▁▁▁▁▁▁⏐︎▁ 5686
asciilifeform: neato mod6 .
asciilifeform: incidentally how did you power yours ?
asciilifeform: (usb hub ? direct plug into pc ?)
asciilifeform: if using hub -- be sure to use powered hub.
mircea_popescu: asciilifeform seems looping over 1g is rather nonsensical.
asciilifeform: mircea_popescu: dieharder's ~tests~ loop.
mircea_popescu: i know. and i don't see why they would.
mircea_popescu: which is why i said should prolly talk to guy see what he says
mod6: I'm using the USB-TTL cable that I bought, separate from the one shipped with the fg. There's a picture of the connection up close in the blog post. It is plugged directly into the computer.
asciilifeform: mircea_popescu wouldja try and get him in here ? pretty pleez?
mircea_popescu: i dont have an email!
asciilifeform: mod6: nice.
asciilifeform: hanbot around?
mod6: lemme see if I can dig him up
mod6: rgb@phy.duke.edu
mircea_popescu: win.
mod6: (Robert G. Brown)
mod6: copyright at this www.phy.duke.edu/~rgb/General/dieharder.php page is denoted as 2017, so he must still be ~somewhat~ active.
mod6: version listed at the site is 3.31.1, which is the same version I'm using.
mod6: (see posted output files on my site)
mircea_popescu: http://btcbase.org/log/2017-04-17#1645221 << were you any good at that ? ☝︎☟︎
a111: Logged on 2017-04-17 21:04 phf: i used to promote a bar in the same block (tom tom's, closed now), so i used to spend a lot of time in this area
asciilifeform: mod6: i have 3.31.1 also.
mod6: ok good to know.
mircea_popescu: http://btcbase.org/log/2017-04-17#1645236 ><< speaking of, there's a whole "ted talks costa rica" "culture". ☝︎
a111: Logged on 2017-04-17 21:18 mats: https://www.eventbrite.com/e/the-art-of-giving-a-tedx-talk-tickets-32306198710
asciilifeform: prng, e.g., aes(aes(...(0))) , or urandom, etc. give uniform distrib. and quickly ideal montecarlo for 'pi', etc. but , interestingly, same sort of dieharder out as FG, in that the level of 'happy' is directly proportional to the size of the collected blob, or rather, inversely proportional to # of the idiot rewinds.
asciilifeform: http://qubit.lncc.br/trng << this d00d must have been very frustrated. i wonder if he understood the above.
asciilifeform: now at the rate dh eats bytes, a fully rewind-free test of, e.g., FG would take...
asciilifeform: ~6 years.
asciilifeform: http://archive.li/SnwEH << recall these ? << illustration of the rewind thing.
mod6: <+asciilifeform> http://archive.li/SnwEH << recall these ? << illustration of the rewind thing. << ah, interesting. who ran that, anyone recall?
asciilifeform: !#s http://wotpaste.cascadianhacker.com/pastes/ehdip/?raw=true
a111: 1 result for "http://wotpaste.cascadianhacker.com/pastes/ehdip/?raw=true", http://btcbase.org/log-search?q=http%3A%2F%2Fwotpaste.cascadianhacker.com%2Fpastes%2Fehdip%2F%3Fraw%3Dtrue
asciilifeform: looks like it was ben_vulpes
mod6: i wonder if that was the one from osx.
asciilifeform: mod6: my point was, small samples look like shit on dh, because 1000s of 'rewind'
asciilifeform: regardless of where you got'em
mod6: ah
asciilifeform: and, apparently, the same tests weaken/fail in proportion to rewinds, regardless of whose rng you used, or even prng.
asciilifeform: which i find quite odd.
asciilifeform: ( i am speaking here strictly of rngs which ~do~ eventually converge -- as defined by 100% 'pass' -- at some realistically attainable sample size. )
asciilifeform: dh also does a very, imho, questionable thing, of throwing away most of the input (yes) on account of insisting on using new segment of the input for each test
ben_vulpes: that last tidbit is interesting
ben_vulpes: also the weakening tests
asciilifeform: ben_vulpes: try it yourself
asciilifeform: (with urandom or whatever else is in your box conveniently)
mod6: wonder if the intent is to provide greater coverage of the pool
asciilifeform: mod6: i still have 0 idea why author did this.
mod6: even if, misguided.
asciilifeform: it'd be easy to patch around. but i wanted to begin with classical dh.
ben_vulpes: the last one strikes me as terrible experimental design
asciilifeform: ben_vulpes: i quite agree
mod6: i'd like to have the discussion with rgb himself and see what the intent was there.
ben_vulpes: sure yeah
asciilifeform: that would be pretty neat.
ben_vulpes bbl, BABIES
mod6: l8r
asciilifeform: the sad thing is, dh was mostly used historically to test prng. so the 'waste' was, i picture, not seen as problem.
mod6: I think I just reached the end of my eatblock test...
asciilifeform: congrats mod6
mod6: height=454254
asciilifeform: lessee the numberz!
mod6: aight, lemme work on getting a post put up here.
mod6: 9181:40 = ~6.375 days
asciilifeform: this was on a what?
mod6: 2x Xeon 5650 (24 core) / 64Gb RAM / Kingston 1Tb SSD
mod6: ok
asciilifeform: ^ all but the last is mostly wasted on trb
mod6: yah. this thing is like single thread.
mod6: save your lulz for the charts
asciilifeform: yeah, i'ma bbl.
mod6: ok c-ya
mircea_popescu: asciilifeform> which i find quite odd. <<
mircea_popescu: mod6 killing it with the quality blog posts
mircea_popescu: !!up joerose72
deedbot: joerose72 voiced for 30 minutes.
asciilifeform: in other lulz, https://archive.is/5L7Ov >> BingoBoingo , qntra ?
asciilifeform: ( https://archive.is/5L7Ov/image gets rid of the js idiocy )
mircea_popescu: eh, bitfinex.
asciilifeform was a little surprised to learn that it was yet alive
mircea_popescu: pretty much been trying to ride the news since day 0. sorta like the divorcee in rhodes island pretending she can't sleep because she's heard all the new york chicks got insomnia.
trinque been out all day, replacing shitty broken laptop
trinque: x1 carbon's got bad ram, actually freezes its own firmware memory test.
mircea_popescu: o.O
trinque: gives one fail on "address test" before it's gone for good.
trinque: I put one of those new AMD chip boxes together today; no moar thinputers.
trinque: mircea_popescu: wot.deedbot.org is my wotviz thinger, needs an SVG capable browser and not much else
mircea_popescu: yes it is splendid. i linked it.
trinque: ah cool
trinque on 80char black and white display atm. obsd doesn't know what to do with this gpu
mircea_popescu: no idea why i didn't before, but anyway. doesn't seem to have gone down from it either, so all good.
asciilifeform: trinque: is that one of those boxes with soldered-down ram ?
trinque: asciilifeform: but of course
asciilifeform: ain't cheap, either. was there an up-side?
trinque: got suckered in by the display resolution, 2560 by something.
trinque: great fonts
trinque: but no, that was the last purchase of that kind of item.
mircea_popescu: getting rid of the whole "laptop" paradigm is great for both wrists and productivity anyways.
asciilifeform: the big one was decent (w540)
asciilifeform: still not a real comp, but you can't fit a real comp in airplane tray
asciilifeform: no soldered-down ram, fwiw.
asciilifeform: (even cpu -- removable)
mircea_popescu: what % of your time last year did you spend in an airplane and what exactly would have happened had you taken a nap / chatted up random chicks / thought about your nails for the interval.
asciilifeform: me -- ~none. but i also didn't use the stinkpad that year.
mircea_popescu: i spent what, 0.3% or so of this year in airplane. had thinkpad with multiple batteries loaded up with films etc by diligent travel assistant / cocksucker. in the overhead tray it stayed.
deedbot: http://qntra.net/2017/04/buttfunexs-fiat-banking-partners-move-towards-becoming-exs/ << Qntra - Buttfunex's Fiat Banking Partners Move Towards Becoming Ex's
BingoBoingo: Ex's in the Butt Fun!
BingoBoingo: In other derp: "The researchers dubbed the reorganization an act of "rapid river piracy," saying that such events had often occurred in the Earth's geologic past, but never before, to their knowledge, as a sudden present-day event. They also called it "geologically instantaneous."" << Ignoring... The Mississippi River's frequent rerouting since the days the French found it. 'Member when the Illinois state capital was in Kaskaski, 2010 ce
BingoBoingo: nsus population 14'
pete_dushenski: asciilifeform: 0 word re where it was for ~2wks << "BitBet Mod 17-04-17 at 14:49 Dear BitBet users, Our apologies for the extended downtime. We underwent a ddos attack, motivated by a rather base extortion attempt (the 4th or 5th this year, but this one was unusually large). We have taken steps with our ISP to buy additional ddos protection capacity and are back to normal operations. All funds under our ☟︎
pete_dushenski: custody remain - as usual - safe, and as our standard policy dictates, not a single satoshi was paid our to the attackers." (from comment on bitbet page)
pete_dushenski: new bbet policy #1 : communicate with terrorists but don't negotiate with them
pete_dushenski: new bbet policy #2 : don't communicate with users but do negotiate with them
mod6: Alright.
mod6: mircea_popescu, asciilifeform, ben_vulpes, et. al: Eatblock test results & blog post: http://www.mod6.net/eatblock-test/
mircea_popescu: o look at that slim checkblock
pete_dushenski: http://www.mod6.net/eatblock-test/charts/Memory.png << whoa ☟︎
mod6: haha
ben_vulpes: pretty cool, barely a week
mod6: thing was /flying/
mod6: did never started on fire... so that was good.
mod6: *disk
mod6: <+mircea_popescu> o look at that slim checkblock << thought it might be cool to do some perl, parse out the stats, do some calcs & provide. ☟︎
asciilifeform: http://btcbase.org/log/2017-04-18#1645414 << systemwide free mem is a 100% worthless stat on linux box, if correctly working box -- it'll be ~0 -- disk cache ☝︎
a111: Logged on 2017-04-18 05:31 pete_dushenski: http://www.mod6.net/eatblock-test/charts/Memory.png << whoa
asciilifeform: mod6: where's the per-block db stall time ? or still perling it out from debug.log?
mircea_popescu: http://btcbase.org/log/2017-04-18#1645420 << might be good to link definitions of "checkblock" etc on the page. ☝︎
a111: Logged on 2017-04-18 05:39 mod6: <+mircea_popescu> o look at that slim checkblock << thought it might be cool to do some perl, parse out the stats, do some calcs & provide.
mircea_popescu: and yes, what alf says. the per block thing
mod6: lemme take a quick look
mod6: I don't see a "stall time".
mod6: Let me illustrate
asciilifeform: mod6: the read/write times
asciilifeform: y'know, when trb sits like idiot and waits for bdb to disk i/o.
mod6: (btw, the full log is available via the blog), but here's a snippit of one block eaten
mod6: http://p.bvulpes.com/pastes/QCBvH/?raw=true
asciilifeform: where's the db read wait ?
asciilifeform: that was in the last patch
mod6: my checkblock stats are mean/median of all of those values
mod6: accept block, process block, and db write wait.
mod6: that was in the odometer?
asciilifeform: http://therealbitcoin.org/ml/btc-dev/attachments/20170226/asciilifeform_blackhole_reads.vpatch <<
asciilifeform: anyway looks like you don't have this
asciilifeform: plot the write times plox ?
asciilifeform: ( the avg is not very useful, it includes megatonne of 0/handful-tx blocks from early years)
asciilifeform: so the two curves would be:
asciilifeform: 1) total ProcessBlock time (vertical) -- block # (horiz.)
asciilifeform: 2) db write wait time (vertical) -- block # (horiz.)
mod6: <+asciilifeform> anyway looks like you don't have this << ahh, sadly, no.
asciilifeform: not a big deal, read+write ~= total ProcessBlock time
asciilifeform: ( from my earlier findings . )
mod6: ok, i'll see what I can do about getting those two curves put together today.
asciilifeform: neato, ty for the sweat mod6
mod6: <+asciilifeform> not a big deal, read+write ~= total ProcessBlock time << makes sense. stangely, i overlooked that last patch
mod6: asciilifeform: np all tall, my pleasure
asciilifeform: in other lulz, 0day in : minicom. >> https://archive.is/fIsCp
mircea_popescu: meanwhile in berkley, http://68.media.tumblr.com/603b50d8351a3ff3b54e4f4ffdc8c702/tumblr_n8nbfqqovl1spdf5ao1_1280.jpg
BingoBoingo: !~ticker --market all
jhvh1: BingoBoingo: Bitstamp BTCUSD last: 1202.0, vol: 5500.07229829 | BTC-E BTCUSD last: 1217.755, vol: 6831.59355 | Bitfinex BTCUSD last: 1263.0, vol: 19196.08120282 | BTCChina BTCUSD last: 1030.177, vol: 3219.74830000 | Kraken BTCUSD last: 1205.6, vol: 2182.4943908 | Volume-weighted last average: 1221.85440241
asciilifeform: BingoBoingo: https://archive.is/m66V5 << lulzies/qntra
BingoBoingo: lulzies, maybe qntra if more relatit shoes drop
deedbot: http://phuctor.nosuchlabs.com/gpgkey/909DFE9E2F3DED604BE70F6DF2B433F01B8893DE2B2FCE532D57511CD10A97BC << Recent Phuctorings. - Phuctored: 1407...9989 divides RSA Moduli belonging to ' (ssh-rsa key from (13-14 June 2016 extraction) for Phuctor import. Ask asciilifeform or framedragger on Freenode, or email fd at mkj dot lt) <ssh...lt>; ' (lvps87-230-33-12.dedicated.hosteurope.de. DE NW)
deedbot: http://phuctor.nosuchlabs.com/gpgkey/909DFE9E2F3DED604BE70F6DF2B433F01B8893DE2B2FCE532D57511CD10A97BC << Recent Phuctorings. - Phuctored: 1630...5117 divides RSA Moduli belonging to ' (ssh-rsa key from (13-14 June 2016 extraction) for Phuctor import. Ask asciilifeform or framedragger on Freenode, or email fd at mkj dot lt) <ssh...lt>; ' (lvps87-230-33-12.dedicated.hosteurope.de. DE NW)
shinohai: https://fossbytes.com/how-to-make-your-ubuntu-linux-look-like-windows-10/
ben_vulpes: gross
asciilifeform: ew
asciilifeform: why am i reading this
ben_vulpes: i had a similar experience recently, asciilifeform
ben_vulpes: this showed up in my house: https://www.amazon.com/Activist-Innosanto-Nagara/dp/1609805399
ben_vulpes: and if you thought the ideological brainrot was bad, the writing is /even worse/.
ben_vulpes: which tbqh i did not think was possible
ben_vulpes: t i shit thee not is for trans
asciilifeform: 'c is for co-op, cooperating cultures, creative counter to corporate vultures.'
asciilifeform: lol!
ben_vulpes: yeah and corporate is capitalized but vulture oddly not
ben_vulpes: zero rhyme, and let us not speak of the reason.
asciilifeform: where'dya get this
asciilifeform: bought for lulz?
ben_vulpes: no i prohibited spending money on it when i saw the cover for the first time months ago
ben_vulpes: someone else, possibly 2/10 troll
ben_vulpes: one of the $libgirls in $othertown
ben_vulpes: bad bad agitprop masquerading as a children's book
asciilifeform: would be lulzy to make a subtle parody of this masterpiece, and sneak into bookstores, with forged barcode.
asciilifeform: 'c is for clit...'
asciilifeform: 'i is for islam..'
ben_vulpes: board books are costly, but yes
ben_vulpes: hah uh huh
ben_vulpes: "Kings are fine for storytime./ Knights are fun to play./ But when we make decisions/ we will choose the people's way!" ☟︎
ben_vulpes: "baby why do you hate it?" "because it's sooo baaaad"
mircea_popescu: aaand in other adventures, http://68.media.tumblr.com/04eca2b5bfb1718f4d4fb491cf67fd37/tumblr_ooizvk3dbH1ust2lpo1_1280.jpg
asciilifeform: mircea_popescu: pretty great scenery
mircea_popescu: great again, you mean.
lobbes: So, I'm midway through my first gentoo adventure. Currently on the compile kernel step (genkernel), but running into funkiness with uClibc errors. My question is: if I abandon uClibc for, say, glibc, will I have issues building trb? (I remember reading in logz that trb doesn't use glibc)
lobbes: Perhaps musl is better option? Fwiw, I posted over on gentoo forumz with my specifics, but am not versed enough to know if the suggestions they gave (e.g. using glibc) will fuck me over building trb or not: https://forums.gentoo.org/viewtopic-t-1062324.html?sid=c3ea68da31445ec3e870e5344a443dd3
asciilifeform: lobbes: uclibc definitely won't work.
asciilifeform: it's missing a bunch of essentials.
asciilifeform: for trb, that is.
lobbes: Aha, well this makes my decisions easier already
asciilifeform: glibc is also not supported for trb.
asciilifeform: ( it fails to link statically )
ben_vulpes: trb builds with buildroot though, does that with which the kernel is compiled affect that pipeline?
asciilifeform: nope
asciilifeform: but you gotta make sure to have musltronic gcc.
asciilifeform: otherwise no dice.
asciilifeform: ( this was, if anyone recalls, the whole reason i brought in the buildroot thing )
asciilifeform: in mod6's buildtron, the kernel isn't even used.
asciilifeform: ( but presumably lobbes is trying for a musltronic linux -- probably oughta ask trinque , iirc he has one going )
lobbes: Well, I'm just trying to stand up a gentoo that'll run trb. Seems like my kernel choice may not be as important as I thought as long as gcc is musltronic?
asciilifeform: whole point of a static bin
asciilifeform: is that it will run, in principle,
asciilifeform: on any reasonable linux.
asciilifeform: regardless of how built.
asciilifeform: and be unaffected by local liquishit.
lobbes: Aha (Sorry, most of this is over my head still)
lobbes: Wow, that makes sense all of a sudden
trinque: https://wiki.gentoo.org/wiki/Project:Hardened_musl
trinque: mostly, it's 1) start from a musltronic stage3 (they're present on the mirrors in iirc "experimental") and 2) install layman, add musl overlay
trinque: if you wanted to start from stage1, there'd be additional steps involving selecting the right portage profile
trinque: but yeah, not required to have musltronic trb
lobbes: Nifty! thanks trinque/asciilifeform
mircea_popescu: http://btcbase.org/log/2017-04-18#1645485 << yeah well, when the people actually get a way, or for that matter a clue, maybe we revisit this. ☝︎
a111: Logged on 2017-04-18 17:16 ben_vulpes: "Kings are fine for storytime./ Knights are fun to play./ But when we make decisions/ we will choose the people's way!"
mircea_popescu: oh, and in "trends and fashions of wedding parties today", https://68.media.tumblr.com/036dc4e5d5e3ac795a421c403de31b76/tumblr_o1m2tuV3ym1uzp3feo1_400.gif
asciilifeform: in other noose, 'Steve Stephens was spotted this morning by PSP members in Erie County. After a brief pursuit, Stephens shot and killed himself.' -- re: recent desperado
ben_vulpes: PSP?
asciilifeform: PA state police.
ben_vulpes: ah
ben_vulpes: 'members' threw me off
mircea_popescu: who was this ?
asciilifeform: some d00d
mircea_popescu: clearly.
asciilifeform: the one in http://qntra.net/2017/04/easter-manhunt-in-us .
mircea_popescu: a
asciilifeform: https://archive.is/ts6e4 << in other manchurian candidates.
asciilifeform: 'The man accused of fatally shooting five people in a Washington state mall last year has been found dead in his jail cell, authorities said Monday. ... When police confronted the suspect, he froze and complied ... was unarmed and silent, "kind of zombie-like," ... emigrated from Turkey and was a legal permanent resident'
mats: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2460568 a comparison of game theory strategies: always-defect, always-cooperate, tit-for-tat and win-stay, lose-shift
Framedragger: mats: why this particular paper? unless no good comparisons till now, but kinda hard to believe? just curious.
asciilifeform: mats: if you're an aficionado, you'll like http://btcbase.org/log/2014-06-22#728931 ☝︎
a111: Logged on 2014-06-22 17:11 asciilifeform: (think: axelrod's classic 'prisoner' tournament - 'on steroids')
mats: Framedragger: came up in a discussion re: dprk engagement, regarding how one might 'error correct' to prevent conflict spiraling
mats: i've little exposure to game theory beyond iterated tit-for-tat so i found this interesting
mod6: im having trouble making a graph with nearly a million datapoints in it. ☟︎
Framedragger: mats: ah, sounds cool :)
mod6: drawing of the graph is crashing LibreOffice Calc.
mats: everything crashes libreoffice calc in my experience
mod6: any other graph tool suggestions?
ben_vulpes: excel
ben_vulpes ducks
mod6: lol
mats: matplotlib, but iirc you're a perl-er
mircea_popescu: mod6 use the tool we use in eulora!
mircea_popescu: recall, all the mining maps
mod6: ah. ok.
mod6: infact, this thing crashed my entire computer. pff
mod6: had to hardpower off the fucking cocksucker
mircea_popescu: in other hardpower lulz, chick's icemaker in the fridge stopped working, just this very sad sound and no ice forthcoming. she called the very helpful super (young guy, i suppose he likes her). quoth he : "sometimes when the power goes out you have to reset these. you know, like computers ? power it down overnight and see in the morning."
ben_vulpes: second gross of the day
mats: https://arabic.sputniknews.com/arab_world/201704181023545107-%D8%AF%D8%A7%D8%B9%D8%B4-%D8%A7%D9%84%D8%A8%D8%BA%D8%AF%D8%A7%D8%AF%D9%8A-%D8%B3%D9%88%D8%B1%D9%8A%D8%A7 ru news reports capture of abu bakr al-baghdadi
ben_vulpes: keep 'em comin boys
mircea_popescu: quoth her : "fuck me, i bet i know what happened. power went out for like 10 minutes yest, musta been the water duct froze in the interval then power being back on it maintains the ice cork and it can't make more ice."
mircea_popescu: quoth me : "so power it down, open freezer door, put glass of hot water close to conduit in question."
mircea_popescu: win-win-lulz.
Framedragger: smooth troubleshooting!
asciilifeform: http://btcbase.org/log/2017-04-18#1645539 << gnuplot ☝︎
a111: Logged on 2017-04-18 18:23 mod6: im having trouble making a graph with nearly a million datapoints in it.
mircea_popescu: aha.
asciilifeform: strictly.
asciilifeform: anything else -- waste of time
diana_coman: asciilifeform, that's what "we use in eulora"
mircea_popescu is sitting pretty on a large ball of synergy, pulling on the threads.
mod6: ok
mod6: is this kinda what you're looking for? http://www.mod6.net/eatblock-test/ProcessBlockTimeVDBWaitTime.png
mircea_popescu: o hey nice
mod6: oh wit.
mod6: *wait.
mod6: this can't be right...
mircea_popescu: oh ?
mod6: well, maybe it is? i see some spikes in the 30000ms range... didn't think it was ever that high.
mod6: looking through the raw data now...
mircea_popescu: worth a check
mod6: $ cat graph.txt | awk '{print $2}' | perl -e '@a=<STDIN>; foreach(@a) { chomp($_); if($_ > 25000) { print "$_\n"; }}' | wc -l
mod6: 10
mod6: oh yah! guess there are. :]
mircea_popescu: :p
mircea_popescu: spotting this sort of thing is why graphs exist in the first place.
mod6: gnuplot is actually really cool. can't believe i never used it before.
mod6: thanks a LOT to diana_coman who helped me :]
mircea_popescu: can't believe you didn't start using it once you saw all of us use it in eulora lol\
mod6: i actually had no idea how you guys were doing the charting.
mod6: should have asked.
mircea_popescu: she made a post about it, lemme see
diana_coman: cool graph mod6
mircea_popescu: http://www.dianacoman.com/2015/10/13/hic-sunt-flotsams-on-eulora-or-the-brand-new-foxymaps/ <
mod6: ya, turned out cool 'eh? thanks for your help. looks sweet.
mod6: oh
mod6 looks
mircea_popescu: well at least didn't hose whole machine eh ? :D
mod6: ahhh right, i remember seeing this.
mod6: haha, yeah, it did grind it down for like 10 seconds, but then it was fine
mircea_popescu: says gnuplot right there
mod6: yeah, stuff falls out of my head sometimes.
mod6: i'll update my blog
shinohai: Heh I remember reading diana_coman 's article and spent whole weekend plotting shit and writing scripts
mircea_popescu: lol
diana_coman can't quite believe that's from 2015 already, ha
mircea_popescu: shit accumulates!
shinohai: I didn't discover it until late last year lol
mircea_popescu: thanks god you took the hour to write it out so two years later i can link it.
diana_coman: there is that
mod6: yeah, srsly.
mod6: and it looks like 69x easier than me gimp'ing over some eulora screenshots.
mod6: [blog updated]
mircea_popescu: mod6 yeah, because it autoprocessed the bot log anyway. you dun gotta do nuttin.
mod6: that's slick
asciilifeform: mod6: see ye olde http://therealbitcoin.org/ml/btc-dev/2015-July/000107.html << there is a pythong and gp script in there
mod6: that's slick
asciilifeform: ( the one i used for the plots )
asciilifeform: oh nm
asciilifeform: looks like mod6 already has
asciilifeform: lol
mircea_popescu: lol
asciilifeform catches up with l0gz
asciilifeform: gotta wonder, what happened in the spikes
asciilifeform: 'stress tests' ?
asciilifeform: ( folx playing with mining 'db-pessimized' blocks... ? )
mircea_popescu: the one concerning point is that the growth has not yet plateau'd.
asciilifeform: mircea_popescu: didn't i tellya it'll be geometric ?
mircea_popescu: so far looks flatly linear. but...
mircea_popescu: early on everything does.
asciilifeform: aha.
asciilifeform: you can trivially derive the fact of it being not at all linear.
asciilifeform: eventually (given the extant turdball) it'll be >10min, and party's over.
mircea_popescu: it's not that far off as it is. bout halfway there.
ben_vulpes: eh we could just give up on verification
asciilifeform: 'we'
mircea_popescu: ben_vulpes has been trying to troll all day, and he sucks at it!
ben_vulpes: this is snark, not trolling
mircea_popescu: snark is trolling.
ben_vulpes: trolling is /fishing/
mircea_popescu: like sarcasm is wit.
mircea_popescu: ben_vulpes that's trawling.
ben_vulpes: one trawls for lolz, trolls for trout.
mircea_popescu: trout is not an ocean fish.
ben_vulpes: it's a freshwater river thing.
asciilifeform: mod6: can haz one moar plot -- just 300,000 and up, plox ?
mircea_popescu: !~google trawler
jhvh1: mircea_popescu: Browse Trawler boats for sale - YachtWorld.com: <http://www.yachtworld.com/boats/Power/Trawler>; Trawler - Wikipedia: <https://en.wikipedia.org/wiki/Trawler>; Trawlers For Sale By Size - Used Trawlers - New Trawlers - Yacht ...: <http://www.denisonyachtsales.com/buying-your-new-boat/trawlers-for-sale-by-size/>
asciilifeform: mod6: also legend oughta read 'db write wait time'
asciilifeform: ( the overall db wait time is ~= the total )
ben_vulpes: !~google trout trolling
jhvh1: ben_vulpes: Trolling For Rainbow Trout In Still Waters - In-Fisherman: <http://www.in-fisherman.com/trout-salmon/rainbow-trout-steelhead/trolling-for-rainbow-trout-in-still-waters/>; Four Ways to Troll for Trout , Salmon, Crappies, Walleyes, and ...: <http://www.fieldandstream.com/articles/fishing/trout-fishing/where-fish-trout/2009/07/summer-trolling>; Trout Trolling Techniques - Game & Fish: (1 more message)
mircea_popescu: meanwhile at berkley, http://68.media.tumblr.com/d132ecec41abba7b99143fb41f52d4c0/tumblr_nfzg9cHlYB1qetydwo1_1280.jpg
mircea_popescu: asciilifeform possibly time to re-do this whole eatblock instrumentation with a per-line timing profiler.
asciilifeform: mircea_popescu: i DID THIS
asciilifeform: omfg
asciilifeform: there's exactly 1 segment that accounted for ~99% of time.
asciilifeform: and i instrumented it.
asciilifeform: thread is in the logs.
mircea_popescu: no, i know. but look at the graph. so what ~exactly~ is this 'db write wait time' that is ~= the total ?
asciilifeform: e.g.,
asciilifeform: ProcessBlock (res == 1) took : 167839ms; db write wait: 130117ms; db read wait: 21201ms
asciilifeform: i posted a bunch of these.
mircea_popescu: can i has corresponding code snippet that was executing ?
asciilifeform: aha!
mircea_popescu: nevermind the bunch of these, i want the whole eatblock'd blockchain.
asciilifeform: http://therealbitcoin.org/ml/btc-dev/attachments/20170226/asciilifeform_blackhole_reads.vpatch
mircea_popescu: ie, ALL, no bunch.
asciilifeform: mircea_popescu: that's what mod6 was doing. except that he forgot the 'read' patch ^ and only had total and write.
mircea_popescu: mod6 you up for a re-do ?
asciilifeform: also fwiw his total block delays (red) match the ones on my ssd node (zoolag)
asciilifeform: roughly.
mod6: i certainly could.
mod6: so just add in this patch: http://therealbitcoin.org/ml/btc-dev/2017-February/000256.html ?
mircea_popescu: would be useful to have the whole saga, and you're all set up for it, so should be easier than reconstructing the setup later.
mod6: i'm up for a full re-eatblock. just want to make sure I'm not missing anything this time :]
asciilifeform: mod6: you need asciilifeform_blackhole_odometer.vpatch + above, aha.
mod6: ok.
mod6: let me get that setup.
asciilifeform: one more thing
mod6: ok
asciilifeform: (optional) make it print total odometers on shutdown
mod6: how do i do that?
asciilifeform: so we find out total % of the node's bringup time spent waiting for bdb
asciilifeform: make a, e.g.,
asciilifeform: int64 dbTotal = 0; somewhere,
asciilifeform: and dbTotal += res ; say, after the print.
asciilifeform: then where the shutdown routine is, print it.
mod6: im certain to screw that up
asciilifeform: anyway this is not essential.
mircea_popescu: lettuce take a moment and make this one proper.
mod6: before I recompile everything. i'll check with you again.
mod6: i gotta recut all the blocks anyway.
mod6: i just deleted them this am
mircea_popescu: oh
asciilifeform: mod6: i recommend to keep'em around
asciilifeform: they are quite handy.
mod6: for whom? it's like 100G
asciilifeform: handy for experiments and what-ifs of many sorts.
asciilifeform: i quite often refer to the raw blox.
asciilifeform: put'em on an old hdd.
mod6: ok, maybe might have to just another drive and drop 'em on there.
asciilifeform: in other noose!
asciilifeform: i may have found a bug in dd.
asciilifeform: of the worst sort, the 'folx in the know, know about it, and it will never be fixed'
mircea_popescu: as a general policy, when doing things of this nature (publishing intertesting stuff in forum) keep the whole echafaudage for day+ while people comment.
ben_vulpes: weeks+ is prudent, review is slow
mircea_popescu: or that.
asciilifeform: apparently, when reading slow or variable-speed source ( such as FUCKGOATS ) dd will sometimes fail to fill an entire block (default block size is 512 on most boxes) and PAD WITH MOTHERFUCKING ZEROS
asciilifeform: the cure, apparently, is option ' iflag=fullblock ' which for some reason is NOT DEFAULT
asciilifeform: MOTHERFUCKERS
asciilifeform: on what planet was this considered acceptable?!!!
mod6: i'll try that next time i run a test.
asciilifeform: 'oh i'ma pad with zeros'
asciilifeform: holy FUCK the sheer gangrene.
asciilifeform: ^ mircea_popescu, mod6 , ben_vulpes , everybody ^
Framedragger: asciilifeform: whaaaaaaaaat
asciilifeform: ikr?
ben_vulpes: p. bad.
Framedragger: why in the fuck would that be useful
Framedragger: it's /dev/urandom but worse
Framedragger: also NOT ITS JOB
Framedragger: :D
mircea_popescu: asciilifeform i thought you thought core utils is fine.
asciilifeform: mircea_popescu: apparently everything is 'fine' until you find the magic corner case.
mircea_popescu: except for any time you try to use them, i find they are fine too! like one of those comedic toolboxes with a hammer that's seemingly attached except if you pick it up the metal falls off etc.
mircea_popescu: magic my foot.
Framedragger: mp is now like "i told ya so!!! tail!!!"
asciilifeform: http://nosuchlabs.com/hardware.html has been updated.
asciilifeform: mircea_popescu: e.g., dd, worx 100% of the time when copying multi-TB ssd, say.
asciilifeform: but apparently doesn't like kB/s variable thing.
asciilifeform: 'unix philosophy' my arse.
Framedragger: it's like the modern website. "i expect good bandwidth. you have low bandwidth? fuck yourself"
Framedragger: that's so *silent* and sneaky tho.
asciilifeform: Framedragger: almost like somebody thought about rng.
mod6: ok blowing away all of my .dat files in ~/.bitcoin
mod6: will recut all .dat files ...
Framedragger: "Note if the input may return short reads as could be the case when reading from a pipe for example, ‘iflag=fullblock’ will ensure that ‘count=’ corresponds to complete input blocks rather than the traditional POSIX specified behavior of counting input read operations." OK
asciilifeform: Framedragger: picture if 'this util will format your hdd unless you supply the iflag=dontformatmyhdd option' buried somewhere deep in docs.
mod6: cutting blocks...
asciilifeform: why the FUCK would anyone specify 'count ATTEMPTS' rather than reads !
Framedragger: apparently it's posix tradition didntyaknow
asciilifeform: barfalicious.
Framedragger: flabbergasted. may as well embrace tmsr's default malice interpretation here. because seriously. ☟︎
Framedragger thought `dd` was one of those few 'no hidden bullshit, what you see is what you get' utilities from better times
asciilifeform: aaaha.
Framedragger: nice find.
ben_vulpes: while we're doing these, i'd be much obliged if folks with a copy of drakma on hand would run `(drakma:http-request "https://untrusted-root.badssl.com")` and let me know what you get
asciilifeform: nooseflash: even, e.g., 'cat', has megatonne of liquishit.
asciilifeform: on 'modern unix'.
trinque: cat thread is by now a yearly tradition
ben_vulpes: cat: https://gist.github.com/pete/665971
asciilifeform: ^ we had this on 2+ threads
asciilifeform: ( spoiler : plan9's is the 'correct' one )
Framedragger: yes well, granted it doesn't support additional command line args, so has to handle fewer things. but then, maybe that, too, is *also* correct...
mircea_popescu: asciilifeform sadly, it IS the posix standard.
ben_vulpes: Framedragger: complexity is a the self-justifying disease of the programmer's mind
Framedragger: i certainly see that, especially looking at gnu cat (omfg)
mircea_popescu: http://btcbase.org/log/2017-04-18#1645740 << i can't iamgine how. the skeleton we found is just about 50 years old. ☝︎
a111: Logged on 2017-04-18 20:09 Framedragger: flabbergasted. may as well embrace tmsr's default malice interpretation here. because seriously.
ben_vulpes: but you don't see that from your career of programmering to date?
asciilifeform: mircea_popescu: ' i only implemented the standard ' has roughly same cachet as ' i was only following orders!11'
mircea_popescu: asciilifeform i betcha it's been in the old dog eared yellowed looseleaf notes the posix copied.
asciilifeform: mircea_popescu: nope
asciilifeform: http://pubs.opengroup.org/onlinepubs/9699919799/utilities/dd.html
Framedragger: ben_vulpes: of course i do. i meant, i see it here, too, and agree. but, yeah, 'tis a big problem
asciilifeform: 2001.
ben_vulpes: hmk
asciilifeform: the 'standard' in question is actually quasi-mythological work, and (yes) SCO's.
asciilifeform: THAT very same SCO.
mircea_popescu: hm.
asciilifeform: btw, here's a lulzy:
asciilifeform: yes | dd of=out bs=1024k count=10 ☟︎
asciilifeform: then ls -l out
asciilifeform: and laugh/cry
doppler: weird.. what's going on there?
asciilifeform: doppler: once you figure out this puzzler, you will grasp what the thread was about
shinohai: See the du -h retardation from previous thread
asciilifeform: ^ that one is at least plain to naked eye
asciilifeform: if you omit the 'h', you get output in martian units
doppler: I agree that -h should be the default in userland utilities
asciilifeform: bbbbutthatwouldntbePOSIX!111
trinque: folks that want a charitable interpretation of these might ponder a while re: who has the most time / resources to shitgnominate
doppler: asciilifeform: haha, yeah. like I have “alias l='ls -hlp'” for example
doppler: asciilifeform: POSIX sure seems to result in a lot of braindamage
jurov: short reads prolly started as unix worse-is-better philosophy "api has these huge warts but as it can be easily fixed by retrying in userspace, it's okay"..posix only snowballed on
mircea_popescu: i believe.
mircea_popescu: was a time when data was a lot more valuable, basically on account of the world not yet consisting of the www pressed shitboard. in that world of little valuable data, having any chunk "lost in the pipes" would have appeared typically soviet wasteful bureaucratism.
asciilifeform: except that the idiocy in question doesn't preserve any data.
asciilifeform: it inserts 0 silently in place of what failed to read.
asciilifeform: 'new jersey philosophy' : 'silent failure is just about as good as success'
mircea_popescu: if i have 3.5 blocks it allows me to read 4 blocks last of which is half 0s
asciilifeform: could just the same do this without the padding.
mircea_popescu: not on their iron.
asciilifeform: read 77 bytes ? write 77, motherfucker
asciilifeform: let the WRITER pad.
asciilifeform: if the iron wants it.
mircea_popescu: wasteful.
asciilifeform: how??
mircea_popescu: padding has now been written.
asciilifeform: i dun get it
asciilifeform: put the padding in when and IF required -- on the end that requires it.
mircea_popescu: do you want to get it ?
asciilifeform: rather than having the reader lie to the operator.
jurov: on the yes test, i got warning:
jurov: dd: warning: partial read (73728 bytes); suggest iflag=fullblock ☟︎
jurov: still, no mention at all data is mutilated
doppler: I didn't get any warnings. you're using coreutils dd, correct?
jurov: yes, sys-apps/coreutils-8.25 (/bin/dd) (acl nls xattr -caps -gmp -hostname -kill -multicall -selinux -static -vanilla USERLAND="-BSD")
doppler: oops, I actually do get that warning. totally didn't see it the first time. coreutils-8.27 here
mod6: <+asciilifeform> mod6: can haz one moar plot -- just 300,000 and up, plox ? << i can do this easy enough, still want this?
mod6: <+asciilifeform> yes | dd of=out bs=1024k count=10 << ya, tried this, 122880 bytes out
mod6: then tried `yes | dd iflag=fullblock of=out bs=1024k count=10` and says 10485760 bytes out
mod6: so yah that's fucky
mircea_popescu: meanwhile at teh farm, http://68.media.tumblr.com/a0620361837d5fee98c28c530b5615d3/tumblr_on2gp6kV7M1w5g497o2_1280.png
mod6: work, work, work.
mircea_popescu: better than the tanning bed.
mod6: for sure
asciilifeform: http://btcbase.org/log/2017-04-18#1645802 << i got no warnings ☝︎
a111: Logged on 2017-04-18 21:20 jurov: dd: warning: partial read (73728 bytes); suggest iflag=fullblock
ben_vulpes: dd on bsd (os x, claims 1994 vintage) does not appear to pad
ben_vulpes: manpage indicates padding must be asked for
Framedragger: this is only done by gnu coreutils dd, it seems. same warning on ubuntu (coreutils), etc.
ben_vulpes: http://p.bvulpes.com/pastes/1rlkx/?raw=true
BingoBoingo: !~ud tiny dancer
jhvh1: BingoBoingo: tiny dancer :: "Tiny Dancer" is the name given to (and used when spoken about in public or otherwise) the 1/2 erection that can be grabbed at the base and twirled around, thus becoming a "tiny dancer." [ex:] "So who wants to go play frisbee?""I do, but wait until my Tiny Dancer goes away." [/ex] | 1.) A drink made with one part Stolichnaya Oranj, one part cranberry juice. Garnish with lime.2.) A (6 more messages)
BingoBoingo: ^ At least ud core util works