log☇︎
160200+ entries in 1.227s
decimation: asciilifeform: yeah I think in this case it was a z80 clone
decimation: in it, he says that he feels that code that is compiled and assembled from a higher level language (higher than assembly) isn't 'his'
decimation: asciilifeform: http://www.theamphour.com/247-an-interview-with-voja-antonic-gerontogenous-galaksija-genesis/ < that podcast is an interview with voja antonic, who invented an 8-bit 'home computer' that was published in a magazine in yugoslavia
decimation: yes, point taken. but what do you call memory that might be a leak, or might not?
asciilifeform: a great many ciphertrons sold by nato cocksucker nations (western europe) were actually electronicized versions of 'enigma'-style rotors ☟︎
asciilifeform: but more broadly against crypto research outside of the castle walls, as a class
asciilifeform: incidentally, it would be a mistake to conclude that nsa was specifically raging against rsa and only it
asciilifeform: (btw it was a very spiffy encyclopaedia and unjustly gets shat on today)
assbot: Logged on 29-05-2015 03:35:55; mircea_popescu: by now i heard the "nsa didn't want strong encryption at all, and lobbied against inclusion of anything like it were microsoft on a bender" from so many people it's settled
asciilifeform: decimation: 'leak' has a precise technical definition
asciilifeform: in that the total footprint does not exceed a certain size
asciilifeform: any that monotonically increases is candidate for a leak. but mod6's test (and mine) suggests that there are no longer leaks in the classical sense
mircea_popescu: by now i heard the "nsa didn't want strong encryption at all, and lobbied against inclusion of anything like it were microsoft on a bender" from so many people it's settled ☟︎
asciilifeform: mircea_popescu: nah this is an entirely other animal, 'igprof.' takes malloc byte counts from a living (!) process
mod6: :] i read a doc from the igprof site that had me confused on what I was looking at.
asciilifeform: there is a hard-ram-bound option, presently unused, and entirely separate from the locks constant. see link in log.
asciilifeform: according to docs, this is not a hard necessity.
asciilifeform: mircea_popescu: i get a ballooning all the way to ~23M during first few hours of runtime, where it stays (based on pmap data)
mircea_popescu: cazalla i think she has a point actually. deed the pages, that's proof enough. and as the derps probably don't know (or don't know they should know), we might even catch a diddled page. maybe.
mircea_popescu: mthreat interesting. so basically they run them looser and with a momentum-consuming groove in one spot.
decimation: asciilifeform: it appears a big chunk of the memory in use is tied up in the CBlockIndex class
asciilifeform: there is absolutely no reason for bitcoind to ever use more than at most a few MB more than when it first boots.
asciilifeform: yes, it's a typo
asciilifeform: then again, it is running on a very generously oversized box
mod6: <+ascii_modem> nonono not leaks << ok i see, so with -mp in the MEM_LIVE output we're just seeing a live heap. not leaks as the other doc page said.
decimation: "MEM_LIVE records the ?live? memory ? memory that hasn?t been freed. If the profile statistics file you are processing came from the end of the application?s run, this will be the memory leaked by the job. If the profile statistics file was triggered during the running of the job, it is a snapshot of the heap, i.e. a heap profile. The statistic is accurate (not statistical) and records the number of bytes allocated and the number of c
mod6: i was just getting started reading and got pulled away for a minute.
decimation: presumably the 'leaks' were found by igprof because he killed it after running a certain time, so memory wasn't freed?
mod6: decimation: ah, no. graphs are from nmon. there is a document that goes along with all of that. it's in the mailing list and in the http://thebitcoin.foundation/test/perf/ root
decimation: mod6: so you did a graph of vmstat calls? Is that what ascii is talking about?
decimation: asciilifeform: all roads lead to home, yes, but home is a good place to start examining each potential pathway
decimation: by my eye, it looks to me that ThreadMessageHandler(void*) is a turd
mod6: asciilifeform: in the igprof_mem_live output, is the "Total" column in aggregate bytes - a sum of all bytes from the call's of said function?
decimation: http://www.reed.com/blog-dpr/?page_id=6 < "One project where my friend and officemate Steven T. Kent (now chief scientist and vice president at BBN, and a chief advisor to NSA) and I lost was our strong argument to put mandatory end-to-end encryption into TCP (and adaptations of the ideas to UDP-based protocols, such as RTP, hich I worked out but abandoned). "
assbot: 10 results for 'referredbyloper' : http://s.b-a.link/?q=referredbyloper
mod6: What's the memory cap that you wanna use on a pogo? 256Mb?
assbot: Logged on 28-05-2015 14:52:47; pete_dushenski: "Broadcom is known as a fabless company. It outsources all semiconductor manufacturing to Asian merchant foundries, such as GlobalFoundries, Semiconductor Manufacturing International Corporation, Silterra, TSMC, and United Microelectronics Corporation." << ahaha. so what the fuck did $37 bn even buy ?
decimation: http://log.bitcoin-assets.com/?date=28-05-2015#1146983 < primarily it bought a whole pile of sunk costs in existing designs, plus serfs who know how to manipulate the tools ☝︎
ben_vulpes: asciilifeform: is there a russian equivalent to cursive?
ben_vulpes: i refuse to take notes on a computer
asciilifeform: this -never- happened at any point when i actually picked up a pen on a daily basis (school)
asciilifeform: i even found myself, as of 4-5 yrs ago, often writing the 2nd letter of a word first
asciilifeform: 'MEM_LIVE records the “live” memory – memory that hasn’t been freed. If the profile statistics file you are processing came from the end of the application’s run, this will be the memory leaked by the job. If the profile statistics file was triggered during the running of the job, it is a snapshot of the heap, i.e. a heap profile. The statistic is accurate (not statistical) and records the number of bytes allocated a
mod6: ok i'll come back to that, it's like a whole process to build that thing.
asciilifeform: btw, after dumping a heap profile snapshot,
mod6: oh, just looking for a version of IgProf to grab... maybe im looking at the wrong thing.
ben_vulpes: anyways, that's a nifty muntzing in the first one
ben_vulpes dreams of a box beefy enough to run sync tests for every patch submitted
mod6: well, was going to launch this for a full sync test.
asciilifeform: mod6: you can use this to create a blockheight vs heap plot
ben_vulpes: "*** Update 7/18/06 *** You have to give the California Department of Motor Vehicles (the DMV) credit for creativity on this one. A DMV insider has disclosed to me that the DMV has made a formal request to a federal agency to rule if my Beetle constitutes a threat to national security based on what could happen if it got into the wrong hands."" << hah news to me since the last time i looked at the
ben_vulpes: <asciilifeform> ben_vulpes: sorta like the proverbial 'jet engine on a ford pinto' situation. << obligatory: http://www.ronpatrickstuff.com/
hanbot: anyway, archive.today the deedbot page consisting of a dump of the page source of an archive.today page? haha!
hanbot: ahaha. mp was calling them scammers a month ago, everyone was like "orly, mp is so crazu". turns out...
cazalla: i'm sure there is a better way to do it but relying on archive.today probably isn't that great an idea but not sure what else to do (for example, all the bitcoin auction stuff has been completely removed from the US Marshall's website) ☟︎
cazalla: so it occurs to me that a #b-a version of archive.today might be useful, really nothing stopping whomever going back to edit archives whereas with our own, we could sign the archives we make
asciilifeform: imagine having a special door in your house through which perfect strangers are invited to dump in up to five thousand tonnes of gravel.
asciilifeform: ben_vulpes, mod6, mircea_popescu, et al: incidentally, even though 'transaction orphanage amputator' may not be a magical pill against oom-on-pogo (my preliminary investigation suggests that it is -not-) it still rips out the idiotic ddos vector that every unbounded cache of anything whatsoever is.
asciilifeform: ben_vulpes, mod6, mircea_popescu, et al: incidentally, now that we have a bounded memory footprint, it may be time to try running pogo with swap space again
asciilifeform: mike_c: slab only really works when we have a small number of known sizes of object
asciilifeform: ben_vulpes: sorta like the proverbial 'jet engine on a ford pinto' situation.
asciilifeform: ben_vulpes: more seriously, the only kind of allocator that is a sure pill against fragging is a compacting one. which means -relocating- !! just picture this in a cpp proggy not originally designed for it
mod6: <+asciilifeform> ben_vulpes, mod6, mircea_popescu, et al: http://therealbitcoin.org/ml/btc-dev/2015-May/000094.html << wow! nice. will review & start up a test on this later tonight. many thanks. :]
assbot: 0 results for 'hoarders buried alive' : http://s.b-a.link/?q=hoarders+buried+alive
asciilifeform: or hiring a housekeeper
ben_vulpes: intentionally filling a 0.5.3.x with orphans?
asciilifeform: 1) and thus, valgrind will not see this as a leak, because the memory is theoretically reclaimable
asciilifeform: our 0.5.3.x has MAX_ORPHAN_TRANSACTIONS but the boojum is that a malicious (or 'cosmic rayed') node can transmit orphan transaction of any size, iirc
williamdunne: Mm should be in Bucharest in a couple of weeks..
mircea_popescu: documenting things is always a good idea.
asciilifeform: mircea_popescu et al: is anyone interested in seeing the (elementary) 4294967297 proof as a proper article ?
assbot: Logged on 28-05-2015 16:29:59; ben_vulpes: (it's a boy)
mircea_popescu: http://log.bitcoin-assets.com/?date=28-05-2015#1146972 << most of the old jewish haunt, from berlin to moscow, is a huge pile of old books looking for new onwners ☝︎
assbot: Logged on 28-05-2015 13:45:46; asciilifeform: http://log.bitcoin-assets.com/?date=28-05-2015#1146846 << my first impression was that the instrument is a very long... spoon
mircea_popescu: http://log.bitcoin-assets.com/?date=28-05-2015#1146904 << it's a crop. ☝︎
ben_vulpes: (it's a boy) ☟︎
mod6: I should be able to get a box for this sometime after the 1st.
mod6: <+gabriel_laddel> mod6: Idk what exactly your goals are on your gentoo quest, but have you reached the conclusion that portage is insane and must go? << this project had to take a pause; I need to get some cash together to buy a physical box to put together steps for configuration on real hardware as opposed to aws or other cloud infrastructure. no, for now, I'm trying to stick with a Gentoo-amd64-uclibc-hardened. That's the goal anyway.
pete_dushenski: "Broadcom is known as a fabless company. It outsources all semiconductor manufacturing to Asian merchant foundries, such as GlobalFoundries, Semiconductor Manufacturing International Corporation, Silterra, TSMC, and United Microelectronics Corporation." << ahaha. so what the fuck did $37 bn even buy ? ☟︎
assbot: 0 results for 'avago' : http://s.b-a.link/?q=avago
pete_dushenski: "Avago, a maker of chips for the wireless and industrial markets, is offering Broadcom shareholders $17 billion in cash and Avago shares valued at $20 billion in one of the biggest deals ever in the chip industry." << i have to admit that i'd never heard of 'avago' before.
pete_dushenski: there are a handful of such stores in pretty much every town here !
pete_dushenski: i saw one such bookstore in manhattan a couple years back, complete with $30,000 copy of le petite prince, but that was just one shop.
pete_dushenski: "Bruno and his team are designing video games, including a prototype puzzle for iPads and smartphones, and looking at virtual reality simulations to help people visualise what are often complex investment portfolios."
pete_dushenski: ""How do you get under the skin of clients today, because they often work on their mobiles and they manage their wealth in their spare time," said Dave Bruno, head of UBS's innovation lab. "It might be in the bathroom, it might be waiting for a flight.""
pete_dushenski: "In a fifth floor office just off Zurich's main shopping street, researchers at UBS (UBSG.VX) are testing dozens of technologies to see what could make the world's biggest wealth manager more appealing as fortunes pass to the next generation."
gabriel_laddel is going for a walk
gabriel_laddel: pete_dushenski: Nope. Consider this a request for a blog post on your travels in these two regions.
pete_dushenski: mhm. i've been to pusan (where i have a cousin), beijing, shanghai, and hong kong.
gabriel_laddel: https://www.linkedin.com/in/richling << Impressive resume. Mitch also has a collection of CL mathematics hacks, mjrcalc.
pete_dushenski: i'd leave at the drop of a hat if push came to shove, but it hasn't yet.
pete_dushenski: it's a hobby more than anything.
pete_dushenski: i just travel for a few weeks here and a few more weeks there.
pete_dushenski: the 6-9 hour train i took across spain and portugal a few years ago most certainly did.
pete_dushenski: a 45 min or even 2 hour trip doesn't eat your day up
pete_dushenski: lol the scenery out the window is prettier. also, tracks are a hell of a lot smoother than shitty alberta roads.
asciilifeform: gabriel_laddel: cars (at least of the petrol variety) underground is a ventilation nightmare. use trains, like sane people
pete_dushenski: good god man get yourself a blog already.
gabriel_laddel: pete_dushenski: nah, it was a message to someone on #lisp
pete_dushenski: this is like trying to build a concrete foundation under a mountain