Show Idle (> d.) Chans
Results 1 ... 90 found in asciilifeform for 'amd'
asciilifeform: copypaste: + amd's psp boobytrap, + whatever undoc'd. kilometres of thrd re subj in logs
dulapbot: Logged on 2022-07-28 11:34:54 asciilifeform: billymg: sadly asciilifeform never properly wrote recipe for xorg -- will say that it 1) muchly depends on the gpu backend yer using ( asciilifeform -- amd ) ; and 2) consisted of ~month of trial&error, manually emerging package by package with --no-deps and walking the barf tree
asciilifeform: billymg: sadly asciilifeform never properly wrote recipe for xorg -- will say that it 1) muchly depends on the gpu backend yer using ( asciilifeform -- amd ) ; and 2) consisted of ~month of trial&error, manually emerging package by package with --no-deps and walking the barf tree
asciilifeform: ( for amd, not even this type of 'cleaner' yet discovered )
dulapbot: Logged on 2022-01-22 09:58:30 asciilifeform: gregory5: otoh for amd's 'PSP' boobytrap, no defusing, partial or otherwise, known atm.
dulapbot: Logged on 2020-08-06 20:14:47 asciilifeform: for the most part gives very little shit re intel irons, either way. nao, a leak of e.g. amd's 'psp' boobytrap would be useful. but none to date.
asciilifeform: 1 of the reasons asciilifeform still uses exclusively amd 'opteron' in x86 boxen
asciilifeform: meanwhile, in the schizo reservations...
dulapbot: Logged on 2022-05-14 11:20:27 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-05-14#1102469 << buy an e.g. amd wx2100 (~100bux) and be done with it
dulapbot: Logged on 2021-11-07 20:24:04 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-07#1063729 << i got 1 of those amd cad card things that doesn't 3d ~at all. but worx with 100% blobless driver, and pushes 3 x '4k', which was all i needed.
asciilifeform: famous scamdrug, massively stockpiled by usg during bush ii's 'anthrax theatre', resulted in huge surplus that had to be pawned off somehow
asciilifeform: recall e.g. amd's swallowing of dec alpha
mats: nvidia ? amd?
billymg: seen in a heathen pit: "amd/intel will stop selling chips to russia, what does this mean? no more nsa backdoors"
dulapbot: Logged on 2022-02-21 10:25:44 asciilifeform: interestingly, some blox still take >20s to verify, and this on amd 3960x
asciilifeform: interestingly, some blox still take >20s to verify, and this on amd 3960x
dulapbot: Logged on 2022-02-14 19:34:32 verisimilitude: I was referring to `Amdahl's law'', in particular.
verisimilitude: I was referring to `Amdahl's law'', in particular.
dulapbot: Logged on 2022-01-22 09:58:30 asciilifeform: gregory5: otoh for amd's 'PSP' boobytrap, no defusing, partial or otherwise, known atm.
asciilifeform: for reference, 'kgpe-d16' is the last amd mb w/out 'psp', i.e. where can boot blobless bios.
dulapbot: (trilema) 2016-10-04 asciilifeform: 'AMD’s PSP is based around a single 32-bit ARM Cortex-A5, with its own isolated ROM and SRAM but has access to system memory and resources. It contains logic to deal with the x86 POST process but also features a cryptographic co-processor.' << they stuffed, finally, their 'fritz chip', into the g-series.
asciilifeform: gregory5: otoh for amd's 'PSP' boobytrap, no defusing, partial or otherwise, known atm.
asciilifeform: punkman: i've not tried 'epic', they were the 1st amd w/ builtin fritz
dulapbot: Logged on 2021-11-07 20:24:04 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-07#1063729 << i got 1 of those amd cad card things that doesn't 3d ~at all. but worx with 100% blobless driver, and pushes 3 x '4k', which was all i needed.
signpost: yeah, using the kernel amdgpu one
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-07#1063729 << i got 1 of those amd cad card things that doesn't 3d ~at all. but worx with 100% blobless driver, and pushes 3 x '4k', which was all i needed.
dulapbot: Logged on 2021-10-31 14:56:43 punkman: https://www.upstreamdata.ca/black-box-presale
asciilifeform: maybe tomorrow amd sells box where cache line is 1777 bytes strictly. somehow.
billymg: asciilifeform: re the gnat build errors, i'm looking at this: https://wiki.gentoo.org/wiki/Project:AMD64/Fixing_-fPIC_Errors_Guide
billymg: signpost: there are two companies working hard to decentralize minging now: http://upstreamdata.ca and http://gam.ai
dulapbot: Logged on 2021-08-01 21:26:37 asciilifeform: blew days on trying to transplant 'AMDGPU' to dulap-gentoo. so far w/out success.
asciilifeform blew days on trying to transplant 'AMDGPU' to dulap-gentoo. so far w/out success.
dulapbot: Logged on 2020-12-17 10:36:49 billymg: http://logs.nosuchlabs.com/log/asciilifeform/2020-12-16#1026433 << re: mining (de)centralization, there are at least two north american companies selling gas powered bitcoin miners to oil fields: https://www.upstreamdata.ca https://gam.ai
snsabot: (trilema) 2015-07-08 williamdunne: What happened to that German socialist girl, she ever come back to continue her derping?
cgra: asciilifeform: i went to explore the (originally mostly unknown to me) arm64 land after realizing amd64 opcode encoding is hideously complex... in the hopes of finding a better place to study binary auditability
thimbronion: asciilifeform: is this still a good guide for building a gentoo system? http://thebitcoin.foundation/gentoo-stage3-amd64-nomultilib-guide.txt
asciilifeform: in misc. lulz, 'amd threadripper' cpu apparently includes.... a miniature torque wrench.
asciilifeform: while on subj -- if any readers know of decent (i.e. take 80mm or larger modern-day fans) hedgehogs for amd 'g34' socket, plox to write in.
asciilifeform: which is pretty much same thing as e.g. 'intel ME', 'amd PSP', google's 'cr50', etc
snsabot: Logged on 2020-12-17 10:36:49 billymg: http://logs.nosuchlabs.com/log/asciilifeform/2020-12-16#1026433 << re: mining (de)centralization, there are at least two north american companies selling gas powered bitcoin miners to oil fields: https://www.upstreamdata.ca https://gam.ai
billymg: http://logs.nosuchlabs.com/log/asciilifeform/2020-12-16#1026433 << re: mining (de)centralization, there are at least two north american companies selling gas powered bitcoin miners to oil fields: https://www.upstreamdata.ca https://gam.ai
asciilifeform: bringing up amd cpu -- or even the realtek nic -- will certainly feature a good many kB of 'magic numbers no one knows meaning of' , sadly.
asciilifeform: qonfluent: good luck w/ that on any known x64 chipset. most of the heft of 'coreboot' in my builds is amd's boilerplate, but w/out which machine does not quite run
asciilifeform: theoretically this is unnecessary on amd-g boards, can use 'sage' + 'car' -- but as qonfluent has found, it is finicky
thimbronion: http://logs.nosuchlabs.com/log/asciilifeform/2020-12-04#1025870 << I don't recall my thinking at the time (this was back in 2016), but I had put some effort into building an amd/tyan box with ecc memory, and wanted to take things a step further and install an opensource bios. Maybe I just thought it would be cool to debug the boot process...
asciilifeform: even amd.
verisimilitude: There's also AMD, yes.
asciilifeform: trinque: sometimes bootloader is not even needed at all -- e.g. on apu1 & other amd g-series boxen, can bake kernel straight into bios
asciilifeform: factually, e.g. statically-linked trb binary has run on erry amd64 linux i've planted it on, incl. various heathen atrocities.
asciilifeform: if yer deploying on a truly ancient (pre-2.6 say) kernel, may have to adjust constants. otherwise worx on 100% of known amd64 linuxen.
asciilifeform: gregorynyssa: imho 'M' is a fairly complete guide not only to mips arch (as was intended) but to asm-on-amd64 in totality.
asciilifeform: shinohai: interesting. i wouldn't've thought these'd be folx starved by spamdemic (they work in restaurants?! or where..)
shinohai: http://logs.nosuchlabs.com/log/asciilifeform/2020-08-07#1018132 <<< still want to get an apu2 "someday" ... scamdemic isn't doing me any favors earning coin to pay for it tho.
asciilifeform for the most part gives very little shit re intel irons, either way. nao, a leak of e.g. amd's 'psp' boobytrap would be useful. but none to date.
asciilifeform: gregorynyssa: correct. and similar problem w/ post-2013 amd products (tho must note, the net mass of vendor liquishit they require to boot, is substantially smaller vs intel's)
snsabot: Logged on 2020-07-09 23:19:01 asciilifeform: aside from this, as recently as 2013 amd was making cpu that could boot 100% blobless. intel, iirc, last did in '08..
asciilifeform: ... or that plenty of amd boxen support debug probe that runs on any linux whereas intel...
asciilifeform: ... or the fact that 100% of amd cpu in past 15y support ecc. whereas intel artifically cripples it outta all but 'xeon'
asciilifeform: aside from this, as recently as 2013 amd was making cpu that could boot 100% blobless. intel, iirc, last did in '08..
gregorynyssa: asciilifeform: I continue to be drawn to Intel over AMD because of their better documentation. any reason to switch to AMD? I know that Intel cooperates with Intelligence but then again, all major American tech-companies probably do.
asciilifeform: (previously they were separate device, amd for instance made'em)
asciilifeform would readily pay coupla thou. for a 30hr+ portable that has no intel, amd, or otherwise x86 in it, boots blobless, readable display, etc
thimbronion: I recently had something of a monster water-cooled PC put together for something unrelated, also happened to have an amd graphics card laying around that meets the minimum requirements.
asciilifeform: draws <15W, fanless amd 'g-series' .
feedbot: http://qntra.net/2020/03/cross-process-cache-sidechannel-found-in-recent-amd-processors/ << Qntra -- Cross Process Cache Sidechannel Found In Recent AMD Processors
asciilifeform: these were (somewhat kludgy) amd64 ports of the original dec alpha emulator.
billymg: heya asciilifeform, just wanted to drop by and mention i've now used your dulap successfully on two machines, an amd FX desktop and a lenovo E545. in both cases it worked like a charm without any modification to the kernel config
asciilifeform: fromloper1121: let's round off w/ the uc datashit and rom lib datashit.
asciilifeform: fromloper1121: back and front photos of typical probe.
asciilifeform: fromloper1121: here is how to attach the testpoints to 'busblaster' jtag instrument.
asciilifeform: 'apu2' later from same vendor had the pad, but probe will not function, amd locked the hdt module and can only be unlocked with their rsa key
asciilifeform: the main difficulty remains -- getting hold of amd boards which actually have hdt header and where not crippled (on all post-2014 chipsets, crippled.) i was able to locate only 3 working examples on open market : 'gizmo 1', 'gizmo 2', and 'pcengines apu1' .
asciilifeform: fromloper1121: last thing -- amd's 'hdt' protocol, which is what the thing speaks, is standard jtag, but the commands are 'vendor commands' i.e. not documented afaik anywhere. imho your best shot to make clone would be to get hold of 1 working unit and properly capture the traffic on the hdt connector pins, and actually determine the format. then replicate with a standard jtag probe (tho iirc the amd speaks 1.2v, which most sta
snsabot: Logged on 2019-12-20 06:19:31 fromloper1121: in this device, however 1 BTC = $7000 - is really too much; moreover, I found a Gizmo being sold together with a SmartProbe for about $500, however this price is also way too high to be justifiable to not-a-millionare, considering there are plenty of coreboot-supported AMD boards which could provide a serial output in case something goes wrong and
fromloper1121: in this device, however 1 BTC = $7000 - is really too much; moreover, I found a Gizmo being sold together with a SmartProbe for about $500, however this price is also way too high to be justifiable to not-a-millionare, considering there are plenty of coreboot-supported AMD boards which could provide a serial output in case something goes wrong and
asciilifeform: unrelatedly: folx interested in smmism (smm was how boobytraps were often implemented prior to proper fritzchip) will find that it is detailed in amd opteron datashit; and that you can singlestep smm handler w/ 'sage' probe on irons where there is connector for same.
shinohai: Formerly was running it on my AMD board, which sadly lost cap over the summer - this is also on list of things to replace/be done.
asciilifeform: 2 x 1Ghz 'amd g series', 4GB/ram, 'samsung 860' ssd.
snsabot: Logged on 2016-10-04 14:41:09 kmalkki: note that RtmPubSigned.key[0x14..0x23] == AmdPubKey.bin[0x04..0x13]
bvt: currently i'm sure that we don't want to have multiple genesises for 3.x-m, 4.4-rockchip, 2.x-amd64 - but it's hard to tell how it will work in practice.
Random(asciilifeform) | Download hourly DB snapshot | Get Source Code