<kristianpaul> biking around
<kristianpaul> oops
<folkz0> hey
<folkz0> anybody in here awake i have a serioius hardware question
<folkz0> server related
<wolfspra1l> awake yes, but server question, let's see...
<folkz0> good finally
<wolfspra1l> wait until you hear "don't know" :-)
<wolfspra1l> shoot your question...
<folkz0> ok heres the example say we had 2 xeon dual core processors and 1 desktop core i7 quad core proccesor
<folkz0> same clock same ram speed etc
<folkz0> now say we had a program that fully utilizies quad core and all it's features, basically makes use of it
<folkz0> wouldn't the server system with 2 dual cores perform better then the single quad core
<wolfspra1l> hypothetical questions, don't we love them...
<folkz0> a single die cpu with 4 cores would be using up all it's caches within each other
<folkz0> 2 seperate die cpu's would have more space
<folkz0> but then somebody else mentioned the issue about the latency
<wolfspra1l> I'd say that depends very much on who designed the board and the many details that make up the eventual performance of your server application
<wolfspra1l> so the answer is no answer
<folkz0> the latency shouldn't be that big of a impact
<wolfspra1l> best is to try and compare :-)
<folkz0> thats what i said
<folkz0> but i mean
<folkz0> i guess on paper wouldn't the cache
<folkz0> have effect
<wolfspra1l> imho you cannot predict how all these things work together
<wolfspra1l> that's probably harder than trying your actual load and comparing
<folkz0> well if anybody can care to jump in
<folkz0> the latency might be a little higher, but again the design of the board, regardless of that
<folkz0> wouldn't those extra 2mb caches being split up by 2 cores instead of 4
<folkz0> give some notieable performance or is it no for both statements
<wolfspra1l> if you have to pick one system and cannot afford to compare with the other one, pick the cheaper one
<folkz0> i could have sworn i read extra cache helped
<folkz0> no i'm not buying anything
<folkz0> i'm trying to prove a dispute
<folkz0> and this information
<folkz0> would be beneifical as well
<wolfspra1l> sounds like :-)
<folkz0> xeon proccesors can communicate to each other
<folkz0> i7's can't
<wolfspra1l> in my experience you cannot easily calculate the performance difference
<wolfspra1l> but that's not what you need, you need something to settle the dispute
<wolfspra1l> buy a bottle of wine, get drunk, dispute solved
<folkz0> i get it you dont know
<folkz0> man anybody awake
<folkz0> ;x
<folkz0> i'm bored
<wpwrak> wolfspra1l: bah, his problem had a simple answer ;) if the system with more cache is already fully loaded and uses exactly _every_ bit of resources, going to a different system will slow you down :)
<wpwrak> wolfspra1l: alas, the same is true for the system with less cache, as long as it's better in some other regard (e.g., communication latency)
<wpwrak> wolfspra1l: thus both lose their bet ... and you get the wine ;-)
<wolfspra1l> wpwrak: well, of course he dismissed my good faith answer. it's a typical question of people who just pick a fight over something, too little practical problems to take care of :-)
<wolfspra1l> once you are down to such details, on any production server, you cannot tell in advance. That's my experience. so pick the cheaper system, and done.
<wolfspra1l> and if it's so important, compare both systems with real load in advance (though that will rarely be economical)
<wolfspra1l> 2 dual core xeon vs. 1 i7 quad, with some cache differences etc. - bah
<wolfspra1l> try the load
<wolfspra1l> the boards probably designed by two different companies, different chipsets, different memory, etc.
<wolfspra1l> what a question
<wolfspra1l> reminds me of those card games I played in school, with cars, motorcycles, airplanes etc. where various specs won over each other :-)
<wpwrak> yeah, of course. the question was nonsensical as asked
<wpwrak> (card games) ;-))
<wpwrak> those were nice. one i remember was with ships. beat a carrier with a speedboat :)
<wpwrak> roh: by the way, why choose acrylic over polycarbonate ? PC should be lots cheaper than acrylic. is it too brittle ?
<wpwrak> roh: and another question: can you also "print" something with a laser cutter/engraver ? i.e., not just make a groove, but change the color (e.g., transparent to black)
<rjeffries> free open FPGA tutorial  http://www.xess.com/appnotes/FpgasNowWhatBook.pdf  
<rjeffries> he also has XuLa a $70 USD little FPGA dev board http://www.xess.com/prods/prod048.php
<Ayla> rjeffries: note that there's also the XuLa-50 which is cheaper
<Ayla> $40
<rjeffries> Ayla cool. is that from the same guy? he makes everything free and open as best I can tell
<Ayla> rjeffries: yes, see http://www.xess.com/prods/prod047.php
<wpwrak> wolfspra1l: how do you see the possibility of shortening things like http://downloads.qi-hardware.com/people/werner/ at some point in the future ? e.g., dl.qi-hardware.com/p/ ?
<wpwrak> ... /p/werner/, of course
<wpwrak> alas, i made the mistake of not simply picking "w". that would already have saved 5 characters
<wpwrak> trimming downloads and people would save 12 characters
<wpwrak> of course, URLs into git are still much worse ...
<kristianpaul> i think there is already a qi-hw.com
<wpwrak> yes, but i don't think it does anything. at least i haven't found any useful url so far that wouldn't fail if changing qi-hardware to qi-hw
<wpwrak> (i usually check this before complaining, just in case :)
<roh> wpwrak: i think pc was bad to get and it doesnt cut well with a laser
<roh> i think it was too 'thermoplastic' ... it melts at much lower temperatures.. ah.. and i cannot cut anything emitting chloride vapour...
<wpwrak> roh: ah, i see. odd. pc should be a lot more common than acrylic.
<wpwrak> open the window ? ;-)
<roh> because that vapour will not only be bad for me but also the machine. it corrodes the mechanics and is bad for the optics.
<wpwrak> yeah, the cutter may dislike it a bit. that is, if a significant amount of Cl is actually liberated.
<roh> i know nobody who puts stuff containing cloride into his cutter
<wpwrak> so just milling then
<roh> for pc? yes.. but actually it depends on what i need to do.
<wpwrak> actually, are you sure that they add Cl to PC ? i couldn't find any reference
<wpwrak> or do you mean PVC ?
<roh> same game
<roh> ah. could be.. not sure entirely (just awoke from a long night after a long party)
<wpwrak> ;-)
<roh> anyhow.. pc isnt avail in colors by choice.
<wpwrak> i see Cl only in the maufacturing process, but not in the final product
<roh> and material costs are not important compared to the machine running cost anyhow
<lekernel> roh, where did you party? :-)
<roh> stattbad
<roh> always a lot of fun if you got a clubmarke and your own 'backstage'
<wpwrak> interesting according to wikipedia, PC is more expensive than acrylic. hmm.
<roh> eh whut? funny
<wpwrak> yeha, i'm not quite sure i belive this just yet :)
<roh> me neither.. i thought pc is what the coca cola bottles are made from. or was the PET`
<wpwrak> may be PET
<lekernel> ah, another one of those berlin clubs that never tell you the price of the entrance in advance. it's only after you've traveled x kilometers to get there that you know. how annoying.
<roh> ?
<roh> its not a club. and the price is different every time
<wpwrak> hm, accordig to at least one shop, acrylic is indeed a little cheaper than PC: http://acrylicparts.com/lexanprice.html http://acrylicparts.com/acrylicprice.html
<lekernel> depending on how you look? :-)
<roh> no. depending on the group organizing it.
<lekernel> well... they put flyers online for each party, why the heck don't they write the price on it?
<wpwrak> and acrylic can be injection molded, too. excellent. seems that i have a winner :)
<roh> well..if i can injection-mold.. i would use abs+pc
<roh> because its 'schlagfest'
<lekernel> this seems to be an annoying Berlin specialty...
<wpwrak> we'll need that for the death metal edition of M1 ;-)
<wpwrak> specs: "must survive impact of fists, boots, musical instruments"
<lekernel> wpwrak, next version will be probably encased in metal :-)
<DocScrutinizer> that's punk
<lekernel> (next major one)
<lekernel> like mixing consoles etc.
<wpwrak> metal would be nice. a lot more stage-proof :)
<wpwrak> roh will need a mean laser then, though. or bring the mill back up :)
<lekernel> I think CNC and a bending machine are more appropriate there
<wpwrak> DocScrutinizer: hmm, i wonder if any of the wilder styles refrains from smashing their instruments. well, at least they should be relatively safe with ambient and synthpop ;-)
<lekernel> how are mixing console enclosures manufactured?
<lekernel> they look like bent/drilled metal to me...
<wpwrak> lekernel: benging machines sound interesting. i have no idea about that technology.
<DocScrutinizer> usually are
<DocScrutinizer> good ole blacksmith business
<wpwrak> yeah, ancient
<DocScrutinizer> simple steel folding machines (err tools) can be found in virtually each steel workshop
<DocScrutinizer> then usually you do "pulverbeschichtung" after drilling & folding
<wpwrak> what does that do ?
<wpwrak> ah, coating
<qi-bot> [commit] Werner Almesberger: labsw/README: described situation with respect to high voltage (master) http://qi-hw.com/p/wernermisc/b2c8481
<qi-bot> [commit] Werner Almesberger: m1rc3/norruption/LOG: latest result (showing failure to reproduce NOR corruption) (master) http://qi-hw.com/p/wernermisc/165d799
<qi-bot> [commit] Werner Almesberger: labsw/fw/labsw.c: improved debouncing and cleaned up button logic (master) http://qi-hw.com/p/wernermisc/a400f9a
<qi-bot> [commit] Werner Almesberger: labsw/README: all opto headers need wider spacing for high voltage (master) http://qi-hw.com/p/wernermisc/f9c4bb6
<qi-bot> [commit] Werner Almesberger: labsw/web/: Web-based overview page (master) http://qi-hw.com/p/wernermisc/28cdb89
<roh> lekernel: mixing consoles etc are usually stamped out metal sheets, then folded and pulverized.. as werner said
<roh> stamping is much cheaper than milling or drilling
<wpwrak> the pulverizing came from joerg
<roh> ah. sorry ;)
<wpwrak> i wonder if we can mill a usable stamp
<DocScrutinizer> no way
<roh> well.. i think so. atleast on our mill. (use hardened steel and active flood cooling) .. but we havent done that yet
<roh> DocScrutinizer: you should visit us at the congress (we will keep raumfahrtagentur open this year)
<DocScrutinizer> :-D
<DocScrutinizer> not sure if I join congress this year
<roh> already mounted the first canvas for the beamers, second one will follow soon.
<DocScrutinizer> err, we're talking about that event between xmas and new year?
<DocScrutinizer> (storm) here as well
<DocScrutinizer> runs closing windows
<DocScrutinizer> YAY
<DocScrutinizer> o/
<DocScrutinizer> away for dinner, hoping to arrive at pub dry
<DocScrutinizer> too late :-/
<roh> ;)
<DocScrutinizer51> FSCK, that's been the last hour of my DSL modem
<DocScrutinizer51> the differential current switch and fuses were ok after refenabling them, and TV survived too :-)
<DocScrutinizer> alive again \o/
<DocScrutinizer> dang, first time the lightning hit my own house
<roh> DocScrutinizer: uh. everything ok?
<DocScrutinizer> halfway back to normal, both my ears and nerves, and the infra
<roh> uh. lucky.. usually quite some infra gets fried on real direct hits
<DocScrutinizer> ZyXEL DSL exploded
<roh> already exchanged complete streets of ntba after only one house got hit
<DocScrutinizer> has loose parts inside now
<roh> heh. propably the overvoltage components *g*
<DocScrutinizer> yeah
<DocScrutinizer> some dimmers fried, here as well as some hoses away
<roh> dsl/landline equipment has often even 'spark-gaps' on the pcb. be sure to take a look if you're nosy
<roh> you're insured? or only cheap small stuff?
<DocScrutinizer> well, it really stinks and got *very* hot while all LEDs were dead
<DocScrutinizer> I got no ensurance
<roh> mmh.
<DocScrutinizer> gonna get me dinner now, finally
<roh> :) have fun. food is good for the nerves *g*
<DocScrutinizer> then back to switching from PC driven PPPoE to a proper router setup
<DocScrutinizer> thanks and o/
<DocScrutinizer> well, it's been rather ok, I planned to hit this ZyXEL P660 with a sledgehammer since long
<DocScrutinizer> and I lured the lightning with my green laser - it *worked* X-P
<DocScrutinizer> (for those curious: ionizing effects on air need a bit more energy than a 200mW laserpointer can drive)
<DocScrutinizer> and it hit the house, not my hand with laserpointer ;-)
<roh>
<DocScrutinizer51> sub-elipsis statement?
<wpwrak> DocScrutinizer: (laser) see, it hit the house because it wasn't powerful enough :)
<DocScrutinizer> yeah, and it took 20s to ripe, after I stopped beaming up
<DocScrutinizer> actually I'm not sure it was our house, or the neighbour 20m right or left
<wpwrak> pretty potent then
<DocScrutinizer> but I'm sure the lightning didn't come from where I pointed at with the laser, but rather from the "backside" of our house
<DocScrutinizer> (potent) it killed a dimmer 60m away across the street
<wpwrak> wow, you have realtime monitoring of all the dimmers in your street ?
<DocScrutinizer> only of those in my fav pubs ;-)
<wpwrak> so the emergency evacuation procedure is to leave home and go to the pub ? sounds nice :)
<DocScrutinizer> err, yes :-D
<DocScrutinizer> given the fact I was about to leave for dinner when I posted [2011-09-11 20:26:01] <DocScrutinizer> too late :-/
<DocScrutinizer> some 5..10min later lightning stroke, so I thought it's be best to finally get dinner after I checked there's no hidden fires anywhere in the house
<DocScrutinizer> hmm, NFC what blew up in this POS ZyXEL
<DocScrutinizer> BWAHAHAHA red water from the tap
<DocScrutinizer> seems the lightning properly cleaned all the rust out of the tubing
<DocScrutinizer> I already wondered which obscure path the lightning took from our 150yo house's roof down to the ground
<DocScrutinizer> seems now I got a clue
<wpwrak> nice clean water ... soon :)
<DocScrutinizer> yeah
<DocScrutinizer> now for the tedious d-link config
<DocScrutinizer> wonders if it might be simpler to access the config webpage of dlink router via WLAN rather than via a temporary set up wiring and network IP addr
<DocScrutinizer> bbl