long streams of vector operations is ideal. it will benefit most from high bandwidth.
<10> For raytracing, low latency
<9> I didn't say this was ray tracing
<9> I'm asking what it is
<9> is high def realistic video considered.. ray tracing? you're confusing me
<1> the main question is whether you can organize your processing into vector operations.
<10> The answers are likely to be the same for other photorealistic rendering methods
<1> and alsom whether there is significant control flow
<2> Maloeran: got a billion poly model rendered yet? :)
<10> AaronWL, vector operations or not, the ram is accessed in a highly random manner
<10> A large cache and low latency memory is what you want
<10> A billion? Hum no, I don't have that handy... :)
<1> by vector operations, i mean operations on large arrays of data.
<2> i wanna see if you can beat FF:AC
<2> think you can?
<1> i dont consider multiplying two n=3 vectors to be vector operations.
<8> so, I'm doing a global substring-replacement in this string -- I'll have to allocate a new string. How do you guys do this?
<10> AaronWL, rays are your lage arrays of data, the scene partitionning technique is typically walked erratically, hoping around memory
<10> lage -> large
<8> In this case I was thinking of using my list-code to push() each offset, then I could calculate the total needed and not have to do str compares again
<2> Maloeran: You've seen FF:AC right?
<10> What is FF:AC, zid_?
<2> eep
<2> http://en.wikipedia.org/wiki/Advent_children
Return to
#c
or
Go to some related
logs:
#cisco
#politics
#politics
#gentoo
#bsd
HDST.PK
maximum fat32 size 131GB
#microsoft
test phd phyisic
chanel choser
|
|