2011-09-11 00:23 biking around 2011-09-11 00:23 oops 2011-09-11 07:01 hey 2011-09-11 07:01 anybody in here awake i have a serioius hardware question 2011-09-11 07:01 server related 2011-09-11 07:01 awake yes, but server question, let's see... 2011-09-11 07:02 good finally 2011-09-11 07:02 wait until you hear "don't know" :-) 2011-09-11 07:02 shoot your question... 2011-09-11 07:02 ok heres the example say we had 2 xeon dual core processors and 1 desktop core i7 quad core proccesor 2011-09-11 07:02 same clock same ram speed etc 2011-09-11 07:02 now say we had a program that fully utilizies quad core and all it's features, basically makes use of it 2011-09-11 07:03 wouldn't the server system with 2 dual cores perform better then the single quad core 2011-09-11 07:03 hypothetical questions, don't we love them... 2011-09-11 07:03 a single die cpu with 4 cores would be using up all it's caches within each other 2011-09-11 07:03 2 seperate die cpu's would have more space 2011-09-11 07:03 but then somebody else mentioned the issue about the latency 2011-09-11 07:03 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 2011-09-11 07:04 so the answer is no answer 2011-09-11 07:04 the latency shouldn't be that big of a impact 2011-09-11 07:04 best is to try and compare :-) 2011-09-11 07:04 thats what i said 2011-09-11 07:04 but i mean 2011-09-11 07:04 i guess on paper wouldn't the cache 2011-09-11 07:04 have effect 2011-09-11 07:04 imho you cannot predict how all these things work together 2011-09-11 07:04 that's probably harder than trying your actual load and comparing 2011-09-11 07:05 well if anybody can care to jump in 2011-09-11 07:05 the latency might be a little higher, but again the design of the board, regardless of that 2011-09-11 07:05 wouldn't those extra 2mb caches being split up by 2 cores instead of 4 2011-09-11 07:05 give some notieable performance or is it no for both statements 2011-09-11 07:05 if you have to pick one system and cannot afford to compare with the other one, pick the cheaper one 2011-09-11 07:06 i could have sworn i read extra cache helped 2011-09-11 07:06 no i'm not buying anything 2011-09-11 07:06 i'm trying to prove a dispute 2011-09-11 07:06 and this information 2011-09-11 07:06 would be beneifical as well 2011-09-11 07:06 sounds like :-) 2011-09-11 07:07 xeon proccesors can communicate to each other 2011-09-11 07:07 i7's can't 2011-09-11 07:07 in my experience you cannot easily calculate the performance difference 2011-09-11 07:07 but that's not what you need, you need something to settle the dispute 2011-09-11 07:07 buy a bottle of wine, get drunk, dispute solved 2011-09-11 07:07 i get it you dont know 2011-09-11 07:07 man anybody awake 2011-09-11 07:07 ;x 2011-09-11 07:12 i'm bored 2011-09-11 10:30 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 :) 2011-09-11 10:31 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) 2011-09-11 10:32 wolfspra1l: thus both lose their bet ... and you get the wine ;-) 2011-09-11 11:26 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 :-) 2011-09-11 11:27 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. 2011-09-11 11:27 and if it's so important, compare both systems with real load in advance (though that will rarely be economical) 2011-09-11 11:29 2 dual core xeon vs. 1 i7 quad, with some cache differences etc. - bah 2011-09-11 11:29 try the load 2011-09-11 11:30 the boards probably designed by two different companies, different chipsets, different memory, etc. 2011-09-11 11:30 what a question 2011-09-11 11:31 reminds me of those card games I played in school, with cars, motorcycles, airplanes etc. where various specs won over each other :-) 2011-09-11 11:36 yeah, of course. the question was nonsensical as asked 2011-09-11 11:36 (card games) ;-)) 2011-09-11 11:37 those were nice. one i remember was with ships. beat a carrier with a speedboat :) 2011-09-11 11:43 roh: by the way, why choose acrylic over polycarbonate ? PC should be lots cheaper than acrylic. is it too brittle ? 2011-09-11 11:47 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) 2011-09-11 16:06 free open FPGA tutorial  http://www.xess.com/appnotes/FpgasNowWhatBook.pdf   2011-09-11 16:06 he also has XuLa a $70 USD little FPGA dev board http://www.xess.com/prods/prod048.php 2011-09-11 16:07 rjeffries: note that there's also the XuLa-50 which is cheaper 2011-09-11 16:07 $40 2011-09-11 16:08 Ayla cool. is that from the same guy? he makes everything free and open as best I can tell 2011-09-11 16:08 The build has FAILED, see log here: http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-xburst.full_system-09112011-0411/ 2011-09-11 16:11 rjeffries: yes, see http://www.xess.com/prods/prod047.php 2011-09-11 16:15 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/ ? 2011-09-11 16:15 ... /p/werner/, of course 2011-09-11 16:16 alas, i made the mistake of not simply picking "w". that would already have saved 5 characters 2011-09-11 16:17 trimming downloads and people would save 12 characters 2011-09-11 16:18 of course, URLs into git are still much worse ... 2011-09-11 16:25 i think there is already a qi-hw.com 2011-09-11 16:27 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 2011-09-11 16:27 (i usually check this before complaining, just in case :) 2011-09-11 16:32 wpwrak: i think pc was bad to get and it doesnt cut well with a laser 2011-09-11 16:33 i think it was too 'thermoplastic' ... it melts at much lower temperatures.. ah.. and i cannot cut anything emitting chloride vapour... 2011-09-11 16:33 roh: ah, i see. odd. pc should be a lot more common than acrylic. 2011-09-11 16:33 open the window ? ;-) 2011-09-11 16:33 because that vapour will not only be bad for me but also the machine. it corrodes the mechanics and is bad for the optics. 2011-09-11 16:34 yeah, the cutter may dislike it a bit. that is, if a significant amount of Cl is actually liberated. 2011-09-11 16:36 i know nobody who puts stuff containing cloride into his cutter 2011-09-11 16:36 so just milling then 2011-09-11 16:37 for pc? yes.. but actually it depends on what i need to do. 2011-09-11 16:40 actually, are you sure that they add Cl to PC ? i couldn't find any reference 2011-09-11 16:40 or do you mean PVC ? 2011-09-11 16:41 same game 2011-09-11 16:42 ah. could be.. not sure entirely (just awoke from a long night after a long party) 2011-09-11 16:42 ;-) 2011-09-11 16:42 anyhow.. pc isnt avail in colors by choice. 2011-09-11 16:42 i see Cl only in the maufacturing process, but not in the final product 2011-09-11 16:42 and material costs are not important compared to the machine running cost anyhow 2011-09-11 16:48 roh, where did you party? :-) 2011-09-11 16:49 stattbad 2011-09-11 16:50 always a lot of fun if you got a clubmarke and your own 'backstage' 2011-09-11 16:50 interesting according to wikipedia, PC is more expensive than acrylic. hmm. 2011-09-11 16:50 eh whut? funny 2011-09-11 16:51 yeha, i'm not quite sure i belive this just yet :) 2011-09-11 16:53 me neither.. i thought pc is what the coca cola bottles are made from. or was the PET` 2011-09-11 16:54 may be PET 2011-09-11 16:54 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. 2011-09-11 16:54 ? 2011-09-11 16:55 its not a club. and the price is different every time 2011-09-11 16:55 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 2011-09-11 16:55 depending on how you look? :-) 2011-09-11 16:55 no. depending on the group organizing it. 2011-09-11 16:56 well... they put flyers online for each party, why the heck don't they write the price on it? 2011-09-11 16:56 and acrylic can be injection molded, too. excellent. seems that i have a winner :) 2011-09-11 16:56 well..if i can injection-mold.. i would use abs+pc 2011-09-11 16:56 because its 'schlagfest' 2011-09-11 16:57 this seems to be an annoying Berlin specialty... 2011-09-11 16:57 we'll need that for the death metal edition of M1 ;-) 2011-09-11 16:58 specs: "must survive impact of fists, boots, musical instruments" 2011-09-11 17:00 wpwrak, next version will be probably encased in metal :-) 2011-09-11 17:01 that's punk 2011-09-11 17:01 (next major one) 2011-09-11 17:01 like mixing consoles etc. 2011-09-11 17:02 metal would be nice. a lot more stage-proof :) 2011-09-11 17:02 roh will need a mean laser then, though. or bring the mill back up :) 2011-09-11 17:04 I think CNC and a bending machine are more appropriate there 2011-09-11 17:04 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 ;-) 2011-09-11 17:05 how are mixing console enclosures manufactured? 2011-09-11 17:05 they look like bent/drilled metal to me... 2011-09-11 17:05 lekernel: benging machines sound interesting. i have no idea about that technology. 2011-09-11 17:05 usually are 2011-09-11 17:06 good ole blacksmith business 2011-09-11 17:07 yeah, ancient 2011-09-11 17:08 simple steel folding machines (err tools) can be found in virtually each steel workshop 2011-09-11 17:11 then usually you do "pulverbeschichtung" after drilling & folding 2011-09-11 17:12 what does that do ? 2011-09-11 17:12 http://de.wikipedia.org/wiki/Pulverbeschichtung 2011-09-11 17:13 ah, coating 2011-09-11 17:47 [commit] Werner Almesberger: labsw/README: described situation with respect to high voltage (master) http://qi-hw.com/p/wernermisc/b2c8481 2011-09-11 17:47 [commit] Werner Almesberger: m1rc3/norruption/LOG: latest result (showing failure to reproduce NOR corruption) (master) http://qi-hw.com/p/wernermisc/165d799 2011-09-11 17:47 [commit] Werner Almesberger: labsw/fw/labsw.c: improved debouncing and cleaned up button logic (master) http://qi-hw.com/p/wernermisc/a400f9a 2011-09-11 17:47 [commit] Werner Almesberger: labsw/README: all opto headers need wider spacing for high voltage (master) http://qi-hw.com/p/wernermisc/f9c4bb6 2011-09-11 17:50 [commit] Werner Almesberger: labsw/web/: Web-based overview page (master) http://qi-hw.com/p/wernermisc/28cdb89 2011-09-11 18:12 lekernel: mixing consoles etc are usually stamped out metal sheets, then folded and pulverized.. as werner said 2011-09-11 18:12 stamping is much cheaper than milling or drilling 2011-09-11 18:13 the pulverizing came from joerg 2011-09-11 18:13 ah. sorry ;) 2011-09-11 18:13 i wonder if we can mill a usable stamp 2011-09-11 18:13 no way 2011-09-11 18:13 well.. i think so. atleast on our mill. (use hardened steel and active flood cooling) .. but we havent done that yet 2011-09-11 18:14 DocScrutinizer: you should visit us at the congress (we will keep raumfahrtagentur open this year) 2011-09-11 18:14 :-D 2011-09-11 18:14 not sure if I join congress this year 2011-09-11 18:15 already mounted the first canvas for the beamers, second one will follow soon. 2011-09-11 18:15 meh. storm (tm) http://www.t-online.de/wetter/info/niederschlagsradar.html 2011-09-11 18:16 err, we're talking about that event between xmas and new year? 2011-09-11 18:16 (storm) here as well 2011-09-11 18:16 runs closing windows 2011-09-11 18:17 YAY 2011-09-11 18:18 o/ 2011-09-11 18:18 away for dinner, hoping to arrive at pub dry 2011-09-11 18:20 http://www.unwetterzentrale.de/uwz/getwarning_de.php?xpos=250&ypos=190&bland=bayern&lang=de 2011-09-11 18:26 too late :-/ 2011-09-11 18:26 ;) 2011-09-11 18:38 FSCK, that's been the last hour of my DSL modem 2011-09-11 18:40 the differential current switch and fuses were ok after refenabling them, and TV survived too :-) 2011-09-11 19:53 alive again \o/ 2011-09-11 19:53 dang, first time the lightning hit my own house 2011-09-11 20:14 DocScrutinizer: uh. everything ok? 2011-09-11 20:15 halfway back to normal, both my ears and nerves, and the infra 2011-09-11 20:15 uh. lucky.. usually quite some infra gets fried on real direct hits 2011-09-11 20:15 ZyXEL DSL exploded 2011-09-11 20:16 already exchanged complete streets of ntba after only one house got hit 2011-09-11 20:16 has loose parts inside now 2011-09-11 20:16 heh. propably the overvoltage components *g* 2011-09-11 20:16 yeah 2011-09-11 20:16 some dimmers fried, here as well as some hoses away 2011-09-11 20:16 dsl/landline equipment has often even 'spark-gaps' on the pcb. be sure to take a look if you're nosy 2011-09-11 20:17 you're insured? or only cheap small stuff? 2011-09-11 20:17 well, it really stinks and got *very* hot while all LEDs were dead 2011-09-11 20:17 I got no ensurance 2011-09-11 20:17 mmh. 2011-09-11 20:18 gonna get me dinner now, finally 2011-09-11 20:18 :) have fun. food is good for the nerves *g* 2011-09-11 20:18 then back to switching from PC driven PPPoE to a proper router setup 2011-09-11 20:19 thanks and o/ 2011-09-11 20:20 well, it's been rather ok, I planned to hit this ZyXEL P660 with a sledgehammer since long 2011-09-11 20:20 and I lured the lightning with my green laser - it *worked* X-P 2011-09-11 20:22 (for those curious: ionizing effects on air need a bit more energy than a 200mW laserpointer can drive) 2011-09-11 20:22 and it hit the house, not my hand with laserpointer ;-) 2011-09-11 20:31 2011-09-11 21:37 sub-elipsis statement? 2011-09-11 22:01 DocScrutinizer: (laser) see, it hit the house because it wasn't powerful enough :) 2011-09-11 22:13 yeah, and it took 20s to ripe, after I stopped beaming up 2011-09-11 22:15 actually I'm not sure it was our house, or the neighbour 20m right or left 2011-09-11 22:15 pretty potent then 2011-09-11 22:16 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 2011-09-11 22:16 (potent) it killed a dimmer 60m away across the street 2011-09-11 22:17 wow, you have realtime monitoring of all the dimmers in your street ? 2011-09-11 22:18 only of those in my fav pubs ;-) 2011-09-11 22:18 so the emergency evacuation procedure is to leave home and go to the pub ? sounds nice :) 2011-09-11 22:18 err, yes :-D 2011-09-11 22:20 given the fact I was about to leave for dinner when I posted [2011-09-11 20:26:01] too late :-/ 2011-09-11 22:21 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 2011-09-11 22:29 hmm, NFC what blew up in this POS ZyXEL 2011-09-11 22:33 BWAHAHAHA red water from the tap 2011-09-11 22:33 seems the lightning properly cleaned all the rust out of the tubing 2011-09-11 22:35 I already wondered which obscure path the lightning took from our 150yo house's roof down to the ground 2011-09-11 22:35 seems now I got a clue 2011-09-11 22:40 nice clean water ... soon :) 2011-09-11 22:40 yeah 2011-09-11 22:41 now for the tedious d-link config 2011-09-11 22:42 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 2011-09-11 22:48 bbl