ChanServ changed the topic of #asahi to: Asahi Linux: porting Linux to Apple Silicon macs | "Does XXX work yet?": https://alx.sh/fs | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Topics: #asahi-dev #asahi-re #asahi-gpu #asahi-alt #asahi-stream #asahi-offtopic | Keep things on topic | Logs: https://alx.sh/l/asahi
hdbngr has joined #asahi
omidmash2 has joined #asahi
hdbngr has quit [Ping timeout: 480 seconds]
omidmash has quit [Ping timeout: 480 seconds]
omidmash2 is now known as omidmash
XeR has quit []
jnn has joined #asahi
mergen has quit [Quit: Leaving]
hdbngr has joined #asahi
jn has quit [Ping timeout: 480 seconds]
hdbngr has quit [Ping timeout: 480 seconds]
XeR has joined #asahi
hdbngr has joined #asahi
hdbngr has quit [Ping timeout: 480 seconds]
hdbngr has joined #asahi
EtherNet642495367544181766 has joined #asahi
EtherNet64249536754418176 has quit [Ping timeout: 480 seconds]
hdbngr has quit [Ping timeout: 480 seconds]
hexdump01 has joined #asahi
hexdump0815 has quit [Ping timeout: 480 seconds]
tobhe_ has joined #asahi
tobhe has quit [Ping timeout: 480 seconds]
cylm has quit [Quit: WeeChat 4.7.0]
skipwich has quit [Remote host closed the connection]
skipwich has joined #asahi
ipatch_ has joined #asahi
ipatch has quit [Read error: Connection reset by peer]
hdbngr has joined #asahi
opticron has quit [Read error: Connection reset by peer]
opticron has joined #asahi
hdbngr has quit [Ping timeout: 480 seconds]
Pokemod97736 has quit []
KxCORP5894000345588131820 has quit [Quit: Bye!]
KxCORP5894000345588131820 has joined #asahi
Pokemod97736 has joined #asahi
john-cabaj has quit [Ping timeout: 480 seconds]
hdbngr has joined #asahi
hdbngr has quit [Ping timeout: 480 seconds]
hdbngr has joined #asahi
alarumbe has quit []
hdbngr has quit [Ping timeout: 480 seconds]
ddxtanx has joined #asahi
hdbngr has joined #asahi
hdbngr has quit [Ping timeout: 480 seconds]
hays has quit [Quit: Ping timeout (120 seconds)]
hdbngr has joined #asahi
hays has joined #asahi
hdbngr has quit [Ping timeout: 480 seconds]
ravikant_ has joined #asahi
FippleStone has quit [Remote host closed the connection]
daneelrolivaw72 has quit []
daneelrolivaw72 has joined #asahi
rizztty has quit [Read error: Connection reset by peer]
rizztty has joined #asahi
tertu2 has joined #asahi
tertu has quit [Read error: Network is unreachable]
hdbngr has joined #asahi
chadmed has quit [Quit: Konversation terminated!]
chadmed has joined #asahi
hdbngr has quit [Ping timeout: 480 seconds]
hdbngr has joined #asahi
hdbngr has quit [Ping timeout: 480 seconds]
hdbngr has joined #asahi
hdbngr_ has joined #asahi
snklp has joined #asahi
hdbngr has quit [Ping timeout: 480 seconds]
snklp has quit [Ping timeout: 480 seconds]
snklp has joined #asahi
anticataclyst has quit [Ping timeout: 480 seconds]
glem8100548899408 has quit [Ping timeout: 480 seconds]
glem81005488994086 is now known as glem8100548899408
kdb42490352912 has quit [Ping timeout: 480 seconds]
kdb42490352912 has joined #asahi
figment has quit [Ping timeout: 480 seconds]
maettu102 has joined #asahi
figment has joined #asahi
m0bius__ has joined #asahi
m0bius_ has quit [Ping timeout: 480 seconds]
mergen has joined #asahi
ansg191 has quit [Remote host closed the connection]
ansg191 has joined #asahi
<dualecton5>
Hi people, I kinda want to buy my first macbook but I hate macOS, I was wondering how much of an headache is linux on mac
<twb>
Kind of depends on what kinds of things annoy you
<dualecton5>
I like to configure my own stuff but if I have to do it for every small thing... yikes
<j`ey>
configuring stuff is the same as non-mac
<twb>
If you're prepared to trust curl|sh then for the supported systems, it Just Worked for me
<twb>
It's not as upstreamed and polished as a thinkpad T series maybe, but it's definitely no worse than a random android handset
<dualecton5>
mh better than i thought, what about the boot loader
<twb>
If you don't care about measured boot, the bootloader stack is FineTM in my personal experience
<dualecton5>
battery life?
<dualecton5>
for basic developer stuff
<twb>
well on my t490s I measure 20 hours idle and 10 hours while active, and on my macbook the same test measured 64 hours idle and 24 hours active
<twb>
But I have not deliberately tried to run it down while doing constant compiles and timing it against the wall clock
<twb>
macOS repored "20" hours regardless because it seems to always report min(20 hours, actual time left)
<twb>
intel powertop works but its reports are less comprehensive than on a thinkpad, e.g. on my thinkpad I can see that like 5W of my 6W total usage is the f***ing radio
<chaos_princess>
m1pro uses approx 40W running an all-core load, 10-ish idle with screen on and high brightness, and approx 3.8W at screen-off idle
<dualecton5>
lmao i see
<dualecton5>
ty
<chaos_princess>
also, the sleep power use is bad on pro/max, it _will_ drain overnight
<twb>
chaos_princess: thanks. That roughly matches my synthetic measurements, being maybe 20% more efficient at idling than my thinkpad
<chaos_princess>
(in linux that is, in macos you can leave it for like a week depending on how much cycles are on your battery)
<twb>
From talking to #macosx, I think a large part of the macos-on-macbook battery hours improvement is just that macos by default aggressively sleeps
<chaos_princess>
hehe
<chaos_princess>
the macos power usage "improvement" is because it actually drives the hw correctly
<twb>
Apparently suspend turns off more components than on an intel system, which is why they hard-code the regime as "suspend for 24h then hibernate" rather than the usual 1h
<twb>
I also found some docs saying it'll wake itself up and like go "hey any push notifications?" then go back to sleep, but I didn't see a lot of detail
<chaos_princess>
this is just normal wake-on-rtc, x86 can do that too
<twb>
Fair
<twb>
I think that was less about saving battery and more about "oh I just opened it and already it has all my mail and photos synced"
<chaos_princess>
pretty much. or, you know, "oh, i just opened and it has cooked itself and drained the battery trying to install windows updates while in a backpack"
<twb>
dualecton5: note that M3/M4 aren't usable at all right now, so make sure you buy an older model
<dualecton5>
If i buy one I'll go with m2 but I don't think so
<twb>
chaos_princess: haha my last armbook I literally never turned it off -- I left it doing TTS inside the laptop bag, with headphones while cycling home in pouring rain and 40°C hot days
<twb>
surface is the worst. What kind of tablet has a f*ing fan
<twb>
And is like 25mm thick
<dualecton5>
ahahahaha yeah
<chaos_princess>
is a fan in a tablet really that bad? And there are thee *s in "fuck"
<chaos_princess>
*three
<dualecton5>
I heard the laptop ones have the worst battery ever
<twb>
chaos_princess: well in my dad's one it was super annoying. Partly because it's a little fan so it's very noisy
<twb>
Not as noisy as 1U rackmount fans I admit
<chaos_princess>
right, fair
<twb>
I will accept having a fan in the system if it's only used when the ambient temperature is >35°C and it's doing fan instead of just slowing the CPU down
<twb>
But it shouldn't be whirring when you're just writing a word document
<twb>
in an air-conditioned room
<tpw_rules>
the only time the fan in my m1 max starts is several minutes into compiling a kernel
<twb>
tpw_rules: yeah that's pretty much how my thinkpad works as well
<tpw_rules>
oh the one in my thinkpad drives me bananas, it will do it if i scroll fast
<tpw_rules>
and it has a particularly aggravating ramp up
<tpw_rules>
but it's an older workstation model so
<twb>
I mostly prefer "slow the system down" method though, because fewer moving parts and dirt ingress, and if you need that much compute power maybe the job should be offloaded to a rack with noise and heat abatment
<chaos_princess>
i might have thought that way too before getting this m1pro
<chaos_princess>
runs amazingly fast for a laptop
<chaos_princess>
still kinda can't believe that it is that fast despite owning it for however long
<sven>
I feel the same about a base(!) m1 MacBook Air
<twb>
I mean I was running on EeePCs and chromebooks until 2020 and I only upgraded to a "real" computer when bloody websites started needing >2GB of system RAM to work at all
<sven>
and it’s almost 5 years old now
<chaos_princess>
yea, i expect at least 7-8 years of service out of mine
<sven>
yeah. and hopefully by then we have m3 working :D
<twb>
I do do compiles on my t490s now, but I feel bad about it
<sven>
my battery is slowly starting to show its age
<sven>
but that's hardly surprising after ~5 years
<twb>
like all those people who have a private motorcar, so the car has to be built and uses all those resources, but is actually only used for 1 pax for like 1 hour a day, and the rest of the time it's also consuming land resources in a parking lot
<twb>
Every server work owns is >12 years old and is less powerful than my laptop :-(
<twb>
which is uh... 6 years old now
<chaos_princess>
Something tells me that if i put my phone into a freezer, it would probably also be faster than quite a few servers
<sven>
lol
<twb>
I read somewhere phones run faster if you hold them upright
<twb>
because if they're lying on the table they can't dump heat into the table as fast as they could into the air
<chaos_princess>
what
<chaos_princess>
air is an insulator?
<twb>
but convection
<chaos_princess>
i would really want to see the measurement setup, not immediately calling bs, but it feels unintuitive
<twb>
I don't have a citation so I'm happy to just retract
<tpw_rules>
there's a great book on heatsinks which somewhat supports that
<tpw_rules>
especially with today's phones that can't lie flat on a table anymore
<tpw_rules>
(hot air rises and heat sinks)
anticataclyst has quit [Remote host closed the connection]
<twb>
I did like that peltier rig intel trialled for a couple of years, where the sensors reported all cores running at 0°C because they couldn't report negative numbers
<twb>
made me want to go back and see the meeting with the guys designing the sensor spec
<twb>
"should we use kelvin" "no what cpu would ever run below 15°C even in anarctica"
<opticron>
chaos_princess, I have a MBP M1 Max and it is by far the fastest machine I own for my regular workload of compiling GCC, GDB, support libraries, and newlib for the target platform
<opticron>
7 minutes for a full fresh build with hot caches where the workhorse dual xeon at work takes 22 minutes for the same compile and even my gaming desktop takes 13 minutes
<opticron>
I suspect that the memory bandwidth is critical for that task
<opticron>
though I'd be interested in someone running the test for a M1Pro
<opticron>
just to see if it differs
<chaos_princess>
Same-ish roughly
<chaos_princess>
Memory bandwidth doesnt matter _that_ much
<chaos_princess>
As in, you can't use all of max memory bw with only cpu
rvalue- has joined #asahi
<twb>
opticron: are you cross-compiling in any/all of those scenarios?
<chaos_princess>
It shouldnt matter much
<chaos_princess>
Except for gcc, but you just do --diable-bootstrap or w/e
rvalue has quit [Ping timeout: 480 seconds]
bgb has quit [Remote host closed the connection]
bgb has joined #asahi
bgb has quit [Remote host closed the connection]
<opticron>
twb, just newlib with the built cross-compiler
bgb has joined #asahi
<opticron>
but yes, the built compiler is always run natively
<twb>
I guess unless there are unit tests it's never *executing* the cross-compiled code under qemu-user so it doesn't matter
<opticron>
correct
<opticron>
all unit tests occur in a different step elsewhere
<twb>
I'm usually doing mmdebstrap so cross-arch is actually running all the post-install scripts through CPU emulation
<twb>
I guess the other obvious(?) guess is whethere MAKEFLAGS=j$(nproc) is similar
<twb>
I remember compiling busybox with -j (i.e. infinitely parallel) would effectively forkbomb myself
<twb>
What's the sensible approach when you have AMP instead of SMP? Do you just hard-code MAKEFLAGS=j4 when you have 4 performance cores and 4 mobility cores?
<opticron>
(for reference, my build test is the rtems-source-builder aarch64 toolchain build)
<opticron>
I think the builder I'm using just uses all cores
<twb>
I wonder if that's optimal
<tpw_rules>
what do you use rtems for?
<opticron>
I work at OAR, so I port it to new boards/platforms/write drivers/etc.
<opticron>
I also do a bit on the OSS side of things for community management and general project steering
<chaos_princess>
twb: why not? Let e-cores help out, once p-cores are out of work, the workload should migrate
Moprius has joined #asahi
Moprius has quit [Remote host closed the connection]
fwg has quit [Ping timeout: 480 seconds]
<twb>
chaos_princess: I guess I naively thought they'd be bad at that job?
<twb>
Like it'd be more efficient mips-per-watt overall to do 4 and force them to stay only on the performance cores
<chaos_princess>
E-cores are also quite good, it isn't some atom trashheap
<twb>
Fair enough I guess
<opticron>
it's not like the e-cores are equivalent to a stm32 :)
<chaos_princess>
(i should check out what rtems even is some day)
<twb>
I've never really looked into how M and A cores are different. I assume it's like "the A cores have additional branch prediction vulnerabilities ^W^W^W performance optimizations"
<chaos_princess>
Wdym by M and A cores?
<opticron>
Cortex-M* vs Cortex-A*
<twb>
aren't the apple arm cores basically like co--- that'
<chaos_princess>
No
<twb>
ohhh
<chaos_princess>
They would both be A cores in arm-speak
<chaos_princess>
P-cores are just wider, clock higher, and have more cache
<j`ey>
M are microcontrollers style chips
<twb>
oh OK
<twb>
I thought they were much more differenter on macbooks
<chaos_princess>
And arm doesnt mix a and m either, their big.little setup is also A-cores, but 2-3 variations thereof
<opticron>
same idea as big.LITTLE on most mobile/phone CPUs
<opticron>
I think they've made that more flexible and renamed it dynamIQ
<twb>
Clearly I know fuck-all about this stuff :-(
<opticron>
chaos_princess, short summary, RTEMS is an alternative to ThreadX/FreeRTOS/Zephyr/VxWorks/etc.
<chaos_princess>
oh, i know that part. It is just that i learned of it's existence from that libogc drama and it is still in an open tab somewhere :P
<opticron>
oh, hah
rivendell has joined #asahi
rivendell has quit [Ping timeout: 480 seconds]
rivendell has joined #asahi
bgb_ has joined #asahi
bgb has quit [Remote host closed the connection]
<maettu102>
I added swiss mac keyboard layout to kbd few days ago :D
hdbngr_ has quit [Ping timeout: 480 seconds]
hdbngr has joined #asahi
hdbngr has quit [Ping timeout: 480 seconds]
john-cabaj has quit [Ping timeout: 480 seconds]
rivendell has quit [Remote host closed the connection]