| + chomwitt (~chomwitt@2a02:85f:9a42:1100:1ac0:4dff:fedb:a3f1) | 00:08 | |
| josch | m) yes, i really did not consider this as a possibility of how it works XD | 00:10 |
|---|---|---|
| josch | (says the one who just soldered wires from said trackball usb to the card manually) | 00:11 |
| - chomwitt (QUIT: Ping timeout: 252 seconds) (~chomwitt@2a02:85f:9a42:1100:1ac0:4dff:fedb:a3f1) | 00:21 | |
| minute | josch: hehe | 00:30 |
| - xktr (QUIT: Ping timeout: 264 seconds) (~xktr@user/xktr) | 00:30 | |
| + xktr (~xktr@user/xktr) | 00:30 | |
| - mjw (QUIT: Ping timeout: 244 seconds) (~mjw@gnu.wildebeest.org) | 01:03 | |
| - aelius (QUIT: Remote host closed the connection) (~aelius@user/aelius) | 01:29 | |
| + aelius (~aelius@user/aelius) | 01:30 | |
| + chomwitt (~chomwitt@2a02:85f:9a42:1100:1ac0:4dff:fedb:a3f1) | 01:46 | |
| - voltaire28_ (QUIT: Ping timeout: 240 seconds) (~jlafon@28.162.2.93.rev.sfr.net) | 02:10 | |
| - chomwitt (QUIT: Ping timeout: 244 seconds) (~chomwitt@2a02:85f:9a42:1100:1ac0:4dff:fedb:a3f1) | 02:21 | |
| - paperManu (QUIT: Ping timeout: 240 seconds) (~paperManu@146.71.9.156) | 02:55 | |
| - nsc (QUIT: Ping timeout: 246 seconds) (~nicolas@143-98-142-46.pool.kielnet.net) | 03:14 | |
| + nsc (~nicolas@i5C74DE29.versanet.de) | 03:15 | |
| - bkeys (QUIT: Quit: With every step we take, danger will follow closely) (~Thunderbi@98.19.128.69) | 05:03 | |
| + bkeys (~Thunderbi@h69.128.19.98.dynamic.ip.windstream.net) | 05:03 | |
| - bkeys (QUIT: Client Quit) (~Thunderbi@h69.128.19.98.dynamic.ip.windstream.net) | 05:04 | |
| + bkeys1 (~Thunderbi@98.19.128.69) | 05:04 | |
| - bkeys1 (QUIT: Client Quit) (~Thunderbi@98.19.128.69) | 05:06 | |
| + bkeys (~Thunderbi@98.19.128.69) | 05:06 | |
| - paperManu_ (QUIT: Ping timeout: 246 seconds) (~paperManu@146.71.9.156) | 05:56 | |
| - spew (QUIT: Quit: nyaa~) (~spew@user/spew) | 06:36 | |
| josch | gordon1: i really appreciate your input on the lte card -- it's not even super expensive compared to the em06-e and if it has connections with the purism phone and is otherwise getting hacked on, it sounds like the perfect fit for a computer like the reform. :) | 07:56 |
| josch | it should fit nicely in the space between display connector and usb ports | 07:57 |
| josch | now that m.inute corrected my understanding of realleif's solution i am also wondering whether doing it like this would not even be better because then the mini pci-e could get populated by the intel ax210 card | 07:59 |
| + chomwitt (~chomwitt@2a02:85f:9a42:1100:1ac0:4dff:fedb:a3f1) | 08:24 | |
| + voltaire28_ (~jlafon@28.162.2.93.rev.sfr.net) | 09:06 | |
| + libercv (~libercv@user/libercv) | 09:08 | |
| - voltaire28_ (QUIT: Ping timeout: 240 seconds) (~jlafon@28.162.2.93.rev.sfr.net) | 09:19 | |
| + voltaire28_ (~jlafon@28.162.2.93.rev.sfr.net) | 09:25 | |
| - voltaire28_ (QUIT: Ping timeout: 245 seconds) (~jlafon@28.162.2.93.rev.sfr.net) | 10:03 | |
| - libercv (QUIT: Quit: Konversation terminated!) (~libercv@user/libercv) | 10:48 | |
| + voltaire28_ (~jlafon@28.162.2.93.rev.sfr.net) | 10:48 | |
| ch | josch: do you know of any FEX packaging efforts? | 11:00 |
| vkoskiv_ | funny timing, I just set out to build FEX from source a few minutes ago :D | 11:05 |
| josch | ch: i packaged box64 and fex because fex needs a rootfs somewhere and how do you ship a rootfs via a package? | 11:06 |
| josch | people keep telling me how fex is so much better than box64 so if i'm wrong with that assumption i'd like to be corrected | 11:06 |
| vkoskiv_ | I don't know which is better, I just watched this yesterday: https://www.youtube.com/watch?v=3yDXyW1WERg | 11:07 |
| josch | ah the 39c9 talk | 11:08 |
| josch | errr 39c3 | 11:08 |
| vkoskiv_ | And, somewhat related, I hope everyone here who was at 39C3 had a good time! I was invited to go but couldn't make the timing work | 11:08 |
| josch | well, it must be nice to get support from valve ;) | 11:08 |
| gordon1 | josch: i would advice against putting foss firmware on it if you value battery life, it sorta broke mine in some way so even after returning vanilla firmware it still cannot enter low-power mode, didn't figure it out yet | 11:08 |
| josch | gordon1: noted. I indeed want to primarily "use" it and not tinker with it. :) | 11:09 |
| josch | i just got the shipping notification of the card :) | 11:09 |
| gordon1 | right, so stick with vanilla then and i'll update you if it is ready for foss :) | 11:09 |
| josch | nice :) | 11:10 |
| josch | gordon1: i also just learned about the source code of leif's pci-e board: https://github.com/leifhelm/mnt-mpcie-usb-hub | 11:10 |
| vkoskiv_ | josch: This is how fex handles the rootfs stuff: https://wiki.fex-emu.com/index.php/Development:Setting_up_RootFS | 11:10 |
| josch | yes, and box64 just uses multiarch | 11:11 |
| josch | how fex does it seems to be a trip back into the past when qemu also still required a rootfs for user-mode emulation | 11:12 |
| gordon1 | nice, i may or may not steal the hub part completely | 11:12 |
| gordon1 | (no license >:o ) | 11:12 |
| josch | (i just asked about that via fedi) | 11:13 |
| ch | josch: i havent checked at all what fex needs, but it seems to be one of the two options for wine arm64->x86 | 11:32 |
| ch | josch: and box64 will also need significant work going forward | 11:33 |
| ch | josch: i understand fex is what steamvr will use, maybe i'm misinformed | 11:33 |
| josch | yes, fex is what steam/valve is using | 11:34 |
| josch | and yes, box64 needs continuous maintenance and patching | 11:35 |
| josch | vkoskiv_: in the 39c3 talk, is a comparison drawn to box64? | 11:38 |
| ch | but also it will need a mingw64 aarch64 clang | 11:51 |
| ch | not looking forward to that either | 11:51 |
| josch | ah and there was that :) | 11:51 |
| josch | sometimes our standards get into the way of practicality | 11:51 |
| - Ar|stote|is (QUIT: Quit: No Ping reply in 180 seconds.) (~linx@149.210.3.83) | 12:09 | |
| + Ar|stote|is (~linx@149.210.3.83) | 12:14 | |
| josch | gordon1: https://social.sdfeu.org/@realleif/115830886760654564 | 12:25 |
| - nsc (QUIT: Ping timeout: 244 seconds) (~nicolas@i5C74DE29.versanet.de) | 12:37 | |
| - voltaire28_ (QUIT: Ping timeout: 240 seconds) (~jlafon@28.162.2.93.rev.sfr.net) | 12:52 | |
| minute | ch: josch: fex seems to be packaged for ubuntu (by fex), i wonder how far away that is from being able to be ported to debian | 13:11 |
| - robin__ (QUIT: Ping timeout: 256 seconds) (~robin@user/terpri) | 13:17 | |
| + robin (~robin@user/terpri) | 13:18 | |
| josch | minute: practically speaking we can package anything -- this is not about the technical rules but about the social rules like DFSG, embedded code copies and other packaging rules | 13:20 |
| minute | josch: ah i'm reading the log in more detail now, sorry :D | 13:21 |
| minute | so the rootfs is a problem | 13:22 |
| minute | somehow i thought the rootfs can be downloaded with a helper script and isn't part of the .deb | 13:22 |
| minute | i wonder if the rootfs is special or it could also be created by debootstrap etc | 13:23 |
| josch | minute: yes and then the package would go into "contrib" and not "main" but at least personally i'm not very motivated to package something for contrib | 13:23 |
| minute | got it, and very understandable | 13:23 |
| + voltaire28_ (~jlafon@28.162.2.93.rev.sfr.net) | 13:28 | |
| vkoskiv_ | fex does have support for accessing the rootfs directly from a squashfs file. It gives an option to extract it under ~/.fex-emu/RootFS/, presumably for performance reasons? | 13:29 |
| + mjw (~mjw@gnu.wildebeest.org) | 13:29 | |
| vkoskiv_ | I just chose the ubuntu 24.04 rootfs, the wiki advised to choose the "most similar" one, if the host distro isn't in the list. | 13:30 |
| vkoskiv_ | Building and getting fex set up went without issue, I can now run x86_64 binaries from under that rootfs with (presumably) binfmt_misc | 13:31 |
| vkoskiv_ | Now I'm trying to figure out how to install the amd64 steam deb :D | 13:31 |
| vkoskiv_ | I did do `dpkg --add-architecture amd64`, but I think I need something more, getting closer now. | 13:34 |
| josch | vkoskiv_: if you built it yourself -- where did you ge the mingw clang build from? | 13:34 |
| josch | vkoskiv_: and you probably want their wine fork as well because: | 13:35 |
| josch | "Upstream wine has various things missing for full FEX/ARM64EC support, while it can be used, my fork" https://wiki.fex-emu.com/index.php/Development:ARM64EC | 13:35 |
| vkoskiv_ | I just followed this: https://wiki.fex-emu.com/index.php/Development:Setting_up_FEX | 13:35 |
| vkoskiv_ | No mention of mingw clang here? | 13:35 |
| vkoskiv_ | And yeah, I was planning on trying linux native amd64 things first, then worrying about wine once I get that going. | 13:36 |
| vkoskiv_ | Good to know they recommend a fork | 13:36 |
| josch | vkoskiv_: "Unlike Linux FEX, the Windows FEX modules purely handle CPU emulation, with syscall translation being done on either the wine side (WOW64) or ABI (ARM64EC)." | 13:39 |
| vkoskiv_ | Okay, my issue was that I downloaded steam_latest.deb, when I should have downloaded steam-launcher_latest_all.deb | 13:40 |
| minute | vkoskiv_: why not apt install steam? | 13:40 |
| minute | ah no, i've confused that with box64's helper script, please ignore :D | 13:41 |
| vkoskiv_ | It doesn't show up in my apt sources | 13:41 |
| * vkoskiv_ -> vkoskiv | 13:43 | |
| vkoskiv | One day I should figure out my IRC setup so my nick doesn't go changing itself :D | 13:44 |
| vkoskiv | I installed the valve steam launcher deb, and it's now launching steam. | 13:44 |
| vkoskiv | "Updating Steam runtime environment" | 13:44 |
| vkoskiv | Guide didn't mention I need to install steam-libs-{amd64,i386}, but I get the feeling I may need to do that. | 13:46 |
| vkoskiv | Quite a few errors in the terminal, but Steam is still quite happily setting things up. It's at "Updating Steam..." now | 13:47 |
| vkoskiv | Fully expecting to fire up Cyberpunk 2077 with RTX: ON any moment now.. :D | 13:48 |
| vkoskiv | I'm seeing a bunch of "panfrost: driver missing" followed by "glx: failed to create dri3 screen", but that doesn't seem to be a problem so far, I've seen several windows pop up. | 13:52 |
| vkoskiv | It's still setting things up, I think it's almost done now. | 13:53 |
| vkoskiv | Login window popped up! | 13:54 |
| vkoskiv | Steam is up and running now. | 13:58 |
| f_ | What's the difference between box64 and FEX is what I never understood | 13:59 |
| f_ | and also what's preventing them from working together | 13:59 |
| vkoskiv | https://vkoskiv.com/screenshots/2026-01-03T13:01:56UTC-t5Ah64bl.png | 14:02 |
| vkoskiv | lol @ "GenuineIntel Cortex-A73" | 14:03 |
| minute | vkoskiv: nice | 14:04 |
| minute | vkoskiv: i wonder why not cortex-a76 | 14:04 |
| vkoskiv | I'm running the a311d | 14:05 |
| minute | ohhh ok | 14:05 |
| + paperManu (~paperManu@146.71.9.156) | 14:07 | |
| f_ | I didn't know Intel made consumer arm CPUs | 14:07 |
| f_ | very genuine | 14:07 |
| vkoskiv | I'll just try to install Portal and see if that works. Steam system info does show it's using llvmpipe, so probably won't be 60FPS. | 14:08 |
| vkoskiv | Processing vulkan shaders | 14:12 |
| vkoskiv | It started! Valve logo & music going on :D | 14:14 |
| minute | heh | 14:15 |
| minute | vkoskiv: do you have working vulkaninfo / vkcube on the arm side? | 14:16 |
| vkoskiv | vulkaninfo says "Found no drivers!" | 14:18 |
| vkoskiv | Portal is running now, in the main menu. It's a i386 binary. The guide mentioned needed workarounds to get 32-bit stuff working, but it's just running? Maybe outdated info then. | 14:19 |
| vkoskiv | Maybe I'll debug gpu stuff next. Is it realistic to expect the mali-g52 to be able to run Portal? | 14:20 |
| minute | vkoskiv: do you have mesa-vulkan-drivers installed? | 14:20 |
| minute | vkoskiv: because you should have vulkan nowadays with uptodate mesa... | 14:20 |
| vkoskiv | In the portal main menu I get 1 frame about every 5 sec, which is to be expected with software emulation, I think. Let me check that. | 14:21 |
| minute | yeah | 14:21 |
| + paperManu_ (~paperManu@146.71.9.156) | 14:21 | |
| vkoskiv | The fact that this is even running is pretty wild IMO. | 14:21 |
| minute | yes indeed | 14:21 |
| vkoskiv | I don't have mesa-vulkan-drivers installed. I'll install that. | 14:22 |
| vkoskiv | Now vulkaninfo shows stuff, and vkcube works too | 14:23 |
| vkoskiv | though vkcube uses llvmpipe again | 14:23 |
| minute | hm for me on rk3588, vkcube shows mali-g610 | 14:24 |
| minute | vkoskiv: one can pass --gpu_number to vkcube | 14:25 |
| vkoskiv | Ah, I see vulkaninfo shows only llvmpipe, no other GPUs. I'll try to reboot perhaps | 14:28 |
| vkoskiv | rebooted, vulkaninfo --summary shows GPU0 as llvmpipe, and no other GPUs. So I'll debug that next. | 14:31 |
| vkoskiv | I should note that I'm still running the same install I've had since ~2022, I've just installed updates, but my config has diverged somewhat from the latest reform system images I've tried. | 14:34 |
| vkoskiv | I should try to harmonize things somewhat at some point, there are some very neat things I saw when I last booted a fresh SD card :D | 14:35 |
| minute | vkoskiv: ah ok | 14:35 |
| vkoskiv | I really like the OOBE stuff that's in there now. | 14:35 |
| minute | cool, that's good to hear | 14:35 |
| minute | g52 is bifrost, which is supported by panvk, but not yet fully conformant i think, maybe it needs an env var to enable? only guessing tho | 14:36 |
| vkoskiv | I'll try some things | 14:36 |
| - chomwitt (QUIT: Ping timeout: 244 seconds) (~chomwitt@2a02:85f:9a42:1100:1ac0:4dff:fedb:a3f1) | 14:38 | |
| + siviq (~siviq@user/siviq) | 14:40 | |
| vkoskiv | I'll actually start by tending to some of these reform-check items I've been putting off ^^ | 14:45 |
| - mjw (QUIT: Remote host closed the connection) (~mjw@gnu.wildebeest.org) | 15:21 | |
| + mjw (~mjw@gnu.wildebeest.org) | 15:21 | |
| minute | > [ OK ] Found device dev-disk-by\x2dlabel-reformsdboot.device - SanDisk_3.2Gen1 reformsdboot. | 15:38 |
| minute | woah woah, qcs8550 can now find reform system image on usb stick | 15:38 |
| ch | josch: re rootfs: 1) for the wine support, apparently no rootfs is needed 2) wondering why the rootfs cant be just / on debian | 15:40 |
| ch | josch: 2) ... provided users install :i386 stuff. will try to understand this in a bit | 15:41 |
| ch | josch: 3) unsurprisingly the wine parts for fex also need the mingw64 aarch64 clang, so that has to happen either way | 15:41 |
| ch | why is there a long moment where the mouse cursor cannot be moved when logging in, after entering the password. is that the time gnome-shell needs to start? | 15:42 |
| f_ | neat! | 15:42 |
| f_ | (re qcs8550) | 15:42 |
| vkoskiv | Is it safe to flash uboot on eMMC on A311D these days? | 15:53 |
| vkoskiv | I've been happily booting with an SD card, so I haven't checked this recently. | 15:54 |
| minute | vkoskiv: yeah | 15:54 |
| minute | vkoskiv: make sure all your reform-tools etc are up to date first | 15:54 |
| vkoskiv | Will do, thanks. | 15:54 |
| f_ | vkoskiv: I'd make sure there's a way to recover first | 15:55 |
| f_ | BPI-CM4 has testpoints that supposedly enable USB boot but that never worked for me/only entered it in their u-boot vendor fork | 15:55 |
| f_ | and also you'd need an actual cm4io board even if it did work, as I understand it USB mode isn't really exposed on reform | 15:56 |
| f_ | normally it's supposed to be exposed to the first usb port enumerated (on the CM4IO that'd be the USB-C port) | 15:56 |
| f_ | Anyway seems like u-boot on eMMC is quite well tested these days so I'd worry less about that | 15:57 |
| vkoskiv | I don't think I do, or I'm at least unfamiliar with that process. As far as I can remember, the risk was it no longer booting off the sd card if the emmc gets flashed with a misconfigured uboot | 15:57 |
| f_ | yes, that's the risk | 15:57 |
| f_ | it's less risky today than it was back in the early A311D Reform days though I guess. | 15:58 |
| f_ | (why BPI-CM4 has an integrated eMMC instead of removable and no convenient way of entering bootROM USB mode is a question I'm still waiting a response for ;) | 15:59 |
| minute | got first 7z b scores for 8550 | 16:02 |
| f_ | does qcs8550 have secureboot enabled | 16:02 |
| minute | no idea | 16:03 |
| minute | also has a hypervisor i think | 16:03 |
| minute | (gunyah) | 16:03 |
| f_ | Yeah I heard of that | 16:03 |
| minute | anyway, first 7zb total score 30937 | 16:03 |
| minute | (without heatsink, and no idea about perf governor) | 16:03 |
| minute | on rk3588, total score 20029 | 16:03 |
| f_ | Not bad! | 16:03 |
| minute | yeah 33% multicore speedup in this benchmark | 16:04 |
| minute | but at the same time doesn't get hot... | 16:04 |
| minute | oh it downclocks to 400-500mhz in idle, that's neat | 16:04 |
| f_ | That's expected, it's supposed to be a phone SoC IIRC? Or based on such | 16:05 |
| minute | yes, sm8550 is in a bunch of phones and qcs8550 is the iot version of that without the modem | 16:05 |
| f_ | Would be unusual if a phone SoC did get that hot | 16:05 |
| minute | yes exactly | 16:05 |
| minute | that's one big reason why i wanna use qcom in the tablet | 16:06 |
| f_ | Especially considering how bloated android is these days | 16:06 |
| minute | but we're gonna use linux and not android ofc :D | 16:06 |
| minute | in this first test, the cores "only" go to 2956mhz, that's less than advertised iirc, so i guess there's more possible | 16:07 |
| minute | with boost, the first benchmark round goes to 34304 | 16:13 |
| minute | lol i can't math, it's of course 50% speedup. up to 70% | 16:17 |
| + AnimaInvicta (~AnimaInvi@88-120-179-216.subs.proxad.net) | 16:18 | |
| f_ | pretty exciting | 16:20 |
| - voltaire28_ (QUIT: Ping timeout: 240 seconds) (~jlafon@28.162.2.93.rev.sfr.net) | 16:21 | |
| f_ | minute: as I understand it (though I'm not a qualcomm expert) secureboot being disabled would allow a bit more control over fw running, like you wouldn't have to chainload u-boot from some other bootloader | 16:22 |
| f_ | Also some of these IoT SoCs are also being used on phones, like the Fairphone 5 has an IoT-version of sc7280 (w/ modem though) | 16:23 |
| f_ | qcm6490 iirc | 16:23 |
| minute | f_: yeah that's the qcm6490. we also have qcs6490 module which is the iot version of that, and brought that up a few months ago | 16:24 |
| minute | f_: but in my benchmarks the qcs6490 is around the same cpu speed as rk3588 | 16:24 |
| minute | f_: so i'm a bit more excited for the 8550 | 16:24 |
| minute | but dunno yet about pricing differences | 16:24 |
| f_ | oh wait you're doing multiple qcom SoC modules? | 16:24 |
| minute | f_: yes because they are mostly pin compat | 16:25 |
| f_ | pretty cool | 16:25 |
| minute | f_: 6490, 8550 and 8750 (don't have 8750 samples yet though) | 16:25 |
| minute | 8750 is the 8 elite with oryon cores | 16:25 |
| f_ | that's really neat | 16:30 |
| f_ | and yeah qc[sm]6490 isn't all that powerful | 16:31 |
| + nsc (~nicolas@i5C74DE29.versanet.de) | 16:39 | |
| ch | minute: is there more of a datasheet for the eremit cells than https://www.eremit.de/p/eremit-3-7v-4-000mah-high-cap ? | 16:51 |
| ch | minute: nevermind, found it on the website | 16:52 |
| minute | ch: nice | 16:52 |
| amospalla | if this is table is relevant, the 8550 more than double the performance of rk3588 in single core. https://www.cpubenchmark.net/compare/4906vs5740/Rockchip-RK3588-vs-QTI-QCS8550 | 16:55 |
| amospalla | while also being much cooler. | 16:56 |
| minute | amospalla: yeah but that's the only benchmark out there... | 16:56 |
| amospalla | true | 16:56 |
| minute | so i wanted to confirm it myself :D | 16:56 |
| amospalla | If you do some single core benchmarks please give some numbers :D | 16:57 |
| minute | but with which tool? | 16:58 |
| amospalla | something like compressing a file? | 16:58 |
| - bkeys (QUIT: Quit: With every step we take, danger will follow closely) (~Thunderbi@98.19.128.69) | 16:59 | |
| + bkeys (~Thunderbi@98.19.128.69) | 17:00 | |
| ch | i wonder if we fuck up Vbatt somehow | 17:03 |
| + voltaire28_ (~jlafon@28.162.2.93.rev.sfr.net) | 17:10 | |
| - bkeys (QUIT: Quit: With every step we take, danger will follow closely) (~Thunderbi@98.19.128.69) | 17:16 | |
| + bkeys (~Thunderbi@h69.128.19.98.dynamic.ip.windstream.net) | 17:16 | |
| vkoskiv | I wonder why my u-boot version shows as "2024.04-dirty" even though I just flashed the latest u-boot to both my SD card and eMMC | 17:19 |
| vkoskiv | I vaguely recall doing something related to u-boot around that time, so it's probably something I did. | 17:20 |
| vkoskiv | I've made reform-check much happier. And also purged a few dozen kernel packages dating back to 5.x days. I guess I forgot about apt purge and what it does. | 17:22 |
| josch | vkoskiv: if the problem with the u-boot version persists, please elaborate | 17:24 |
| josch | vkoskiv: with purge you also remove config files. By default, configs are retained on package removal. | 17:24 |
| vkoskiv | reform-check reports: 'I: Version of U-Boot: 2024.04-dirty' | 17:25 |
| vkoskiv | It also reports that both the eMMC and SD card contain the latest u-boot version, 2024-12-23. | 17:26 |
| vkoskiv | My expectation was that I'd see the u-boot version from that /proc file show 2024-12-23 after I've rebooted, but it stayed the same. | 17:26 |
| josch | vkoskiv: oh that's part of how u-boot is built | 17:27 |
| josch | this is fixed with the next reform-tools release | 17:27 |
| vkoskiv | Ah, so that's not my doing then, okay. | 17:27 |
| josch | the next reform-tools release is blocked on https://source.mnt.re/bugs/bugs/-/work_items/50 | 17:28 |
| amospalla | minute: this script replicates the test from https://openbenchmarking.org/test/pts/compress-gzip : https://paste.debian.net/hidden/75c5a5e2 | 17:28 |
| amospalla | if you find it relevant to use. | 17:29 |
| amospalla | my intel 14500 takes 26.5 seconds, my rk3588 53 seconds. An intel n100 takes 48 seconds. Not bad for the rk3588. But as this is only a gzip test, and does not mean anything besides that. | 17:35 |
| minute | amospalla: isn't this io bound? are you running it on /tmp ? | 17:37 |
| amospalla | I did on the pocket | 17:37 |
| amospalla | It compresses twice the kernel, so 1.6GiB or so of data to read. | 17:38 |
| - bkeys (QUIT: Quit: With every step we take, danger will follow closely) (~Thunderbi@h69.128.19.98.dynamic.ip.windstream.net) | 17:38 | |
| + bkeys1 (~Thunderbi@98.19.128.69) | 17:38 | |
| - bkeys1 (QUIT: Client Quit) (~Thunderbi@98.19.128.69) | 17:38 | |
| amospalla | I did not on the desktop, but the results it gives is what you would expect accordingly to the openbenchmarking website results list. | 17:39 |
| vkoskiv | ran reform-emmc-bootstrap, and now it happily boots without an SD card. Nice! | 17:39 |
| + bkeys (~Thunderbi@98.19.128.69) | 17:39 | |
| - mjw (QUIT: Remote host closed the connection) (~mjw@gnu.wildebeest.org) | 17:43 | |
| josch | i'm happy that tool works for you -- it is the most fragile of all by far | 17:43 |
| vkoskiv | Okay, vkcube uses the gpu if I set PAN_I_WANT_A_BROKEN_VULKAN_DRIVER=1 | 17:54 |
| vkoskiv | josch: I did very carefully review the output and also save a copy to my desktop in case something did go wrong, but it went smoothly. | 17:57 |
| + libercv (~libercv@user/libercv) | 18:03 | |
| - siviq (QUIT: Quit: Client closed) (~siviq@user/siviq) | 18:08 | |
| + siviq (~siviq@user/siviq) | 18:08 | |
| - siviq (QUIT: Client Quit) (~siviq@user/siviq) | 18:12 | |
| + spew (~spew@user/spew) | 18:21 | |
| - voltaire28_ (QUIT: Ping timeout: 264 seconds) (~jlafon@28.162.2.93.rev.sfr.net) | 18:42 | |
| - libercv (QUIT: Quit: Konversation terminated!) (~libercv@user/libercv) | 18:44 | |
| + libercv (~libercv@user/libercv) | 18:49 | |
| + Chwoka (~Chwoka@216.73.127.21) | 18:53 | |
| - potash1 (QUIT: Quit: The Lounge - https://thelounge.chat) (~potash@user/foghorn) | 19:07 | |
| - libercv (QUIT: Quit: Konversation terminated!) (~libercv@user/libercv) | 19:37 | |
| + libercv (~libercv@user/libercv) | 19:38 | |
| - libercv (QUIT: Quit: Konversation terminated!) (~libercv@user/libercv) | 19:49 | |
| + liberto (~Liberto@user/libercv) | 19:54 | |
| vkoskiv | Getting back to working on refmon (again), I think I figured out why the waybar battery gauge doesn't always match what the lpc gauge percentage says. Waybar's battery code doesn't account for /sys/class/power_supply/BAT0/charge_empty | 20:05 |
| vkoskiv | My refmon battery code didn't either until now, and now mine matches what the keyboard battery page reports exactly. | 20:05 |
| josch | my rk3588 classic reform rebooted out of nowhere while running an "apt full-upgrade" and having cura render something at the same time | 20:35 |
| josch | this is not the first time that rk3588 shuts off or reboots | 20:35 |
| josch | funnily i do not have these issues with my rk3588 pocket reform | 20:35 |
| + voltaire28_ (~jlafon@28.162.2.93.rev.sfr.net) | 20:37 | |
| - liberto (QUIT: Quit: WeeChat 4.8.1) (~Liberto@user/libercv) | 21:12 | |
| + mjw (~mjw@gnu.wildebeest.org) | 21:41 | |
| + siviq (~siviq@user/siviq) | 21:44 | |
| - angelwood (QUIT: Remote host closed the connection) (~angelwood@user/angelwood) | 22:21 | |
| ch | can someone enlighten me on the L1 calculation on the pocket charger board? MP2650 datasheet 'typical application' has it as 1.5uH, but on page 61 it gives 2.2uH for 20V Vin. tbh following the formulas i dont arrive at the values in the table | 22:21 |
| + angelwood (~angelwood@user/angelwood) | 22:22 | |
| minute | ch: usually there's a range of values that works, with different drawbacks | 22:24 |
| ch | minute: right. i'm mostly wondering how to check the calculations there, given the datasheet often seems to assume Vin=15V, but we can select up to 20V | 22:26 |
| - bkeys (QUIT: Quit: With every step we take, danger will follow closely) (~Thunderbi@98.19.128.69) | 22:34 | |
| + bkeys (~Thunderbi@98.19.128.69) | 22:35 | |
| - bkeys (QUIT: Client Quit) (~Thunderbi@98.19.128.69) | 22:39 | |
| + bkeys (~Thunderbi@h69.128.19.98.dynamic.ip.windstream.net) | 22:39 | |
| - bkeys (QUIT: Quit: With every step we take, danger will follow closely) (~Thunderbi@h69.128.19.98.dynamic.ip.windstream.net) | 22:51 | |
| + bkeys (~Thunderbi@98.19.128.69) | 22:51 | |
| ch | looks like gnome crashed sometime when charging to 99% | 22:57 |
| ch | weird | 22:58 |
| ch | restarting gdm (from ssh) worked | 22:58 |
| - siviq (QUIT: Quit: Client closed) (~siviq@user/siviq) | 23:20 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
