log
160 entries in 0.163s
asciilifeform: diana_coman: one painful way in which ic fab differs from softwarisms is that failure is ~very~ expensive. buggy proggy you can throw away, you lost a day or however long it took to write. buggy ic is a coupla 100k (usabux) in the hole.
asciilifeform: this kinda thing is where fab capability would be +9000 handy. but even without it, still doable, imho.
asciilifeform: concludes by summarizing (with illustration) the actual chem of the fab process.
asciilifeform: you can't make chip fab cheaper by asking it to live in group house 15 cucks to a room and to eat recycled noodles.
asciilifeform: what asciilifeform is looking for, re ic fab breakthrough, is specifically a process that would be to ic what the cd recorder was to cd
asciilifeform: looking at the 'process steps' docs in the linked page, it seems to be a straight wikipedization of ordinary schoolbook description of ic fab process ☟︎
asciilifeform: in fact, genius shoemaker pissed right into the mouth of the 'comcon' sanctions, bought semiconductor fab line straight from ti
asciilifeform: ( consider, sovok faced much simpler problem, when ~it~ resorted to fab -- that amerireich ~wouldn't sell'em~ the goodstuff. today, can't even ~buy~ the goodstuff if yer sultan of brunei )
asciilifeform: http://btcbase.org/log/2019-03-13#1902255 << i find it increasingly harder to avoid the thought that computing is going down to bottom of the sea unless and until we fab ☝︎
asciilifeform: i dun have a fab in my kitchen, i promise that would say if i did
asciilifeform: ( yep they have a private 2uM fab going. is why the old 'warthog' still flies, and other 'vhs'-era tech )
asciilifeform: ( at 'modern' micronicities, fab process is no longer standard but custom to the chem, and entirely diff at each fab house. hence the proprietary tooling crapolade. )
asciilifeform: but for some reason no ic fab does.
asciilifeform: and they fab photoclones of ameri-crud.
asciilifeform: picture what those folx could've done with modern fab. ☟︎
asciilifeform: imho the classical fab is an overwhelmingly incatronic tech, it centralizes unhealthily. ☟︎
asciilifeform: to briefly revisit upstack, asciilifeform's interest in ic fab largely revolves around http://btcbase.org/log/2018-04-30#1806521 . ☝︎
asciilifeform: btw the reason, afaik, why erry fab house forces 'standard cells', is that they have proprietary tweaks to their process , and have lib of cells ( kept seekrit ) that are known to work with said process.
asciilifeform: mircea_popescu: let's suppose we make the req'd contact. what wouldja want to fab 1st ?
asciilifeform: ( the other sub-$1mil fab is mosis co., but it is in usa, serves primarily usg, and dun publish prices, in the past i tried to get estimate from it but without success )
asciilifeform: granted in past 2y since i last looked, the mythical beast of 'small fab' could have been born
asciilifeform doesn't have a 'clean room' in the chip fab sense, but did install a multistage circulating filter thing in machine room, it actually worx pretty well re cutting downtime for fan cleanings etc
asciilifeform: however i suspect that 'alt-juniper' would be small change, if we had fab friends, oughta go straight to general-purpose washington-free comp.
asciilifeform: would need 1) fab capability 2) substantial time, unless mircea_popescu has a coupla qualified pairs of hands up his sleeve to assist
asciilifeform: they dun ~own~ fab, of course, just as google doesn't. they contract the fab.
asciilifeform: for one thing, if you can't fab ic, you inevitably 'hot glue konsoomer GB nics'
asciilifeform: unless i misunderstand somewhere, this needs ic fab.
asciilifeform: mats: mircea_popescu's 'peltier relay' would be quite trivial if we had fab resource, but we aint there yet
asciilifeform: world's largest fab house, nominally
asciilifeform: there's a fab in usa where, believe or not, they still make 2um classics like cdp1802 for usg. but they dun take walkins. ☟︎
asciilifeform: esthlos: i recommend to read whole set of search output earlier, as part of your log eating , there was even an instance where asciilifeform went and got actual quotes from small fab houses
asciilifeform: esthlos: http://btcbase.org/log/2016-04-22#1455187 << e.g. fab thread. ☝︎
asciilifeform: if you can build equiv of, say, dec vax ( as su , in its last yrs of life, did ) you have something like fab-sovereignty.
asciilifeform: !#s from:asciilifeform fab
asciilifeform: vertical integration is a thing, google bought a few power plants; but afaik not fab yet
asciilifeform: owning fab is not usually +ev
asciilifeform: lotsa folx have no own fab . e.g. sun microsystems didnt
asciilifeform: ( they dun have, afaik, a fab, the 'TWN' suggests they contracted out to e.g. infineon )
asciilifeform: rrright because when asciilifeform's fab house sent back proto-FG repeatedly with poor reflow and dud analogue board, asciilifeform complains 'from certain mental structure' and not because imbeciles all the way down like turtles on elephants.
asciilifeform: mircea_popescu: if we had any fab capacity to speak of, these'd be the priority items : 1) large homogeneous fpga 2) otp roms 3) 1+2 ☟︎
asciilifeform: ( iirc not long ago there was thread re the world's only 10nm fab, belonging to usg )
asciilifeform: well ~somehow~ he conceived of the notion that 'fab dun have to cost $bil' but took it no further than the folx discussed in the 'gas tank additive THEY dun want you to know!' thread
asciilifeform: d00d has nfi that classical ic fab is technopolitical dead end.
asciilifeform: ( for the sake of thread-completeness, what would the ~alternative~ to this story look like? i suggest -- it'd be a process which does to ic fab what 'polaroid' process did to colour photography. find way of etching the circuit from prefab 'sandwich' without caustic baths, sputtering, etc... ) ☟︎☟︎☟︎
asciilifeform: point being, d00d made 0 fundamental breakthrough that would enable 'cheap ic fab' or anything even with stretch of imagination describable as such a thing.
asciilifeform: but we dun have a cpu fab yet.
asciilifeform: caaddr: afaik not fully published ( in the can build fab immediately sense ) either
asciilifeform: the only cure for this is to fab own fpga.
asciilifeform: if it only needed pcb, everybody would just call up favourite pcb fab and then stfu
asciilifeform: asciilifeform lives next to a large reservation for these creatures. lab that makes usg-approved, 'law-abiding' drones; ic fab which makes safely-useless rubbish; electron microscope where asciilifeform is not invited within cannon's breadth, even for money; etc
asciilifeform: picture you as the enemy. you control the fab. what is your next move ?
asciilifeform: fib is interesting to be strictly from home-fab pov
asciilifeform: or desktop ic fab
asciilifeform: http://btcbase.org/log/2017-08-22#1701786 << asic miners. the original 'asics' were actual shaved fpga. the next gen -- were 'hardcopy fpga', the cheapest fab available, where you just get to define metallization layer and naught else ☝︎
asciilifeform: ( they didn't own a fab, either, contracted )
asciilifeform: because fab ain't cheap.
asciilifeform: at least until we have something like a fab
asciilifeform: incidentally, if there were such a thing as an honest ic fab, the scintillator-cum-photodiode-cum-debiaser could easily be an ic
asciilifeform: mircea_popescu: 'reasonable cost' in this context is ' asciilifeform's experimentation budget.' which presently does not stretch to ic fab.
asciilifeform: it is what intel's rdrand, via'd chips, etc use. tightly married to the vagaries of the fab process, and very difficult to tie to a specific physical effect
asciilifeform: i thought mircea_popescu was of the notion that you don't need economy of mass to do si fab.
asciilifeform: recently i found that even a plebe could, in theory, afford a few thou. units of 1980s-style fab.
asciilifeform: makes'em in modern fab, cheap, too, i have a bunch of samples here
asciilifeform: ( and picture if it could be made in modern fab, at GHz+ ! ) ☟︎
asciilifeform: kinda disgraceful, not like there isn't a usg 15nm fab or two still standing in upstate new york
asciilifeform: because the cost, to someone who ALREADY has fab capability , is miniscule
asciilifeform: or hm, nm, you need atomic layer epitaxy rig. you will need to infiltrate a semiconductor fab.
asciilifeform: these are 'kalash', they can be brought up with some patch wire, in a pinch, not even needing pcb, and exist in ~limitless junkyard supply, as does their ram and glue logic; and can be manufactured with sov-era fab tech, in turn can be re-created for less than what a dozen garbage trucks cost.
asciilifeform: for this sum one could aaaaalmost get a chip made at that french fab.
asciilifeform: the one 'republican ic' i would immediately put into production if i had a fab, is the humble antifuse rom !
asciilifeform: now! it ~is~ 'republican' in the sense where -- iirc -- moore eschews the traditional '1000 terabyte and 10x10^8 dollars' fab toolchain in favour of his own
asciilifeform: phf: no rush, it is not as if we had a chip fab in which to make clones, or even particularly want to make clone (i have 0 interest in perpetuating 20 years of mit undergrad young cocklet crapolade code)
asciilifeform: and tmsr-fab, and many other things.
asciilifeform: this'd be more of a 'why does restaurant kitchen have mice, looksy, intel's fab floor doesn't' 'are you ready to pay $50,000 a gram for the potato?'
asciilifeform: http://btcbase.org/log/2016-09-17#1544495 << go and make even a thing that helloworlds on a pci bus, without si fab. ☝︎
asciilifeform: (si fab etc.)
asciilifeform: in other nyooz, re 'ic fab in kitchen',
asciilifeform: drawing fiber is only slightly easier than chip fab
asciilifeform: and for some reason mircea_popescu is reluctant to confront the fact that even he can only remotely afford fab time because it is subsidized by the drool of a billion 'konsooomerz'
asciilifeform: unless you are mircea_popescu and can take over a whole fab.
asciilifeform: does mircea_popescu grasp how ic fab tooling works ?
asciilifeform: the only problem is that he has 100% of the fab capacity
asciilifeform: costs of chip fab as we know it, and of orbital launch, are even comparable. ☟︎
asciilifeform: nope, the latter - fab.
asciilifeform: they actually fab in usa.
asciilifeform: it would have to begin with time rented on ordinary fab, or the like.
asciilifeform: http://log.bitcoin-assets.com/?date=08-03-2016#1426499 << mircea_popescu found a chip fab between his sofa cushions yet ? ☝︎
asciilifeform: http://log.bitcoin-assets.com/?date=13-01-2016#1368908 << 'give si fab and then we talk' ☝︎
asciilifeform: given that i presently lack a fab
asciilifeform: go and fab cpu 'nights&weekends with bare hands.'
asciilifeform: the really galling thing, to me, re: the 200mil sc4mz0rz, is that for prolly less than half of that, a sane man could build a perfectly serviceable chip fab in the pampas or the like
asciilifeform: funkenstein_: not chip fab, either
asciilifeform: and si fab requires a fortress.
asciilifeform: so far, EVERY SINGLE SI FAB ON THE PLANET and ALL THE DESIGN TOOLING (with the exception of chuck moore's...) runs on winblowz ☟︎
asciilifeform: unless mircea_popescu pulls a cpu fab out of his shirt pocket, there is nothing preventing this disease from progressing.
asciilifeform: he assumed that men would remain equally capable while si fab advances
asciilifeform: space launch, si fab - no.
asciilifeform: but if anyone presently alive could fab a muller gate fpga, it is probably moore.
asciilifeform: (i.e. fab)
asciilifeform: also gotta point out that the $redacted analogue ic on the rng was prolly baked in a cn fab