Wednesday, 2023-03-01

*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC01:09
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc01:39
sadoon[m]<programmerjake> "well, it's at least better..." <- Yeah I guess01:57
sadoon[m]Finally, now talking about gentoo-sffs:01:58
sadoon[m]0 failures rebuilding the base installed packages for proto-sffs (it's what I'm calling sffs until we make it an official compiler target)01:58
programmerjakenice!01:59
sadoon[m]Oh good lord I did something really stupid02:04
sadoon[m]Ignore what I said lmao02:04
sadoon[m]I edited the wrong file and waited 4 hours for it to finish building02:04
programmerjakeoops...02:05
sadoon[m]Also just discovered that I can build many packages all at once so I guess that's good news02:05
programmerjakeparallelism ftw!02:05
sadoon[m]You'd think I'd know this already02:06
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC02:32
*** josuah <josuah!~irc@46.23.94.12> has quit IRC02:57
*** josuah <josuah!~irc@46.23.94.12> has joined #libre-soc02:58
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc03:01
*** markos <markos!~Konstanti@static062038151250.dsl.hol.gr> has quit IRC03:46
*** markos <markos!~Konstanti@static062038151250.dsl.hol.gr> has joined #libre-soc03:46
sadoon[m]Too many failures in important stuff like ncurses, openssl, etc03:51
sadoon[m]A lot of these configure scripts are expecting power9 to have vsx03:51
sadoon[m]Unless there's already a compiler target that supports power9 without altivec or vsx I think the next step should be patching gcc to incorporate sffs03:52
sadoon[m]I'm off to work now though03:53
sadoon[m]Hopefully things aren't too chaotic since I took a small vacation lol03:53
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC04:05
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc04:29
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC05:06
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc05:30
*** Ritish <Ritish!~Ritish@60.243.42.218> has joined #libre-soc06:01
Ritishlkcl, I do have root access, but I dont know why some files have been missing, hmm06:03
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC06:12
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc06:36
*** Ritish <Ritish!~Ritish@60.243.42.218> has quit IRC06:52
lkclsadoon[m], watch out for particularly large packages: *all* linkers are completely borked, and it doesn't matter if you have 192 GB of RAM, if ever a batch of linking goes into swap-space it *will* make your machine fall over09:33
lkclRitish: it's rare but occasionally apt will retry retry retry give up. there's no warning. you just have to start the devscripts again.09:35
*** Ritish <Ritish!~Ritish@60.243.42.218> has joined #libre-soc09:39
Ritishnoted sir :)09:39
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC10:58
*** Ritish <Ritish!~Ritish@60.243.42.218> has quit IRC11:55
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc11:56
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC12:10
*** josuah <josuah!~irc@46.23.94.12> has quit IRC12:56
*** awilfox <awilfox!~awilfox@kelsey.foxkit.us> has quit IRC12:56
*** josuah <josuah!~irc@46.23.94.12> has joined #libre-soc13:00
*** awilfox <awilfox!~awilfox@kelsey.foxkit.us> has joined #libre-soc13:00
*** ghostmansd <ghostmansd!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC13:11
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.174.44> has joined #libre-soc13:12
sadoon[m]<lkcl> "sadoon, watch out for particular..." <- I'll take a look at the logs but it's definitely not a OOM issue, it rebuilt for power9 just fine before the sffs options13:17
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.174.44> has quit IRC13:41
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.174.44> has joined #libre-soc13:42
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.174.44> has quit IRC13:50
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@nat.222-104.maryno.net> has joined #libre-soc13:50
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@nat.222-104.maryno.net> has quit IRC14:07
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.174.90> has joined #libre-soc14:08
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.174.90> has quit IRC14:13
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@nat.222-104.maryno.net> has joined #libre-soc14:14
*** ghostmansd <ghostmansd!~ghostmans@nat.222-104.maryno.net> has joined #libre-soc14:32
*** ghostmansd <ghostmansd!~ghostmans@nat.222-104.maryno.net> has quit IRC14:42
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@nat.222-104.maryno.net> has quit IRC15:36
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.174.250> has joined #libre-soc15:37
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.174.250> has quit IRC15:54
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.55.3> has joined #libre-soc15:56
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.55.3> has quit IRC15:59
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.55.3> has joined #libre-soc15:59
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.55.3> has quit IRC16:08
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc16:08
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC17:09
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc17:46
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC18:05
lkclsadoon[m], when you go to parallel builds that's when you may run into the possibility of thrashing.  and it'll be really, *really* bad thrashing.18:05
lkclif you performed the entire P9 builds single-package-at-a-time you have so much RAM that every single link-phase would have been resident-in-memory18:06
*** ghostmansd <ghostmansd!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc18:10
sadoon[m]I'll rebuild the ones that failed one at a time18:29
sadoon[m]It all boils down to "openMP" support not being detected18:50
sadoon[m]Basically all the packages are giving me that error18:51
sadoon[m]I *can* try to disable openMP but these packages will not have parallel computing support I guess? https://packages.gentoo.org/useflags/openmp18:53
sadoon[m]Worth a shot since it seems to be the only blocking error which is good news :)18:53
sadoon[m]It's more likely that I borked gcc itself by disabling p9 specific features when building it, luckily there's VM snapshots for that :D19:01
markosthere seems to be just one Tyan 10-core P8 server left at https://www.ebay.com/itm/134327782109 (I bought the one before last :)19:11
markosin case anyone's interested19:11
markoson offer even19:11
sadoon[m]That's extremely cheap19:11
sadoon[m]Might pick it up19:11
sadoon[m]Hell this might help me finally replace my x86 server to a power one19:12
markosno memory/disks mind you19:13
markosbut these should be relatively easy to find19:13
markosand I asked them what kind of disk trays the server takes19:14
markosso that I would get the same19:14
markosthey had no idea19:14
markosI'm hoping they take compatible 3.5" ones19:14
markosand I'm pretty certain I will do a noctual/silent fan mod on this system19:15
markosI can't bear the rack fans noise levels19:15
sadoon[m]Ordered :D19:27
sadoon[m]DDR3 RDIMMs are very cheap these days19:27
sadoon[m]And since I'm replacing my current server I can reuse a lot of parts19:28
sadoon[m]Including storage19:28
sadoon[m]This means my website will hopefully be running on POWER819:28
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc19:28
sadoon[m]Thanks for the tip btw!19:33
markosnice, np :)19:53
markosmy plan exactly, I want to move most/all of my services to power19:53
markoswell, the ones I can anyway19:53
sadoon[m]Me personally I do believe all my services can be on power20:19
sadoon[m]My two servers are one for a website, gitea, email, nextcloud, and matrix, which all run on power just fine20:19
sadoon[m]The second is offline for storage only, which also works20:19
sadoon[m]A good opportunity to also run each service in a virtual machine on the first server instead of my current monolith20:20
sadoon[m]We have a meeting in 40 minutes correct?20:22
Ryuno-KiAndrJaen> <@sadoon:soulserv.xyz> My two servers are one for a website, gitea, email, nextcloud, and matrix, which all run on power just fine20:23
Ryuno-KiAndrJaen> The second is offline for storage only, which also works20:23
Ryuno-KiAndrJaen/me starts wondering whether sadoon will migrate to Forgejo at some point20:23
* sadoon[m] looks up Forgejo20:23
Ryuno-KiAndrJaenhttps://forgejo.org20:23
sadoon[m]Ah, fork of gitea20:24
Ryuno-KiAndrJaenI'm involved there 😻20:24
sadoon[m]Then I will definitely look into it :)20:24
Ryuno-KiAndrJaenDrop-in replacement (to a large extent).20:24
sadoon[m]Awesome thanks20:25
Ryuno-KiAndrJaenIf you are using Codeberg you already had contact with Forgejo20:25
Ryuno-KiAndrJaenCarry on.20:26
Ryuno-KiAndrJaenExcuse the interruption20:26
sadoon[m]No worries :)20:28
programmerjakeforgejo fixed one of the concerns luke had about switching to gitea...gitea is on github which isn't foss20:39
programmerjakethe other major concern he has is that he doesn't know go so isn't confident he can fix any emergency server issues that may arise20:40
programmerjakeso we're staying on gitweb and gitolite for now20:41
sadoon[m]For me something like gitweb or gitolite is actually more than enough, but it's fun having a nice interface so that's why I'm currently using gitea heh20:42
sadoon[m]At the end most of my work is done through git commands anyways20:43
programmerjakei prefer something with a nice ui like gitlab or forgejo20:44
sadoon[m]Is forgejo's UI any different from gitea?20:44
programmerjakemuch nicer when trying to read code on the web20:44
programmerjakenot that i know of20:44
sadoon[m]Being a drop in replacement I'll probably just try it out20:44
sadoon[m]Anyways, meeting in 15 right?20:44
programmerjakei don't have a meeting on my calendar, the openpower sync meeting is next week20:47
sadoon[m]Ah alright20:47
programmerjakeluke may have scheduled something else tho, icr20:47
sadoon[m]So I just found out part of the problem I'm having with sffs is missing _ieee128 data type, I tried -mno-float-128 (I think it was) but it didn't work20:48
sadoon[m]Seems like something I can fix with an option right?20:48
programmerjakeweird, float128 is supposed to be softfloat by default...maybe power9 enabled hw f128 by default and you forgot to change it back to softfloat?20:50
sadoon[m]But sffs should be hard float right?20:50
programmerjakenot for f12820:50
sadoon[m]Ah20:51
programmerjakejust for f32/6420:51
sadoon[m]So with it being -mcpu=power9, it's expecting hard 128 floats20:51
sadoon[m]Ok I forgot to set the right -mcpu anyways, let me try to fix that20:52
sadoon[m]Though I doubt it fixes it20:52
programmerjakethis is making me think we need to add -mcpu=generic-sffs20:52
sadoon[m]Yes, soon20:53
sadoon[m]That's what I was thinking20:53
programmerjakeidk about gcc, but i can try to add that to clang/llvm and rustc20:54
programmerjakenot right now tho20:54
sadoon[m]We could split the work then, I'd look into gcx20:54
sadoon[m]gcc*20:54
sadoon[m]Ok so it's -mno-float128-hardware but that also doesn't fix it20:54
programmerjakemaybe it's trying to shove f128 values into vsx registers which we disabled?20:55
sadoon[m]I can't tell but it's possible20:56
sadoon[m]"If you use the ISA 3.0 instruction set (-mpower9-vector or -mcpu=power9) on a 64-bit system, the IEEE 128-bit floating point support will also enable the generation of ISA 3.0 IEEE 128-bit floating point instructions. Otherwise, if you do not specify to generate ISA 3.0 instructions or you are targeting a 32-bit big endian system, IEEE 128-bit floating point will be done with software emulation. "20:56
sadoon[m]Targetting power8 doesn't seem to fix it anyway though20:58
sadoon[m]This is nice, I'm building a package that I know for a fact fails with our options and it's tiny (eix for gentoo)20:59
sadoon[m]This way I can test this package specifically before doing anything else20:59
programmerjakein other words gcc assumes isa v3 implies hw f128 even though they're separate features? sadness...21:20
sadoon[m]Hell even p721:28
sadoon[m]Very weird21:28
programmerjakehttps://gcc.gnu.org/wiki/Ieee128PowerPC21:43
sadoon[m]I'll have a read in a bit21:52
lkcltoshywoshy, ping PM (ISA WG)23:03
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC23:26
*** lkcl <lkcl!lkcl@freebnc.bnc4you.xyz> has quit IRC23:39
*** lkcl <lkcl!lkcl@freebnc.bnc4you.xyz> has joined #libre-soc23:45
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc23:50

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!