2026-02-01 22:57:13 <iggy> https://milkv.io/jupiter2 no info on when it'll be available though (on their site or any others I looked at)
2026-02-01 22:57:39 <iggy> there is some more info on the cnx-software article about it
2026-02-01 22:58:13 <iggy> apparently it's only 8 x100 cores and 8 a100 AI cores, so not 16 generally usable cores
2026-02-01 22:59:09 <iggy> it supposedly does support hypervisor extensions, so could be useful for dev VMs or something
2026-02-01 23:00:49 <iggy> looks a lot like that spacemit devkit k3 pico-itx board
2026-02-01 23:02:43 <iggy> https://www.cnx-software.com/2026/01/30/jupiter-2-an-rva23-compliant-sbc-powered-features-spacemit-k3-octa-core-risc-v-ai-soc-up-to-32gb-ram-256gb-ufs/ some more info on the cores in the comments
2026-02-02 08:13:41 <mps> > iggy| apparently it's only 8 x100 cores and 8 a100 AI cores, so not 16 generally usable cores
2026-02-02 08:14:41 <mps> yes, there was few days ago discussion about that on #riscv irc channel
2026-02-06 16:22:00 <mps> so no one except ikke and me have bananapi-f3 and can post TLV values from eeprom (tlv_eeprom command in u-boot cli)
2026-02-06 22:29:56 <iggy> I only have the vf2 and licheepi 4a... I'm curious what the k5 is going to look like, otherwise, I might get a k3 based board once they start shipping
2026-02-06 22:34:13 <iggy> the k3 is listed at 60 TOPS of AI performance which I assume they attribute to the 8 A100 cores since there's no NPU listed... so the k5 is listed as 120TOPS, so it could be 16x A100 cores (instead of 8 X100 + 8 A100), 16 X100 + 16 A100, or something totally different
2026-02-09 12:45:07 <mps> https://www.akeana.com/akeana-tapes-out-highest-performance-rva23-alpine-test-chip/
2026-02-09 12:45:28 <mps> alpine chip
2026-02-09 12:57:29 <ikke> I suspect that has nothing to do with Alpine Linux
2026-02-09 13:38:50 <mps> i think so but was funny when saw it
2026-02-09 14:10:14 <ikke> A chip dedicated to Alpine Linux would be funny
2026-02-09 14:12:42 <mps> iirc there was arm64 SoC named Alpine
2026-02-09 14:15:14 <mps> besides fruit juice and clothes maker here :)
2026-02-10 01:01:58 <iggy> I was about to ask how they were getting 4nm litho, but then noticed they are based out of santa clara
2026-02-11 21:56:15 <runxiyu> could someone check out why https://gitlab.alpinelinux.org/runxiyu/aports/-/jobs/2215982 is failing? ty
2026-02-11 22:02:37 <ikke> 5Building Go toolchain2 using go_bootstrap and Go toolchain1.
2026-02-11 22:02:39 <ikke> SIGILL: illegal instruction
2026-02-11 22:03:15 <runxiyu> well, yeah, that's there
2026-02-11 22:03:27 <runxiyu> what's the cause of *that* and such, would be beyond my knowledge
2026-02-11 22:03:28 <ikke> Yeha, just providing context
2026-02-11 22:03:35 <runxiyu> ah
2026-02-11 22:04:19 <ikke> Maybe a go bug
2026-02-11 22:05:59 <ikke> or assumption that a certain extension is available
2026-02-11 22:06:08 <runxiyu> i dont have access to a riscv64 machine to debug it at all
2026-02-11 22:06:12 <runxiyu> :/
2026-02-11 22:07:37 <ikke> Let me try to get it to run on a different runner
2026-02-11 22:25:33 <achill> looks like it was just a fluke
2026-02-11 22:25:37 <achill> a retry worked
2026-02-11 22:25:53 <achill> i love go
2026-02-11 22:25:54 <achill> and riscv
2026-02-11 22:32:38 <runxiyu> :D
2026-02-11 22:33:03 <runxiyu> random SIGILL build failures are... definitely not a problem at all™
2026-02-18 11:21:07 <mps> booted alpine on spacemit musebook
2026-02-18 11:22:58 <mps> now have to make it good
2026-02-18 11:23:54 <mps> machine have 16GB ram and 512 GB nvme
2026-02-18 11:24:23 <mps> keyboard is fine
2026-02-18 11:24:55 <mps> display is - well - ok
2026-02-18 11:25:20 <runxiyu> how performant is the cpu?
2026-02-18 11:25:51 <mps> didn't tried anything heavy yet
2026-02-18 11:25:53 <runxiyu> the only things i know about riscv64: (1) the ISA is increadibly clean; (2) wow this builder is underpowered
2026-02-18 11:26:43 <mps> but their OS, debian based is good
2026-02-18 11:27:00 <runxiyu> huh?
2026-02-18 11:27:02 <mps> but i removed it
2026-02-18 11:27:05 <runxiyu> did you say you just booted alpine
2026-02-18 11:27:06 <runxiyu> ah
2026-02-18 11:28:05 <mps> can't understand how people could use such heavy distros ;)
2026-02-18 11:28:58 <mps> oh, wifi is very fine
2026-02-18 11:29:28 <mps> sound is ok-ish
2026-02-18 11:33:06 <runxiyu> i use alpine because its boring
2026-02-18 11:33:55 <mps> in what meaning
2026-02-18 11:37:56 <runxiyu> it doesnt surprise me and i dont have to worry about my distro
2026-02-18 11:38:08 <runxiyu> packaging things is really easy
2026-02-18 11:38:30 <ikke> Boring is considered a good quality for software
2026-02-18 11:38:58 <mps> yes
2026-02-18 11:39:30 <mps> and boring life is best life
2026-02-18 11:41:27 <mps> but again with alpine i always something to fix, improve, polish et cetera
2026-02-18 11:41:58 <mps> l always have*
2026-02-18 11:42:32 <mps> s/l/i/
2026-02-19 09:50:41 <mps> my first riscv as workstation with alpine \o/
2026-02-19 09:51:33 <mps> firefox is not so fast but bearable
2026-02-19 09:52:03 <mps> mpv works
2026-02-19 09:53:00 <mps> happily i don't use heavy graphic software
2026-02-19 09:53:49 <mps> anyway have to work to make gpu working
2026-02-20 07:03:57 <lindsay> mps: Congratulations! Does the GPU work out of box?
2026-02-20 07:19:25 <mps> lindsay: not with alpine
2026-02-20 07:20:53 <mps> kernel have driver and firmware but mainline mesa do not 
2026-02-20 07:21:54 <mps> there is SDK with mesa driver but it is behind NDA
2026-02-22 17:10:34 <mps> edge builders hung
2026-02-22 17:14:44 <ikke> mps: issue with coreutils test suite
2026-02-22 17:14:47 <ikke> I'm debugging it now
2026-02-22 17:16:27 <mps> ah ok. thanks for info
2026-02-23 14:45:45 <mps> nice, kernel 7.0 have 'riscv userspace control flow integrity (RISCV_USER_CFI)'
2026-02-25 11:55:50 <ncopa> has anyone tested the 6.19 kernel with bananapi or orangepi rv2?
2026-02-25 12:02:54 <mps> ncopa: you mean vanilla kernel
2026-02-25 12:03:19 <ncopa> yes
2026-02-25 12:05:28 <mps> I doubt it can works. maybe 7.0-rc1 can, i plan to test it in next few days
2026-02-25 12:06:54 <mps> first have to upgrade opesbi/u-boot spacemit
2026-02-25 12:08:58 <mps> this one works https://github.com/jmontleon/linux-spacemit/ except mmc
2026-02-25 13:41:46 <mps> 7.0-rc1 boots on bpi-f3 but mmc doesn't work
2026-02-25 14:17:38 <mps> ncopa: for me on bpi-f3 and musebook stable is only bianbu kernel 6.6.63
2026-02-25 14:58:08 <ncopa> ok, nice
2026-02-25 14:58:30 <ncopa> I suppose bianbu kernel is the best option for spacemit at this point
2026-02-25 15:03:02 <mps> yes :(
2026-02-25 15:04:20 <mps> but work is going to upstreaming drivers and 7.0 is promising
2026-02-26 20:25:29 <mps> do we really need new vim version every day :)
2026-02-27 17:43:42 <mps> https://ubuntu.com//blog/canonical-and-ubuntu-risc-v-a-2025-retro-and-looking-forward-to-2026
2026-02-27 17:44:22 <mps> interesting, they anounce K3 support
2026-02-28 11:09:58 <mps> I finished new bootable image for K1 SoC with new version of opensbi/u-boot (bianbu ones)
2026-02-28 11:11:25 <mps> it should boot on different boards but tested only on musebook
2026-02-28 11:13:26 <mps> it have activated usb keyboard and mipi-dsi in u-boot so can be tweaked in uboot cli shell
2026-02-28 11:39:14 <mps> ncopa: i'm thinking to add in testing linux-k1-bs (bs - bianbu source) to have stable kernel for K1
2026-02-28 11:39:41 <mps> or linux-spacemit-bs
2026-02-28 11:40:46 <mps> opinions/comments?
2026-02-28 13:44:31 <mps> algitbot: retry
2026-02-28 13:45:39 <mps> uh
2026-02-28 13:50:56 <mps> uh, made a mess