asciilifeform: 'Tarrio believes that unless President Donald Trump steps in, the de-platforming and dehumanizing of conservatives will continue. “He needs to step in, not only because if he doesn’t he will lose in 2020 with all of his supporters being kicked off social media, but because it’s the right thing to do,” Tarrio finishes.' << lol!!☟︎
asciilifeform: historically i found that o3 is moar or less unusable for benchmarking anyffin, it risks to 'optimize away' unexpected pieces ( for instance, quite possibly your loop-and-mod-2 )☟︎
asciilifeform: mircea_popescu: this is why i like 'run 9000 shots, and discard outliers' (granted this only applies when the item is expected to go in constant time, goes without saying)
asciilifeform: ( depending on what you got in there )
asciilifeform: if yer proggy takes >=12h, this is certain to happen, it'll run up against cron
asciilifeform doesn't like benchmarks where 1 shot takes moar than coupla min, for the reason that unix box is a shitfest, liable to decide 'hey nao is the time to munge the fs' etc. and suddenly eat cpu for multiple min, whenever
asciilifeform: mircea_popescu: possibly was 'the peak of insanity'
asciilifeform: ( why -- cannot say until diana_coman returns with brief description of the irons )
asciilifeform: i'ma guess box ground to a halt during the 5-loops run, for a spell
asciilifeform: that's quite a few, even if you had nuffin in the middle of it
asciilifeform: what's hard to compute tho. you put sumthing on www, erry 2bit spamplagiarismbot mirrors it in 10sec, not even speaking of actual humans laffing
asciilifeform: i had nfi he was a proud , eh, son of persia
asciilifeform: (this was prior to mold going public re his 'kompyooting' )
asciilifeform remembers the roosh-v d00d, but for something quite else, iirc it was him who at one pt had asciilifeform's www linked as -- insultingly but lulzily -- 'the mencius moldbug of computing'
asciilifeform: mircea_popescu: negation << >> 'нет ни хуя' etc
asciilifeform: ( so far impression is, roughly similar to '80s su, tho with ~less~ obv. leakage of suppressed laffter through the pg , oddly , i expected moar )
asciilifeform: it's funny , when asciilifeform went to 'museum of communism' in timis, pawed through buncha ancient lulzlit there, and photo'd 'a la minox' to read through laters and chew on the 'limba de lemn'☟︎
asciilifeform: asciilifeform often points out that we have buffer overflows on acct of the iron having nfi what 'integer', 'character', 'string', 'code' are , just pushes untagged words around; but it is also the reason why we have e.g. thread deadlocks (on an iron which 'knows' when yer e.g. assigning a variable, deadlock is ~impossible)
asciilifeform: imho it is a curse on programming folx, that spackling seems to work as well as it does, really it is the equiv of eating 'tasty' pb paint chips, leads to grief always
asciilifeform: mircea_popescu: thinking about it -- 'zxc' strikes me as a classic case of http://btcbase.org/log/2018-10-25#1866191 and impedence mismatch generally. it was clearly written as attempt to 'deterministic scheduler on ??? iron/os', but fails, cuz you can't actually spackle away impedence mismatch b/w the underlying platform and the proggy☝︎☟︎
asciilifeform: mircea_popescu: i don't specifically know, but suspect that the eggog happens when parent thread dies and (for moment) the children keep fandangoing
asciilifeform: imho zcx isn't actually of use on the current gnat, at all
asciilifeform: mircea_popescu: if (and only if) yer scheduler actually runs on the naked iron, a 'polled' sys gives ~less~ nondeterminism than a fully asynchronous one. but nobody's got this iron of yet.
asciilifeform: a linear process can still be heavy enuff (e.g. a 8192-bit primality test on ffa) that you wouldn't want to wait for it to finish when killing a thread
asciilifeform: if you have cppola ( or even, e.g., an asm-enhanced ffa ) in yer loop, it won't be poll-killable.
asciilifeform: errr, where yer proggy (rather than gnat) is on naked irons
asciilifeform: pollism may be suitable for an embedded sys, where gnat is running on naked irons, and the entire contents of the machine was built in 1 shot in the same gnat. but errywhere else it imho is liable to break.
asciilifeform: bvt: rright. my observation is, it aint suitable for a general-purpose kill switch , the prob orig. posed by mircea_popescu , cuz polled mode can only give ability to kill items which 'agreed' to be killable
asciilifeform: where you had to explicitly yield a thread, or whole machine ground to halt
asciilifeform: 'polled' mode seems to resemble the 'cooperative multitasking' found on ancient crapple and win9x
asciilifeform: bvt: my reasoning is based on the docs description of how implemented. but certainly worth a test, if anyone can be arsed ( given the 'all loops gotta be doctored' thing, my understanding is that 'polled' proggy would choke if wedges inside a cppism )
asciilifeform: bvt: whole point of multithreadism is to spread out the cpu-bulky work over cores, so i expect most cycles would be spend inside where enabled
asciilifeform: not to mention, the s.mg folx invoke external cppisms , and these won't have poll markers inside'em no matter what you do
asciilifeform: currently i suspect that polling (even if can be made to work errywhere) will be 9000x moar expensive than longjmpism ( the latter only adds to cost of exceptions, the former -- to ~all~ loops )
asciilifeform: http://btcbase.org/log/2019-02-12#1895192 << before this gets lost -- irix has reputation as stable, but porous as all fuck ( sgi never apparently gave half a shit ) ; not that we're at risk of cribbing from it, as you pointed out, 0 src☝︎
asciilifeform: ( 1 moar reason why i'd like that msdos cross-gnat.. on msdos, the wall clock time is invariant )
asciilifeform: it's perfectly good, but to get decent resolution you end up needing gigantic # of repeats ( and unixism machine churn begins to drown the signal )
asciilifeform: ( so as to get actual ticktimes for this & other empiricals )
asciilifeform: i really gotta genesis my rdtsc asmism thingie for gnat
asciilifeform: ( would work so long as you take care to return. but i think that is obv )
