robclark changed the topic of #aarch64-laptops to: Linux support for AArch64 Laptops (Chrome OS Trogdor Devices - Asus NovaGo TP370QL - HP Envy x2 - Lenovo Mixx 630 - Lenovo Yoga C630 - Lenovo ThinkPad X13s - and various other snapdragon laptops) - https://oftc.irclog.whitequark.org/aarch64-laptops
alpernebbi has quit [Ping timeout: 480 seconds]
alpernebbi has joined #aarch64-laptops
<HdkR>
robclark: Some minor improvements there. Benchmarks only show up to a 60% perf improvement but some tighter loops should be closer to 2x :P
<robclark>
"only show up to a 60% perf improvement"
<robclark>
I'd be pretty happy for 60% improvement ;-)
<HdkR>
It'll be pretty niceIt's pretty nice when testing games
<HdkR>
It's a shame X1C is a weird inbetween core that has performance problems with unaligned stores. It's likely hitting us hard
<HdkR>
sm8650 also has flagm2 and afp extensions which gives another free performance boost :D
<robclark>
I'm not sure how much we know about the nuvia cores, but presumably/hopefully qc has been listening to requests from however works on the win x86 emu stuff
<robclark>
since they probably need/want all the same things that fex does
<HdkR>
We can only hope that Microsoft is screaming from the rooftops to give us a TSO memory model enable bit like Apple hardware.
<HdkR>
Since thus far the LRCPC1/2/3/4/5/6 extensions have proven worthless
<HdkR>
Although a 4.3Ghz ARM core that claims to be as fast as Apple's M2 is already a big win.
<HdkR>
Enabling TSO emulation and it dropping to 1/8th the perf like a Cortex wouldn't be nice :P
<robclark>
yeah, I think that is the rub with tso
<steev>
ohhhh that x1c
<steev>
i'm used to people calling the Thinkpad X1 Carbon, an x1c
<robclark>
yeah, cortex-x1c
hexdump01 has joined #aarch64-laptops
hexdump0815 has quit [Ping timeout: 480 seconds]
<steev>
hmm
<steev>
actually, i'm not sure that mutter patch is fully correct
todi1 has joined #aarch64-laptops
todi has quit [Ping timeout: 480 seconds]
newusr has joined #aarch64-laptops
newusr has quit []
<travmurav[m]>
steev: do you have some testcase where it fails?
<jhovold>
clover[m], neobrain: youtube playback works fine here with arch and firefox 120.1
<jhovold>
could pipwire be involved? i'm using pulseaudio here
<jhovold>
I assume at least clover is using steeve
<jhovold>
...steev's branches and the laptop_defconfig (just pointing out differences in setup)
<jhovold>
and I run xorg, not wayland
<jhovold>
also does it depend on the content? do you have video acceleration enabled?
<Jasper[m]>
jhovold: This was a problem before it was enabled in ff
<jhovold>
ok, thanks, one thing to rule out
<Jasper[m]>
Could be pipewire, I was using that when I had the problem
<jhovold>
seemed like a possibility since you said it went away at some point
<neobrain>
I'm using pipewire indeed
<jhovold>
neobrain: perhaps you can try switching to puleaudio and see if that helps
<neobrain>
On it :)
<jhovold>
:)
<neobrain>
Not an issue with pulseaudio-proper indeed, interesting
<neobrain>
Volume is still variable, but I'm assuming that's the speaker firmware regulating heat or something
<jhovold>
neobrain: thanks for confirming, sounds like someone should report this to the pipewire devs
<jhovold>
haven't really noticed the volume changing here, are you sure you're using the latest ucm files etc?
<neobrain>
ah, the ucm files may have been overwritten by system update meanwhile
<jhovold>
yeah, it's really unfortunate that those hasn't been released by upstream yet
<jhovold>
i've built and installed that package on arch myself (after similarly having my local hacks overwritten at some point)
<neobrain>
Where can I find the ucm files again? I always forgot how to find that wiki page that mentions all this stuff D:
<neobrain>
I think my ucm files are up to date, but actually turns out the issue happens to be noticable in particular past 80% volume and in the one specific song I first tried, heh
<jhovold>
no, sorry, i really can't stand hip hop, i'll just have to take your word for it :)
<Jasper[m]>
That hurts :(
<Jasper[m]>
Subgenre's of Hip-Hop can be stellar
<Jasper[m]>
(I haven't listened to what neobrain shared so idk about that :^))
<jhovold>
heh, maybe something really old like public enemy or even gil scott-heron, but not that gangsta stuff
Caterpillar has quit [Quit: Konversation terminated!]
<Jasper[m]>
Kendrick Lamar is a good lead-in I'd say. Conscious hiphop is great. If you like more funky/jazzy instrumentals I'd recommend To Pimp A Butterfly, one of the best albums ever made (I'm genuinely not pulling this out of my ass)
<Jasper[m]>
Otherwise A Tribe Called Quest is also very solid, they peaked a while ago, but it's more in line with Public Enemy. Fat chance they sampled Gil Scott-Heron at some point :^)
<jhovold>
i'll try to give it a listen
<neobrain>
haha I considered adding a disclaimer about the genre
<_[m]1>
<HdkR> "Although a 4.3Ghz ARM core..." <- thisis what you want right? it's not some product launch I missed?
<HdkR>
_[m]1: sc8380xp can boost two of its CPU cores to 4.3Ghz
<_[m]1>
ha yes those oryon core, hot
<HdkR>
M3 Max can boost its CPU cores to 4Ghz, so not too far off :)
martiert has quit [Quit: WeeChat 4.1.1]
martiert has joined #aarch64-laptops
<_[m]1>
still to be seen the adoption
mynery[m] has joined #aarch64-laptops
<mynery[m]>
okay, so x13s here. i have an issue where if i do add an external display, everything works great, but if i close the laptop and open it later, the laptop screen just stays off. sadly, i have no clue where i could find relative information. i did found some issues with screens staying black after suspend, but this isn't really the case.
<mynery[m]>
however, the suspend thing can apparently be a hardware issue, so i just want to ask if anybody is aware whether this is one in that case or not
<jhovold>
mynery[m]: known issue, the display controller fails to resume properly if you have an external display connected
<jhovold>
if you're in a VT console you can just disconnect and reconnect the external display to get the internal display back
<jhovold>
with x/wayland you'd need to some more, but you should be able to log into the machine remotely and restart them
<mynery[m]>
huh, i could try switching to a tty and then disconnect
<jhovold>
do you really want to suspend when closing the lid with external display connected? Otherwise, i think there's some systemd option to allow you to just shut the lid without susending in that case
<mynery[m]>
i'm not sure what the results of that would be. like, the device itself doesn't go to suspend obviously, but if i don't make the display suspend (if this is an option), will gnome still figure out that it wants to move the windows to the external screen?
<jhovold>
if you don't suspend, everything should work as with the lid open, except you won't see any windows on that screen, not sure if the backlight would still be on though...
<jhovold>
mynery[m]: in case you need it: logind.conf, #HandleLidSwitchDocked=ignore
<jhovold>
seems to be the default here
janrinze has joined #aarch64-laptops
<clover[m]>
mynery: you also need a patched version of mutter or you will get issues with gnome and external displays
<mynery[m]>
i read about that on the ironrobin repo, but i have honestly no idea what the described issue even is
<steev>
travmurav[m]: i say that mostly because it seems to not recognize when i disconnect my adapter anymore
<steev>
previously if i would disconnect my adapter, all apps that were on the external screen would move to the primary display, now, they just sit off in nowwhere land
<travmurav[m]>
oh, interesting, let me see if this happens for me too...
<travmurav[m]>
steev: But no, seems like I can't reproduce this - just opened a terminal window and moved it to an external display: it moves back to the laptop when I unplug the display and returns to the old place when I reconnect it
<strongtz[m]>
I wonder if sc8380xp (or whatever it's called) supports SVE2
<steev>
travmurav[m]: dumb question but... using wayland?
<travmurav[m]>
steev: yep
<travmurav[m]>
I actually think x11 is completely broken for me, or at least it was at some point
Caterpillar has joined #aarch64-laptops
<konradybcio>
plasma wayland works just fine fwiw
nerdboy_ has joined #aarch64-laptops
nerdboy_p has quit [Ping timeout: 480 seconds]
Caterpillar has quit [Ping timeout: 480 seconds]
<_[m]1>
<konradybcio> "plasma wayland works just fine..." <- not on arch for me, on debian yes
<steev>
for me, unplug, suspend, bring back from suspend, and it's good, however... i plug the external display back in and the monitor never seems to get a signal... unless i log out of gnome, switch to sway, and then log out of sway and back in to gnome, it works
<steev>
and it's not specific to that hdmi dongle
<steev>
but i can move the mouse to the display as if it's there, and can see it in gnome settings, just no signal
<_[m]1>
hm but that sucks if you have external storage connected through the same dock the hdmi is connected
<_[m]1>
that's odd behavior though
<steev>
yeah, i just umount it beforehand
Caterpillar has joined #aarch64-laptops
Caterpillar has quit [Ping timeout: 480 seconds]
<craftyguy>
steev: which system? I see something similar to that on the x13s, but when I suspend with an external display then on resume it isn't detected unless I replug the cable (and sometimes only after rotating the connector)
<steev>
gnome 45, with patched mutter
<craftyguy>
ah there is something in dmesg, not much... "dwc3-qcom a4f8800.usb: HS-PHY not in L2"
<steev>
yours is a known issue
<craftyguy>
ah, is there more info about it? just for my own curiosity
<steev>
jhovold replied about it above to mynery[m]
<craftyguy>
oops, I'll scroll up
<mynery[m]>
is there an aur package with that patch?
<craftyguy>
jhovold: FWIW I often suspend with the lid closed and an external display connected, to reduce power usage when I'm not using the system
<craftyguy>
no clue, I use sway/wlroots and things work well except for that resume from suspend thing
<steev>
the patched mutter? i do not know, you'd have to ask clover[m], the resident arch expert ;)