log☇︎
900+ entries in 0.078s
diana_coman goes to test with --RTS=sjlj, can't hurt
bvt: during the gnat build, the sjlj runtime is built, so it should be possible to switch to it and test. ☟︎
diana_coman: i.e. if anything, you can say that my test code should wait a bit and then check
diana_coman: mircea_popescu, the check is in the TEST code
diana_coman: bvt's investigation of my ada-won-t-stop test program: http://p.bvulpes.com/pastes/sRqWV/?raw=true ☟︎
mod6: The other weird thing is that 'usb hub' as shown in the kernel.debug file. I do own a USB hub, as I bought one to test multiple FGs at the same time, but it's been just sitting on my desk for weeks.
asciilifeform: will prolly take asciilifeform a coupla days to hoist it on the pedestal and test, tho, conveyor is currently packed
asciilifeform: diana_coman et al : http://p.bvulpes.com/pastes/4SpwP/?raw=true << built, replicated test jig; i'ma lay the thing out on vivisection table properly after sleep ( spent today , among other things, fixing a furnace, bent wrists out , gotta let'em snap back )
a111: Logged on 2019-02-10 20:31 diana_coman: asciilifeform, and anyone else interested in testing Ada's failure to abort, minimal test setup: ossasepia.com/available_resources/test_tasks_ada.zip
mod6: Well, if your frustrated with me, I apologize. I'm just trying to get this up and going so I can test Keccak TRB on here.
trinque: I'm certain the cuntoo script is *not done* which is why I'm having folks test it, but I want more out of these tests than "ow, it wasn't a debian installer"
diana_coman: asciilifeform, and anyone else interested in testing Ada's failure to abort, minimal test setup: ossasepia.com/available_resources/test_tasks_ada.zip ☟︎
asciilifeform: diana_coman: can you post plox the whole test jig, i'd like to reproduce the effect and see wtf , in gdb ( prolly won't get to it until nightfall tho )
asciilifeform: i dun recall whether trinque mentioned an x test tho.
mircea_popescu: trinque "it's a test of musl" in the sense that if something breaks it'll likely be that, not reasonable to expect that because you took out whatever solitaire version gets bundled with gentoobuntu then therefore your build crashes.
asciilifeform: really it's primarily a test of muslism
mircea_popescu: well, put something like httpfox on a browser or w/e you use for that, BingoBoingo can set you up a test acct on their shared environment, and you can capture the request it makes and use that in the future.
mircea_popescu: after all, the whole fucking point of "test this" is the "somebdoy other than me" outlook. otherwise what, "worked for me" ?
mod6: it has a name, 'trb-test1', a simple gentoo box to run trb test builds.
a111: Logged on 2019-02-09 07:54 feedbot: http://thewhet.net/2019/02/hanbots-cuntoo-bake-test-notes-part-i/ << The Whet -- hanbot's Cuntoo Bake Test Notes - Part I
a111: Logged on 2019-02-09 07:54 feedbot: http://thewhet.net/2019/02/hanbots-cuntoo-bake-test-notes-part-i/ << The Whet -- hanbot's Cuntoo Bake Test Notes - Part I
feedbot: http://thewhet.net/2019/02/hanbots-cuntoo-bake-test-notes-part-i/ << The Whet -- hanbot's Cuntoo Bake Test Notes - Part I ☟︎☟︎
a111: Logged on 2019-02-08 00:10 mircea_popescu: "i was nitpicking re strength, not the model per se". wth, it's not necessary the model airplane used to test out the worker hive's going to contain the sort of seating inside you can ~rely on~ to see in the finished item.
mircea_popescu: "i was nitpicking re strength, not the model per se". wth, it's not necessary the model airplane used to test out the worker hive's going to contain the sort of seating inside you can ~rely on~ to see in the finished item. ☟︎
asciilifeform: in other cuntooisms : if anyone is short a x64 box to test-fire cuntoo with, asciilifeform has a surplus disposable box, 'lenovo s10-3' , with that same chipset as in x60 etc period ( https://archive.is/Dny84 ) , if anyone in l1 wants, it's yours for the cost of postage ( has a mechanical hdd in it, i fughet of what size ) ☟︎
asciilifeform: this will be good test for 'can light up cuntoo from in a heathen gentoo'
BingoBoingo: In one hand I have a very enviable problem in a repeat customer wanting a box stood up. With that box the customer intends to have very capable hands under his control test the cuntoo build process. As a bonus they are offering to molest USB sticks into cuntoo propagators.
mircea_popescu: BingoBoingo but i mean, i have to buy a ssd to test out cuntoo ?
trinque: building it on cuntoo might be a good test of your script's isolation. possible it relies on something from the host.
asciilifeform: mircea_popescu: thing doesn't (as far as anyone knows) contain any writables on board. so test of new unit is limited to BingoBoingo's visual inspection and the ordinary smoke test.
asciilifeform: BingoBoingo: in http://qntra.net/2019/01/constant-time-miller-rabin-test-added-to-finite-field-arithmetic/ , plox s/informating/informing
mircea_popescu: koch is implementing fermat because fermat ~is~ the test, not because fermat ~does~ something.
a111: Logged on 2019-01-30 16:43 asciilifeform: ( as i understand, in koch world fermat test uses ~exactly same # of cycles as 1 shot of m-r )
asciilifeform: ( as i understand, in koch world fermat test uses ~exactly same # of cycles as 1 shot of m-r ) ☟︎
feedbot: http://qntra.net/2019/01/constant-time-miller-rabin-test-added-to-finite-field-arithmetic/ << Qntra -- Constant Time Miller-Rabin Test Added To Finite Field Arithmetic
feedbot: http://www.loper-os.org/?p=2978 << Loper OS -- Finite Field Arithmetic. Chapter 16A: The Miller-Rabin Test.
a111: Logged on 2019-01-15 22:09 mircea_popescu: but tbh, the "stochastic convergence test" seems to me worth A LOT more than any bs "ent" ultimately meaningless "squares of pi" bs.
asciilifeform: or, in better-known turdolade, 'The Go programming language (GoLang) 1.10.3 [Goo18] created at Google in 2009 is an open source project including arbitrary-precision arithmetic and cryptographic functionality... ...the pseudorandom number generator used in this primality test is seeded with the tested number n.'
asciilifeform: 'If either LibTomMath or TomsFastMath are selected, the pseudoprimes described in Section 4.9 (see Appendix I) will always be declared prime by the primality test.' << for the innocent : 'tommath' is 1 of those 'independent, not openssl' arithmetrons...
asciilifeform: will be interesting to test on '9000' koch-generated primes, and see if any... aint
shinohai: xpost for trb-ists: http://btcinfo.sdf.org/blog/trb-keccak-regrind-test-results-and-notes.html
asciilifeform: ^ ftr, the added cost of the piecewise-defined (correct for all possible inputs of N and Witness) , vs. yesterday's 'naked' variant, is <0.001s (on std test iron.)
mircea_popescu: asciilifeform it is counterintuitively actually better to test with smaller integers than larger ones.
asciilifeform: this type of test is impossible on systems where m-r eats witness straight from rng, without possibility to override by hand.
asciilifeform: per the proof, it is seemingly harmless ( a carmichael number has 1/4 of the integers as 'liars' ) but what it does is to prevent simple manual test with small numbers , which is imho quite typically kochian
asciilifeform: depends on 'benefits' neh. you can express any 'divisibility test' as a gcd with fixed param.
a111: Logged on 2019-01-22 04:07 asciilifeform: briefly upstack to http://btcbase.org/log/2019-01-22#1889021 << possibly i'm thick, but it ~also~ never made sense to me why a ~router~ would fall down, either. seems like if yer pipe is e.g. 100mb/s , and incoming enemy crapola at 1000mb/s, then you simply oughta get (from pov of arbitrary test peer) 90% packet loss. rather than a smoking crater where router was.
asciilifeform: briefly upstack to http://btcbase.org/log/2019-01-22#1889021 << possibly i'm thick, but it ~also~ never made sense to me why a ~router~ would fall down, either. seems like if yer pipe is e.g. 100mb/s , and incoming enemy crapola at 1000mb/s, then you simply oughta get (from pov of arbitrary test peer) 90% packet loss. rather than a smoking crater where router was. ☝︎☟︎
asciilifeform: i simply can't bring myself to buy any to test, i dun have any pigs that need autoclassifying
asciilifeform: ( and before anyone loses appetite, this experiment was on a test box in torture room, not in pizarro )
a111: Logged on 2018-08-08 16:58 asciilifeform: BingoBoingo: i'ma test new kernel today, then tomorrow will ask BingoBoingo to power down rk-c and put disk in dulap to i can snap new master img, then it will go into service ( and can reimage BingoBoingo also ) .
asciilifeform: a test with independent (from other 5) power supply, also returned same.
asciilifeform: notably, a test using the auxiliary (usb2.0) jack returned same result.
a111: Logged on 2019-01-15 22:09 mircea_popescu: but tbh, the "stochastic convergence test" seems to me worth A LOT more than any bs "ent" ultimately meaningless "squares of pi" bs.
mircea_popescu: but tbh, the "stochastic convergence test" seems to me worth A LOT more than any bs "ent" ultimately meaningless "squares of pi" bs. ☟︎☟︎
asciilifeform: BingoBoingo et al : re uy1: if it continues humming along, as it presently appears to be, we gotta proclaim a maintenance window and properly stress test the thing, load cpu to max for coupla hrs
asciilifeform: in other olds, 'The 256-bit integer q = (2x + 1)(4x + 1) with x = 0x400286bac15132db85b1c936709f369b passes 15 rounds of GMP’s primality test mpz_is_probab_prime_p...' ( Albrecht et al, aug. 2018 , https://eprint.iacr.org/2018/749.pdf / http://p.bvulpes.com/pastes/3QLGp/?raw=true )
asciilifeform: ( for folx who haven't yet read diana_coman's series -- the theorem diana_coman turned up, turns m-r into a deterministic ('aks but fast') test, iff generalized-riemann )
asciilifeform: achtung pizarro subscribers : if 1 of you is running saturation test on the pipe, plz nao & in future ~announce prior~, ty.
asciilifeform: very handy test tho
mircea_popescu: but then agai nthe space of shit is SO VAST it's not feasible to exhaustively test.
mod6: While Republicans are reviewing the above, I'm going to continue working on my TRB HOWTO updates -- I think it's pretty much finished, but I want to test it first before I start handing it around for testing by all. After this, and blessings of the regrind from TMSR~, I'll put a date on when thebitcoin.foundation site will roll over to the keccak vtree exclusively.
mod6: http://blog.mod6.net/?p=26 << Here's my TRB Keccak Regrind blog post. I have a tarball linked very near the bottom that contains my reground trb keccak vtree, as well as seals for each. These are posted so one may test with these (I need the help!).
asciilifeform: observe that by this scheme, we also avoid hardcoding primorials for 'G' test. author of tape is responsible for including a primorial ~for his chosen ffawidth~ if he intends to use G litmusing.
diana_coman goes to read some more and test even more
asciilifeform: ( it dun use any osisms other than the cmdline args, which can be hardcoded for test runs; as for i/o, can be rs232, and guesswat, 'os' )
asciilifeform: aaand to round off : it vanished on the test box also. culprit appears to have been a running raid-verify job... ☟︎☟︎
asciilifeform: (the test inputs are ~100MB ea.)
asciilifeform: and the orig. anomaly disappears on irons other than the test box
asciilifeform: btw the thing that set asciilifeform's ears standing originally, is that the gcd test battery appeared to give +/- 30% variant runtimes depending on hamming weights. but in 20 repeats of the trial set, the same thing showed up, and with no correlation to hamming weight. and a http://p.bvulpes.com/pastes/wherN/?raw=true ( soft barrelshifter ) version now ends up showing... same thing. i suspect that some peculiarity of the pipeline is
a111: Logged on 2019-01-10 01:52 asciilifeform: hm , prolly oughta test both left and right, and make the operand nonzero (in case the iron does 'clever' with zeros), so gotta redo..
asciilifeform: hm , prolly oughta test both left and right, and make the operand nonzero (in case the iron does 'clever' with zeros), so gotta redo.. ☟︎
mod6: *nod* I appreciate that asciilifeform, your eyes on such things always are a Good Thing. Not long after I post it, I'll be looking for people to help test the updated HOWTO guide too if anyone feels so inclined. (Please reach out if interested.)
mircea_popescu: asciilifeform IF rm doesn't take much time, THEN i don't see the point of doing any gcd. because r-m is a complete test for what we need, and gcd is not.
asciilifeform: well at least not in the initial sieve ( one might still want it for e.g. pocklington's test )
asciilifeform: if primality test ( which consists of GCD ~and~ m-r, in order to constant-time ) does not exceed 0.0356sec, then on machine with 1 FG it can be considered that the FG is the limiting reactant.
mircea_popescu: or, to quote from fetgirl, "mistey25 19F Switch Also an intelegence test? What makes you think you have what it takes to judge if someone is smart or not. There’s a quote from Albert Einstein “if you judge a fish by its ability to climb a tree it will live its whole life believing it’s an idiot”. And you’re judging me based off of things you’re pulling out of your ass. Why don’t I give you a bit of judgement. Yo
mircea_popescu: and then the conversation flows by itself, "suppose i test one of them artefacts and it matches turbine ?" "then you have ~made~ a turbine ~out of a~ religious artefact"
mircea_popescu: in the very strict terminology of the turing test : if you get a mystery box you can't explain whether it contains type 1 moron or type 2 moron.
mircea_popescu: the last time anyone tried this test was specifically http://trilema.com/2017/when-did-america-end/#selection-103.0-103.11
mircea_popescu: and if you do not believe, the test is spelled out in ye olde http://trilema.com/2014/the-problem-of-enforcement/ : start some shit about hot-topic-du-jour, see how "your boss" suddenly isn't what he seemed.
BingoBoingo: <asciilifeform> rright but douchebag passed, at least initially, turing test, and these 'people' generally do not << Lacking tits douchbag has to try to brand and personality. Having tits in the Femoocracy, girl just has to scrape and paste.
mircea_popescu: asciilifeform that's because your test is biased against women.
asciilifeform: rright but douchebag passed, at least initially, turing test, and these 'people' generally do not
asciilifeform: ( and if you montgomery, then you gotta either test whether gcd(N, modulus) == 1 , or ~assume~ , the latter is a mine that user will step on. unlike div0ism , it is not an inexpensive test . )
asciilifeform: ( why this is, is because for certain types of pubkeycrypto, you want to test adjacent nums for primality. rsa in particular. )
asciilifeform: oh hm i recall nao. ( it was because operator 'P' wants to be a general-purpose primality test, valid for any input whatsoever that fits in the ffawidth, rather than simply 'generate prime' )
mircea_popescu: then test against them.
asciilifeform: on top of this, if you actually carry out a diff stream of instructions for 2047 and 2048, you leak the bitness of the integer under test.
mircea_popescu: (i suppose if indeed you want to test MORE small primes than fit in one 8kb, you'll have a number of such composite numbers to test about. however many it takes. and yes, you can clever the knobs so they're not in strict order so that the composites are each exactly 8192 bits)
mircea_popescu: test as many small primes as their product is as many digits as your proposed large prime and be done with it, daykin will work ok for same bitness
asciilifeform: to entirely round off the chalkboard, mircea_popescu may be interested to know that no one has called the 1-800 since we last spoke of it. the current summed cost stands at 12 orcbux, and is set to increment by 22 orcbux/mo ( we've spent the vendor's 'test drive' bait). i'ma cover the lunch money cost of this item until given to know that it aint wanted.
asciilifeform: ffa is designed to be used via pcode (aka 'peh') but i'm not about to tell folx that they absolutely must. given the stated reqs (you gotta test for div0ism, we dun do it internally given as it's thermonuke performance) it can be safely used directly.
mircea_popescu: i imagine she's going to bake a test as time permits. it's not a top priority item but then again she moves fast.
asciilifeform: i promise to come back to this thrd, if asked. but would like to not lose the subthread earlier : diana_coman mircea_popescu : is short-term plan to test ffaistic diana_coman-r-m ? and if so, what glue is needed for this from asciilifeform , i'ma bake.
asciilifeform: mircea_popescu: i didnt try patching out the test ( given as it's got montgomery, the answer will be soup ) but did try for odd exp, got ~same~ as for the nonconstanttime gmp knob ( which is pretty hilarious, means that they dun do anything differently for 'constant time' variant, evidently )
mircea_popescu: patch out the test, i guess. they're "clever" like that.
asciilifeform: BingoBoingo: that's the usual heathen hypothesis (aka 'hormesis'). mircea_popescu seems to favour the other one (and seems moar plausible to asciilifeform) , where straight 'burn-in test'
asciilifeform: in other noose, earlier this wk , asciilifeform tried to repeat http://www.loper-os.org/?p=2906 test , but using 'gmp', the 'uncut' version of koch's thing, with asmism etc. but lo and behold, it is apparently impossible to repeat the full test battery, because :
mod6: Then can re-evaluate/re-test a bit if it's helpful.