700+ entries in 0.349s
adlai will have to implement his own, so it runs on his '
pogo'
☟︎ punkman: does
pogo have problems with big mmaped files?
mircea_popescu: even "
pogo-ready btc" might be good enough. definitely WOULD have been good enough, in 2015.
ascii_field: (and it HAS to be cured before
pogo can be a thing)
phf: jurov: the goal is to get bitcoind working with
pogo's limited memory. the problem is that a running bitcoind grows in memory use as a result of normal operations. we know that some processes claim a lot of memory by design, like mempool, so first step is to get a reliable way of cleaning out mempool. ascii wrote that patch, but discovered that in practice zapmempool doesn't reduce memory use.
shinohai looks longingly at his little
pogo on the shelf above.
shinohai: jurov seems to work here, I cat'd a few over to
pogo and it syncs against it ^^
jurov: btw, is it possible to continue running a blockchain synced on amd64 on
pogo?
mircea_popescu: Thing 6. We're trying to create a sane computer (other than
pogo). Part of this consideration is some in-principle fundamental research work into ternary to be maybe attached to Thing 3.
mircea_popescu: Thing 2. We're trying to create an actually statically linked, weaponized bitcoind that's suitable for the
pogo.
trinque: asciilifeform: thinking the kernel's automount mech. for
pogo's external disk slot?
mircea_popescu: kinda have a) a nix ; b) a bitcoin and c) an obvious
pogo-merge for these two.
mircea_popescu: actually the rotating tape is not such a bad idea. not just for
pogo-log, but in general. wild notion, but, asciilifeform how about memory allocator that simply overwrites the beginning ?
assbot: Logged on 05-11-2015 02:54:26; asciilifeform: on a machine like
pogo it does not make sense to produce the log at all
phf: in that case you'd better truncate it doing run too, because if that
pogo doesn't go down, there's no cap on the log
phf: should there even be logs on
pogo?
mircea_popescu: the best thing you can do for it is to get the
pogo ready to go.
shinohai steals ascii_field 's build notes for
pogo...
ascii_field: busybox is typically used in small machines, e.g., routers, our '
pogo', etc.
assbot: Logged on 01-11-2015 18:16:41; mod6: hmm. so yeah, the key will be to get these deps compiled statically. i might be wrong, but alf's
pogo related things might have some sample configurations for getting these statically built on ARM.
mod6: hmm. so yeah, the key will be to get these deps compiled statically. i might be wrong, but alf's
pogo related things might have some sample configurations for getting these statically built on ARM.
☟︎ ascii_field: weird chip (similar to
pogo) for which the only available os was a systemdized linuxlike
adlai: somebody get mathias a
pogo ascii_field:
pogo is something quite like an actual computer
BingoBoingo: <mircea_popescu> Power Supply 10 x 1600 W Form Factor 11 U rack << reasonably even out power savings from
pogo BingoBoingo:
pogo may end up better suited to simpler #b-a services that emerge than serving bitcoin proper
ascii_field: BingoBoingo: we don't have anything remotely close to a
pogo-capable bitcoind.
BingoBoingo starting to wonder if
pogo could ever run bitcoind and a form of p2pool at the same time
phf: ascii_field: you're too sane, to comprehend this madness. the way stupid thinks in this case, "want to sell bitcoin device, can't really make
pogo, because no "value add", so must add some hardware to make it seem like useful hardware." of course they don't have enough sense to say add a ecc sig verifier asic, so add probably stolen miner asic code, which is not going to mine anything for anybody, hence the dodgy justification code.
shinohai: So what does that thing do that a
pogo can't ?
BingoBoingo: ascii_field: This is a newer stinkier shit
pogo assbot: Logged on 11-08-2015 13:36:02; asciilifeform: shinohai: see, i'd post a full recipe and signed binaries today, but we don't quite yet have a
pogo-capable - that is, non-ramguzzling - bitcoind
shinohai: Nope they didn't get in the library, just the little room where my routers and spare computers are. they left my
pogo lol.
shinohai: K jurov. mats sent me sufficient btc to purchase a
pogo, but I only got an inferior drive with it. Planned on getting something a little more robust
shinohai: Thx jurov. On my list of items to purchase when I actually get a working
pogo binary.
ascii_field: these measurements can be deceptive re:
pogo in particular, on account of the latter's sensitivity to cpu and especially bus starvation
jurov: okay but i wanted something cheap and still performant enough to use with
pogo mircea_popescu: but in any case : it's not clear we can't stick it in
pogo. so... there is that.
BingoBoingo: <asciilifeform> signal with farts. << If I understand Edgerouter and
Pogo are "farts" that signal ingroup. GIMP is a shit that doesn't work by spz to signal ingroup?
jurov:
pogo is useless without hard drive, either (and no, we don't have antifuse rom)
assbot: Logged on 17-08-2015 20:33:50; mircea_popescu: incidentally re
pogo and cpu - i am well persuaded the problem is in how the software checks sig, not in the cpu limit.
mircea_popescu: incidentally re
pogo and cpu - i am well persuaded the problem is in how the software checks sig, not in the cpu limit.
☟︎ mircea_popescu: the whole
pogo thing was a few k's. this thing is k's of btc
ascii_field: if it becomes possible to have blocks that selectively disable, e.g.,
pogo, enemy has a useful new capability.
mircea_popescu: ascii_field looky : blocks would only stay full if needed full. so our
pogo is a piece of shit in that case, you can't expect them to accomodate us.
mircea_popescu: ascii_field let us calculate the actual case of the
pogo, shall we ? as an exercise, mostl;y for my own edification.
ascii_field: which means, he can spend $fixint and now ~no
pogo is guaranteed to work~
assbot: Logged on 11-08-2015 19:12:55; hanbot: shinohai: hanbot compiled
pogo ? O.o << nein. have built stator "manually" as of last night, tho still need to verify runs as expected, after which will put together notes. have not
pogo-ed, rotor trial to come. (ubuntu 10.04 32bit + mod6's gracious help)
hanbot: shinohai: hanbot compiled
pogo ? O.o << nein. have built stator "manually" as of last night, tho still need to verify runs as expected, after which will put together notes. have not
pogo-ed, rotor trial to come. (ubuntu 10.04 32bit + mod6's gracious help)
☟︎ shinohai retrieves his
pogo from the disposal and puts it back on the shelf.
shinohai: Compiling ON the
pogo is pointless.
mircea_popescu: shinohai maybe i don't understand crosscompilation, but a) shouldn't it work on
pogo if you ever got it to compile/link anywhere ? and b) wasn't the idea to avoid compiling for the
pogo ON THE
POGO anyway ?
shinohai: Shitty old AMD laptop: 1, Shiny
pogo: 0
trinque: even if you end up with a working
pogo, if you don't understand the steps which resulted in that, you've gained very little
shinohai: asciilifeform: with the
pogo binary from the mailing list I got 3 connections xD
shinohai: binary still doesn't run on mah
pogo so I did something wrong.
shinohai: At the very least, I have not cursed God one time in all my failures with this
pogo danielpbarron, so maybe there is hope for me yet.
shinohai: trinque: *for* the
pogo. *on* the
pogo is just dumb.