asciilifeform: re lappies : recently someone wrote to asciilifeform re an rk lappy supposedly in pre-production .

asciilifeform has witnessed chinese 'microfibers' that still managed to leave crapola behind, however

asciilifeform: 'microfiber' aha

asciilifeform brb:meat

asciilifeform: bvt: in principle oughta

asciilifeform: !Q later tell phf plox to snarf sig in http://bvt-trace.net/2019/05/unrolled-assembly-comba/ ( and diana_coman 's earlier ) at your earliest , into btcbase, ty

asciilifeform: re-grease (alumina, not silicone, greases), tighten.

asciilifeform: 1st with medical alcohol , then dry.

asciilifeform: use lint-free tampons on stick, the same kind as for rifles

asciilifeform: some folx dun clean, result is 'petrol station toilet' inside of lappy.

asciilifeform: when buying vintage lappies, dun fughet to scrub inside, and (if can get) replace fan.

asciilifeform: then again i do clean these things yearly.

asciilifeform: mine fwiw happily runs for week+ (never had occasion to longer) under 100% cpu, and no issues

asciilifeform: bvt: sounds like you have a dud inductor, rather than overheat

asciilifeform: iirc it is possible , with the sad coreboot, to peg fan to run at 100% forever. but then you have not a lappy but an 'air cleaner' ..

asciilifeform: cuz it has working fan control.

asciilifeform: with vendor's, never once overheated, under any load

asciilifeform: bvt: i had it overheat ~immediately when put under load with the 'rms coreboot', so went back to vendor's ( with above patch, or can't use civilized nic in it )

asciilifeform: who sold it to feed his dope habit

asciilifeform: at one time i had 'x61' also, but pronounced it sad and sent it to, iirc, gabriel_laddel

asciilifeform: ( aaalso subj of the infamous 'rms laptop' thrd. )

asciilifeform: http://btcbase.org/log/2017-01-11#1601544 << subj of above. ☝︎

asciilifeform: ( http://btcbase.org/log/2016-03-03#1421281 << orig. , i think, thrd re this specimen ) ☝︎

asciilifeform: the 1 i have , is this.

asciilifeform: iirc we had a thrd about this at one pt. x60 apparently came in at least 3 variants

asciilifeform: i.e. 1 of their 1st amd64 clones

asciilifeform: and yes running in longmode

asciilifeform: t7200 @ 2.00Ghz ; 4096kB cache

asciilifeform: aah there it is :

asciilifeform: brb

asciilifeform blows dust off x60...

asciilifeform: bvt: i'ma have to boot mine up an' see , then, cuz it's defo running a 64bit errything.

asciilifeform: btw spyked , outta curiosity, didja have to pay eurotax on that thing ? how much ? ☟︎

asciilifeform: ( 'toy' lappy , approx. size of vhs tape )

asciilifeform: x60 is sadly rare nao, but the rubbish box i sent to spyked has same chipset, he can say later if confirms this benchmark.

asciilifeform: lol

asciilifeform: 1 of the last, if not ~the~ last, intel chipsets w/out the 'me' boobytrap

asciilifeform: ( x60 , at least the 1 in my collection, is an early 64bit intel 'T2400' )

asciilifeform: fwiw

asciilifeform: bvt: with straight (no asmisms) 64bit ffa, i found ~exactly 2x slower on x60 than on the 2393

asciilifeform: this is trivially verifiable experimentally : build for 32bit ( 'MachineBitness' in http://www.loper-os.org/pub/ffa/hypertext/ch18/iron__ads.htm ) and get ~exactly 2x slower.

asciilifeform: ( and really, as observed by mp_en_viaje in prev. thrd, 'bus width bound' )

asciilifeform: in the parlance of seymor cray et al, ffa is 'cpu-bound problem'

asciilifeform: speed of ram, interestingly, has ~0 effect on ffa timing, given as the working set ~always fits in cache

asciilifeform: ecc ram etc.

asciilifeform: these won't win any speed records, but i like'em, they go years b/w downtimes no prob ( i take'em down strictly to clean fans )

asciilifeform: http://www.cpu-world.com/CPUs/K10/AMD-Third%20Generation%20Opteron%202393%20SE%20-%20OS2393YCP4DGI.html

asciilifeform: 512kB cache.

asciilifeform: 2393SE , 3.1ghz

asciilifeform: even mp_en_viaje's 'fx' i suspect.

asciilifeform: there's still plenty of usable irons faster than asciilifeform's tho.

asciilifeform: imho you defo wouldn't want to ffa on a post-2013 amd, or any intels etc aha

asciilifeform: ( note that asciilifeform's 'standard tester' box, where currently, sans any asmisms, ~1.7 s for same, is 3-4x slower than 'typical irons' of recent years )

asciilifeform: i expect in the end , say <100msec 4096bit modexp, will be achievable on typical irons

asciilifeform: so errything as expected.

asciilifeform: aha

asciilifeform: http://bvt-trace.net/2019/05/unrolled-assembly-comba/#selection-93.193-97.1

asciilifeform: pre-barrett -- small effect. with barrett -- substantial.

asciilifeform: mno

asciilifeform: mp_en_viaje: pre-barrett ffa , when rsaing, spends most of wallclock time in fz_mod , is all. which dun multiply. in 14+ , i fully expect 2-3x boost from asmism (possibly higher, with massage w/ magnifying glass for pipelineism etc )

asciilifeform: then can have 'speed demon' branch of the thing.

asciilifeform: there's no reason not to bake similar asmisms for e.g. add/sub/neg, shifts, etc. tho.

asciilifeform: bvt: iirc 80+% by wallclock

asciilifeform: i fully expect that the asmistic branch of ffa will be sewn out of bvt's item

asciilifeform: at any rate it's great, bvt did already 60-70% of what asciilifeform thought he would have to do himself after errything else

asciilifeform: from multiplication per se aha

asciilifeform: aha iirc in last ch, said, with exact figure.

asciilifeform: in < 14 , it'll ( and iirc he confirmed experimentally ) be just scarcely palpable diff.

asciilifeform: mp_en_viaje: bvt's asmism is correct near as i can tell ( i have not walked with magnifying glass. ) the 1 gotcha is that the actual win from'em aint measurable until bolt'em onto ch14b+ ( where knuth's division is no longer the bulk of the cpu load )

asciilifeform: neato

asciilifeform: ( ... or was this the 'partidul mp' hq thing..? )

asciilifeform: lol, portable en_viaje garden ?

asciilifeform: http://btcbase.org/log/2019-05-12#1912683 << asciilifeform was dumb enuff to buy a couple of these. opened, 'fuck, this is same material as in blog!' they're propping up monitors nao. ☝︎

asciilifeform: or to distill rotgut

asciilifeform: BingoBoingo: what's wrong with apple trees ?

asciilifeform: лесоповал , really not good for much else

asciilifeform: and yea can readily see mp_en_viaje's pt re the serfs

asciilifeform: at any rate i'ma let the folx who actually have 200k, think about whether good deal, for asciilifeform sadly academic q entirely.

asciilifeform: was re indoor sq m

asciilifeform: complete with serfs.

asciilifeform: would expect , in orcistan, a small town, for 200k

asciilifeform: strictly b/c price

asciilifeform presently w/out cow, and, believe or not, w/out own reactor

asciilifeform: and i suppose reactor also needs care or meltdown

asciilifeform: 3x is approx how often i cut it here lol

asciilifeform: 'paint wall shade #3 of blue cuz usg commissar proclaims ' - ugh

asciilifeform: meaningful maintenance , have no prob with

asciilifeform: to keep walls standing - yes

asciilifeform: mp_en_viaje: yes and it's reason # 9000 why asciilifeform wants out

asciilifeform: over here indeed fine by city etc

asciilifeform: keep jungle off walls is enuff in same jurisdictions.

asciilifeform: why even mow. what, gotta play football on it ?

asciilifeform: year aint 1700 nomoar, 'clean' dun have to mean 'chix on knees with tongue an' toothbrush'

asciilifeform: mp_en_viaje: robovac sweeps 100 m^2 / hr or so.

asciilifeform: iirc there were others also, with variations on theme.

asciilifeform: http://www.loper-os.org/pub/uy_sovok_1.jpg http://www.loper-os.org/pub/uy_sovok_2.jpg << subj .

asciilifeform: http://btcbase.org/log/2019-05-12#1912579 << i recall this, poster even had marx an' lenin on it, to draw orcs ☝︎

asciilifeform: with decent airflow - yea

asciilifeform: the chinese (fraudulently stamped 'taiyo yuden') caps, rot even when switched off.

asciilifeform: it also helps if board was cleverly designed , such as to put'em reasonably far from the 200 (400, in one of mine.. 2x) watt of cpu