>2020+5
>Desktops still not ARM

I blame video games

>>105934417 (OP)
> 2020+5
> computer illiterate street shitting child rapists shilling ishit can't get more than 7% of industry
> why doesn't everyone else just start shitting in the street and use ARM, a CPU for mobile phones?
get back into your oil drum and row back up your designated shitting street and stay there, ramapoo.

>>105934417 (OP)
Good.
>x86 - one driver per component
>ARM - one driver (device tree) per device

>>105934546
What an edgelord

https://www.youtube.com/watch?v=AshDjtlV6go

The performance isn't exactly there yet but one day, OP.

>>105934417 (OP)
Why is everyone pushing laptop and mobile CPUs for desktop lately? I mean sure if you're just gonna browse JewTube it's fine but just get a laptop at that point.

>>105934925
My phone runs genshin better than my pc

>>105934417 (OP)
Use case?

x64 just works

>>105934940
I'm sorry to hear that

>>105934968
while consuming 10 times less power

>>105934983
Yeah, mobile phone CPUs uses less power, very impressive of you to notice. It doesn't mean much though since when you get down to performance per watt x86 is still on top

>>105935058
Apple Sillicon blocks your path

>>105934417 (OP)
>buy x86 PC
>can still install the latest software and OS 10+ years later

>buy ARM PC
>manufacturer drops support after 3 years
>updates stop working after 5 more years

>>105935087
cope

>>105934417 (OP)
I'd rather have RISC

>>105934556
What does that mean?
>>105934925
Performance per Watt.
The x86S idea by Intel wasn't bad per-say. They should leave one "legacy" core in place for all the old shit (the entire x86 schtick is that it is backwards compatible as fuck) and rice the fuck out of the rest.
Add a good helping of opening up x86S for licensing production like arm does and they would beat arm into the dirt.

>>105935082
M series chips are so good people aren't upgrading their macs. Apple won't let this keep going.

>>105934417 (OP)
Can't run gayms

>>105935186
>Even faster macs
Good lord, have mercy. But I don't think they'll be able to keep pace, the industry will move to a two-year release cycle. Even with phones. I can see small additions being made like new colors each year, but other than that I cannot see how we're going to keep pace.

Or AI will boost moore's law.

>>105935082
They aren't making anything serious so I'm not sure how to compare it against anything in the areas people actually care about performance per watt.
>>105935163
Yeah I get that people are meming about performance per watt but why? I'll never take a weaker CPU cause it's more energy efficient, power is cheap so who cares? Especially meme tier stuff like ARM that is very efficient when it's for a phone but as soon as you want to build real systems with it, it's fucking DOA trash.
And yeah it was a good idea but I think we're stuck with the current X86 until we move on to something that's actually competitive.

>>105935204
>CAD, Video, Rendering, Special effects
>Basically the working man's computer of choice

>>105934546
Bold

>>105935163
x86S reduces engineering burden of maintaining ancient poorly tested features but it would have very minimal effect on power usage. All x86S does is remove legacy mode thus having the machine boot directly into long mode. One might think this would remove all the "legacy cruft" but this is not the case, the vast majority of the legacy stuff is still valid in long mode.

>>105935204
>Especially meme tier stuff like ARM that is very efficient when it's for a phone but as soon as you want to build real systems with it, it's fucking DOA trash.
So meme, that Apple M chips are beating the crap out of competing x86 chips? Efficiency is everything the current landscape. If you only need 10 watts to perform the same FLOPs as your competitor does with 65 watts, you can turn a knob and go 65 watts to fuck your competitor so hard in the ass, their noses start bleeding uncontrollably. That 's the reason why Apple keeps jumping from ISA to ISA (Power -> x86 -> Arm, maybe RISC-V when it is ready next?) without a single hint of fear of breaking old tools and investing into tools like Rosetta to port those other old tools.
The same dumb reasoning
>hurr what do I need efficiency for, I've got POWAH
is why Apple pulls those insane numbers and there's nothing x86 can do to match it in PRICE / PERF OR PERF / WATTS. Sure you can pull out a fucking EPYC or Threadripper to outperform it, but you'll pay $10k for it and it will guzzle 300 watts. That's just the CPU, no system.
And what's to say Apple couldn't decide to actually go into the server space to get some of that sweet sweet AI money on a whim? For now, the fact that they stopped caring about servers a decade ago. But I know nothing stops nvidia doing arm on the cloud and nvidia does take efficiency VERY seriously for AI (gaming is a meme for them).

Look, I'm a Linux x86 4evaR guy (pic related, that's my neofetch) myself, but at this rate x86 will become a relic FAST, even EPYC is not going to keep x86 alive forever.

>>105935326
The reason why arm chips are capable of being more efficient than x86 is because arm has fixed width instructions. If x86S could ditch variable length instructions, they would gain a lot of silicon real-estate. Make the legacy x86 a co-processor that intercepts the variable length shit, and you're set. You wouldn't even need that co-processor on mobile applications like

>>105935378
>cont.
phones because there is literally no legacy shit for phones that could even run on x86. I can see the case for a new ISA that would be even more efficient than arm for phones.

But with RISC-V on the horizon, I don't see it on mobile to be entirely honest.

>>105935378
The cost of the x86 decoder has been way overblown. It might have been a significant portion of the chip once, but these days with 256 bit AVX registers and hundreds of hidden rename registers the decoder is only a small part of the cost now.

In regards to x86S being a different ISA. That would not in anyway be x86 anymore, you have just invented a new CPU. Also, its not possible to create a co-processor to implement decoding, a decoder is deeply integrated into a cpu it can't realistically be a separate thing.

>>105935283
Ah, you're one of those "working" people who needs a "computer"
>>105935378
I'm not saying X86 is gonna be the thing forever, but right now there is no competition if you want real compute power. And yeah it's super cool that Apple made a good laptop chip, but since you have to use their OS on it or the most ghetto rigged Linux distro in existence it's kind of a non-choice for anything outside of what artfags need. And if Apple goes into servers I'll probably kill myself, imagine having to use and manage their walled garden as your actual infrastructure, you know they'll do it. Maybe in 5-10 years ARM will be useful but right now it's just laptop CPUs and I have no need for that.
I'd much rather see RISC-V be the next thing but it's not jew-compatible so that'll probably not happen.

>>105935087
macos was already doing this with intel macs

>>105935508
https://www.nvidia.com/de-de/data-center/grace-cpu/

>>105934417 (OP)
Any gsind to hardware will be lost because because the implementations are all different and require dynamic recompilers or comparability layers.

But they were.

>>105935525
>German
Go away
Unreleased product and 2 years delayed

>>105935538
And they are.

ARMTODDLERS BTFO

>>105936068
Enjoy having aids

>>105934417 (OP)
But they are thoughever?

>>105935577
I haven't touched RPi's in about a decade, how viable are they as desktop daily drivers these days? I remember some time ago someone told me that RPi emulation boxes suck because there's too much input lag for anything that isn't a jarpig, but that was before covid.

I'd definitely plug in my own keyboard and use my own case though. I've seen a joke build before where someone put a Raspberry Pi into a tower case, I guess that's what most people envision when they think of widespread ARM adoption (they don't know modern crapple chips are ARM and are used in their desktops)

>>105934417 (OP)
ARM can suck my big, paki cock

>>105936850
Better off spending the price of a new rpi on a surplus sff box on ebay. It'll be vastly more powerful and responsive. Assuming you want it for desktop stuff and aren't using the GPIO.

I honestly don't get the hard on for ARM
What's going on in your life where ARM would significantly improve it?
I have zero ARM devices in my house besides my phone. PC, router, and home server are all x86.

ARM in its implementations just remind me of a future that isn't very good.

>>105936731
>>105936744
>>105936763
>>105935082
>>105935186
get a toilet currynigger

>>105936763
how much cheese can you grate with that?

>>105937030
a lot