Dasnap@lemmy.world
on 16 Oct 2024 10:20
nextcollapse
Really they need to work on power usage and temperature of x86 so the chips are easier to use in mobile devices without a fan and dying in 3 hours. Stationary devices seem to be chugging along with x86 comfortably, but the chips are currently impractical otherwise.
leisesprecher@feddit.org
on 16 Oct 2024 10:28
nextcollapse
The new Intel chips already addressed that, at least for notebook class devices.
Realistically, there wasn’t really a reason for Intel and AMD to be super power efficient, simply because there wasn’t any competition for quite a while. It took Apple Silicon to show how powerful arm can be and how easy the transition could be.
InverseParallax@lemmy.world
on 16 Oct 2024 11:17
collapse
Apple took all the old tricks Intel was always way too cheap to use, and turned them to 11.
Nothing magic, nothing special, just balls and the willingness to spend silicon.
frezik@midwest.social
on 16 Oct 2024 10:44
nextcollapse
You’re not going to see phones with x86. The architecture just isn’t going to scale down like that. Not if you want something faster than a Pentium III.
InverseParallax@lemmy.world
on 16 Oct 2024 11:15
nextcollapse
It actually can, the thing we learned is that the unpleasant bits of x86 scale well, so we spent 30% of the die doing uop decode, but that’s now just 1-2% because we blow so much on more registers and cache.
Also we can play games like soft-deprecating instructions and features so they exist, but are stupid slow in microcode.
We used to think only risc could run fast at low power, but our current cisc decoded to risc works fine, Intel just got stupid lazy.
Apple just took all the tradeoffs Intel was too cheap to spend silicon on and turned them to 11, we could have had them before but all the arm guys were basically buying ip and didn’t invest in physically optimized designs, but now that tsmc is the main game in town (fallback to gf was nice for price), there’s a lot more room to rely on their cell libraries.
Intel got so insanely arrogant, just like Boeing and all the other catastrophic American failures right now, we just need to correct for that and we can be decent again.
frezik@midwest.social
on 16 Oct 2024 12:01
collapse
It’s hardly just Intel. There are two other x86 licenses out there. One gave up. The other is kicking ass, but Apple didn’t go with them, either.
Meanwhile, Intel themselves kept the 80486 alive until 2007 as an embedded processor. It outlasted the Pentium III by a few months. It was never as popular as PIC or ARM or z80 devices, but it found some kind of niche.
I’ll grant that in theory, it could be done. But why? There are millions of smartphones running fine with ARM, and they don’t have any backwards compatibility needs to x86. Why pick an ISA that can only legally be designed by three companies? Why pick an ISA that hasn’t been as well tested on mobile device OSes? ARM will hand a license to anyone who shows up with some cash, and if you want to take a plunge into a different ISA, then RISC-V is sitting right there. There doesn’t seem to be a single real benefit to x86 over what mobile device creators have now, and plenty of reasons not to.
InverseParallax@lemmy.world
on 16 Oct 2024 14:24
collapse
No, it doesn’t make sense to do it.
I worked on platform enablement for armv8, bringing all the ecosystem to 64 bit arm. Was an everest, so much code was expecting x86, lots of secret asm and other assumptions like memory model.
But once it was done, we did it again for riscv in no time, all the work was done, it was basically setting defines, maybe adding tsc/rdcycle (now rdtime).
Architectures don’t really matter anymore, but also the overhead of architectures are pretty minor, riscv will probably win because it’s basically free and single thread performance isn’t as critical on client devices, lot of work goes to the GPU too, and servers do other heavy lifting. Qualcomm scared everybody too, and China is going their own way which means even more riscv.
Basically, nothing matters except cost now, we’ll figure out how to run things on a potato, we’ve gotten good at it.
heartfelthumburger@sopuli.xyz
on 16 Oct 2024 11:22
nextcollapse
ISA doesn’t matter as much as most people think it does. It’s all about how you implement it.
frezik@midwest.social
on 16 Oct 2024 12:33
collapse
How do you hire people who can implement it right? There are three companies that can make x86. One is failing, one gave up years ago, and the third is kicking ass but seems uninterested in this part of the market. All the people who know how to do x86 well already work for one of them. That third company that nobody talks about gave up because by 2010, they lacked the ability to make a worthwhile product.
It’s an incredibly difficult ISA to work with, and all the talent is already busy. Due to its closed nature, there is little hope of significantly growing that talent base. Not unless you want the early 2000s version of x86-64, which is patent free.
They were comparable to the rest of the phones at the time. Not great, not terrible. Compared to anything in 2024 they were obviously trash, but that’s mostly because we’ve made 10 years of progress since then.
frezik@midwest.social
on 16 Oct 2024 19:07
collapse
The Samsung Galaxy is 15 years old, and it was excellent. First Android device I felt had decent performance. I think my first one was an Galaxy S III, which would have been 2012.
The Zenphone that’s out now uses a Snapdragon. There’s probably a good reason for that.
It seems that they’re finally taking that seriously though so it’s good to see. They never really had any incentive to put too much effort in making x86 more efficient for consumer devices since their server chips have much, much higher profit margins.
Lunar Lake and AMD’s Z1 is a good start and it’s interesting to see where this goes.
fuckwit_mcbumcrumble@lemmy.dbzer0.com
on 16 Oct 2024 13:01
collapse
It’s amazing what a modern process node and not cranking clock speeds to high hell will do.
Peter1986C@lemmings.world
on 16 Oct 2024 14:04
collapse
You forgot about ditching more of the chipset etc. in favour of integrating everything into the CPU die.
TheGrandNagus@lemmy.world
on 16 Oct 2024 10:48
nextcollapse
I’m conflicted on ARM.
The additional competition is great, but it presents a great risk of PCs becoming more locked down. They don’t have an open, standardised BIOS/UEFI like x86 systems do.
Booting alternate OSes on ARM systems can be a nightmare. Usually it’s straight up not possible.
I don’t want PCs to be like smartphones. I don’t want locked bootloaders.
EDIT:
FFS people. I know there are some ARM devices that allow booting of non-official OSes. That’s why I said usually.
Even for those devices though, they typically have to use non-standardised firmware (you can’t just take an OS for device A and use it for device B in the same way you can take an .iso and install it on any x86 machine), and it requires the OEM to want the device to be open.
Your desire to go “umm ackshully…” and be technically correct over a point I never made in the first place is blinding you to the point I was actually making: x86 is fairly open, standardised, and modular by default. ARM isn’t. And all it takes is a look at the phone/tablet market to see that OEMs don’t want them to be.
I worry, and I don’t think unreasonably, that ARM becoming the standard could mean a further erosion of the openness of PCs.
I mean, different use cases yeah? There’s certainly a big market for people that just do the basics on their devices, ie email, web browsing, documents/spreadsheets etc that don’t need a full blown powerhouse computer, nevermind that they have no clue what an operating system even is nor do they care, as long as they get their memes and cat videos in between work tasks.
I’ll bet there will always be an x86 segment of the market for gamers, power users, tinkerers, and the like. Though, that market may unfortunately shrink in the coming years that could lead to vendors abandoning the space, which could lead to fewer choices and higher component prices. On top of which, major venders might see it as an opportunity for lock-in and advertising, so yeah it’ll be interesting to see what happens.
The sliver of hope here is that the hacking community has always found ways around proprietary bullshit, and we can only hope and support that those efforts continue, lest we further our race in to a stupid corporate dystopia.
PetteriPano@lemmy.world
on 16 Oct 2024 11:52
nextcollapse
Standardized firmware isn’t something that’s specified in the ISA, is it? It’s just shitty phone manufacturers.
Asus had some x86 phones a few years back. I haven’t dug into them, but I doubt they had a full bios/efi.
pine64 arm devices have u-boot, while a bootloader does fullfil a subset of the uefi spec.
gravitas_deficiency@sh.itjust.works
on 16 Oct 2024 11:52
nextcollapse
Tbh I really want to get my hands on a snapdragon X laptop at some point just to play around with it. The energy efficiency alone makes me very curious.
I was under the impression that most of the issues around getting Linux to work on them was around driver support. As in: people are absolutely able to install an arbitrary OS, but the functionality is just super janky in most cases. Is that not accurate?
prosthetiknow@sh.itjust.works
on 16 Oct 2024 20:12
collapse
You’re definitely right in terms of arbitrary OS installation, some folks have got Ubuntu running on Lenovo snapdragon laptops recently.
The lack of “portability” though is a bit troubling, it seems each device (tree) has to be manually added, developed, tested, and have an install image created for it, unless I’m missing something. And this will be arduous and potentially problematic for corner cases or small numbers of adopters of a particular machine model (so basically the same as right now I guess).
As well, software packages have to be ported to the new architecture, which in some cases is easier said than done. Sure basic Unix utilities are portable enough, but more advanced and complicated software might have some issues, unless an efficient compatibility layer could be developed.
demesisx@infosec.pub
on 16 Oct 2024 12:14
nextcollapse
Don’t be conflicted. RISC-V or GTFO.
mesamunefire@lemmy.world
on 16 Oct 2024 13:14
nextcollapse
Yep it keeps getting faster and faster.
fuckwit_mcbumcrumble@lemmy.dbzer0.com
on 16 Oct 2024 14:38
collapse
“it keeps getting faster and faster” isn’t really saying much when it’s 1/10th the performance of a raspberry pi.
LedgeDrop@lemm.ee
on 16 Oct 2024 15:02
nextcollapse
Yeah, but RISC-V also costs 1/10th the price of a Pi.
fuckwit_mcbumcrumble@lemmy.dbzer0.com
on 16 Oct 2024 15:23
collapse
To license the arcitecture it costs a whole lot less, but when it comes to getting an actual usable computer they cost the same or more as an ARM machine, and perform worse.
Oh, I absolutely agree. Licensing is where the big difference is at, but that makes sense though, as ARM and RISC-V are both RISC based processors.
It’s loosely akin to comparing AMD vs Intel. Of course, you cannot pop-out an RISC-V and replace it with an ARM. However, the PCB’s should contain all the same parts, meaning they’ll have both have a similar price.
Unlike Intel/AMD, which you’d need extra capacitor, heat sinks, whatever - to help it handle all that extra power those CISC processors need (which results in heat).
mesamunefire@lemmy.world
on 16 Oct 2024 15:19
collapse
The pi foundation is starting to roll out chips with risc v processors built in. It’s extremely cheap and open spec.
One example is RP2350.
More options in computing is better for everyone.
ZILtoid1991@lemmy.world
on 17 Oct 2024 08:45
nextcollapse
Someone just needs to work on some open-source GPU for it, otherwise it’ll still have some of the usual shortcomings of many ARM SoCs.
tophneal@sh.itjust.works
on 16 Oct 2024 12:19
nextcollapse
Not all ARM chips are in phones, nor are they all locked down like one. There are several ARM devices and SBCs now where switching OSes is as easy as swapping out an SD card. Most do use uboot as a standard and some are even capable of utilizing UEFI.
BrianTheeBiscuiteer@lemmy.world
on 16 Oct 2024 13:56
nextcollapse
I was gonna say I thought Raspi was ARM.
BearOfaTime@lemm.ee
on 16 Oct 2024 13:58
nextcollapse
But it’s not standard.
What made PCs take off was the BIOS war, which occurred because manufacturers were dependent on 3rd party OS’s, which were still competing for dominance.
SynopsisTantilize@lemm.ee
on 17 Oct 2024 02:24
collapse
Where can I watch or read about this? Like what can I Google for more?
Peter1986C@lemmings.world
on 16 Oct 2024 13:59
collapse
Some SBCs only boot from said SD card though, while some do support more robust media. However, too many images are presuming you boot from SD which is a pita.
With or without Das Uboot, they still rely on board specific firmware (even Uboot is customised for many boards to make it work). OSes that state they do support aarch64, often require to have UEFI on your system so no way they are gonna boot on e.g. your Raspberry Pi.
Add to that, that is unlikely that browsers compiled for arm64 will have feature parity with their x86-64 counterparts. Goodbye Digitale Rights Management, and with that goodbye services like Tidal or Spotify (unless you run an OS that is still supported by their apps).
vext01@lemmy.sdf.org
on 16 Oct 2024 14:24
nextcollapse
This! Exactly!
todd_bonzalez@lemm.ee
on 16 Oct 2024 14:46
nextcollapse
Booting alternate OSes on ARM systems can be a nightmare. Usually it’s straight up not possible.
Raspberry Pi owners: 🙄
TheGrandNagus@lemmy.world
on 16 Oct 2024 18:01
collapse
“usually”
I am a raspberry pi owner (actually I’ve owned like a dozen and still own 8). And a pinephone owner.
And even on those very open devices, booting alternate OSes is less straightforward than on x86.
werefreeatlast@lemmy.world
on 16 Oct 2024 14:48
nextcollapse
As a simple user of Linux, I totally understand what it means for me…less choice, more google-android-like shit hardware. No thanks 👍.
LedgeDrop@lemm.ee
on 16 Oct 2024 15:01
nextcollapse
I don’t want PCs to be like smartphones. I don’t want locked bootloaders.
I’m sorry to burst your bubble, but since Microsoft made TPMmandatory for Windows 11+, locked down bootloader are on their way.
Basically, TPM allows (Windows) software to validate/verify the integrity of the OS and hardware. This also (could) include the bootloader/bios if Microsoft chooses to do so.
TPM is the equivalent of attestation on Android, which is the exact reason why your Banking App won’t work on your rooted/custom Android Phone.
That being said, we should embrace ARM. X86/AMD has 30+ years worth of “history” baked into each ( CISC) chip. This complexity is why your PC draws soooo much power and generates soooo much heat.
neo@lemmy.comfysnug.space
on 16 Oct 2024 18:21
nextcollapse
I would like RISC-V to also get ‘good enough’, considering that it seems to be a shitton more open than ARM is.
UEFI exists on arm and windows on arm devices can boot other OSes through it just like on x86.
stoly@lemmy.world
on 16 Oct 2024 14:49
nextcollapse
LOL this is the biggest sign out there that ARM is making a superior product. Once people start going protectionist, it’s time to move on to the new thing.
Treczoks@lemmy.world
on 16 Oct 2024 15:20
nextcollapse
This is a sign of ARM approaching the “enough” level. I remember the times when it was actually important to buy the latest PC at least every other year to have enough power to run a basic office suite or similar programs with acceptable speed.
Nowadays, you can staff offices with about any PC off the shelf - it is powerful and big enough to fulfill the needs of the majority of users. Of course there are servers, there are power users, engineers running simulations, and of course gamers who need more power, and who still fuel the cutting edge of PC building. But the masses don’t need to be cutting edge anymore. A rather basic machine is enough.
Here comes the ARM: For many years, ARM-based chips were used as SOCs, running anything from washing machines to mobile phones. But they have grown bigger and faster, and I can see them approaching the point that they can cover the basic needs of the average office and home user - which would be a damn big chunk of the market. It would be enough for those needs, but it would be cheaper and in many aspects less troublesome than Intel and AMD. Take for example power consumption in relation to computational power, where ARM is way better than the old and crusty x86 architecture. And less power leads to less cooling requirements, making the machines smaller, more energy efficient, and less noisy.
I can see ARM-based systems approaching this enough level, and I can see that Intel and ARM are deadly afraid of that scenario.
Valmond@lemmy.world
on 16 Oct 2024 15:35
nextcollapse
This!
I have wondered for a long time when we’ll hit that ceiling (ssd size, cpu power, ram, …) and I think it’s about right now. There are not many exciting PC hardware news nowadays is another sign IMO.
I also windered for a long time why I shouldn’t have a mobile phone PC, or more like “where are they?”, I have an old Xiaomi redmi note pro 9, 4+4 core with 6+2GB RAM (Whatever that +2 means), 128GB storage and, well, graphics. For not expencive.
It could be an OK home computer.
A little bit of interesting times ahead!
Treczoks@lemmy.world
on 16 Oct 2024 17:44
collapse
I think it will not take long until there is a cell phone/PC hybrid: you plug your cell phone into a base and can use it with a normal desktop interface on a screen with mouse and keyboard. A bit like the Nintendo switch.
ZiemekZ@lemmy.world
on 16 Oct 2024 18:31
nextcollapse
Samsung DeX? My Galaxy Note 9 already has it.
pearsaltchocolatebar@discuss.online
on 16 Oct 2024 18:48
collapse
The Steam Deck is a similar concept
But, my prediction for the majority of users is that the device will just connect to a vdi infrastructure that you pay monthly for.
schizo@forum.uncomfortable.business
on 16 Oct 2024 17:26
nextcollapse
basic needs of the average office and home user
I mean, ARM chips have been at that level of performance for at least a decade by now. Normal people’s most strenuous activity is watching Youtube, which every cellphone since what? 2005? could do.
power consumption in relation to computational power
The thing is that’s very much not the actual situation for most people.
Only Apple really has high performance, very low power ARM chips you can’t really outclass.
Qualcomm’s stuff is within single-digit percentage points of the current-gen AMD and Intel chips both in power usage, performance, and battery life.
I mean, that’s a FANTASTIC achievement for a 1st gen product, but like, it’s not nearly as good as it should be.
The problem is that the current tradeoff is that huge amounts of the software you’ve been using just does not work, and a huge portion of it might NEVER work, because nobody is going to invest time in making it behave.
(Edit: assuming the software you need doesn’t work in the emulation layer, of course.) You might get Photoshop, but you won’t get that version of CS3 you actually own updated. You might get new games, but you probably won’t get that 10 year old one you like playing twice a year. And so on.
The future might be ARM, but only Apple has a real hat in the ring, still.
(Please someone make better ARM chips than Apple, thanks.)___
IMALlama@lemmy.world
on 17 Oct 2024 01:55
collapse
Qualcomm’s stuff is within single-digit percentage points of the current-gen AMD and Intel chips both in power usage, performance, and battery life
The problem is that the current tradeoff is that huge amounts of the software you’ve been using just does not work, and a huge portion of it might NEVER work, because nobody is going to invest time in making it behave.
I agree with the sentiment, but IMO this is a PC and Windows problem. I would also extend this beyond pure comparability. I say this for a few reasons
I lose about 5% charge/day with my laptop asleep. It does wake up very quickly, but 5%/day feels like a lot. At this point, I don’t think Microsoft has a strong incentive to really optimize the kernal for efficiency
Historic massive variability in hardware across devices also makes it hard to optimize efficiency, although the current crop of snapdragon x laptops seem to have less variability
One of the strengths of windows is that it can run applications written 20+ years ago fairly reliably. There’s a ton of software that’s still floating around that hasn’t been actively supported in years. I don’t see all of these software companies desiring to port their code over, especially without guarantees that the market will adopt ARM (the Apple approach) or until they see the ARM adoption rate go up (the current Windows approach)
All that said, I’ve had zero issues with emulation so far. I never personally used a M1 max when they launched, but from reports of that era the current Windows experience is at least as good as that.
pearsaltchocolatebar@discuss.online
on 16 Oct 2024 18:47
nextcollapse
It always drives me insane when I have to spec out a $4k system for execs that use it mostly to browse Facebook and LinkedIn. At least the devs get the same systems.
Treczoks@lemmy.world
on 16 Oct 2024 19:02
nextcollapse
I’ve seen worse. A group at the university was using the IBM mainframe for basically everything from their terminals. To reduce load on the mainframe, the university spent a load of money to buy a cluster of workstations with crazy specs and software, each one more expensive than a big new car back then.
I visited them shortly after they got those killer machines. For comparison: in our university department, we had green serial terminals connected to an old VAX 11/780. They had those shiny new workstations with GUI on high-resolution (for that time) color monitors. My friend there logged in - and his autostart just opened two terminal sessions on the IBM mainframe, where he did all his work just like before. He was happy that he had the terminals in a windowed environment, though, so he could easily open and handle several sessions on the machine at the same time.
trolololol@lemmy.world
on 16 Oct 2024 21:15
collapse
Jira is the new reason we need to constantly upgrade
ZeroPoke@lemmy.ca
on 16 Oct 2024 19:21
nextcollapse
Valve is currently working on some tech to let x86 games work on arm64.
AceFuzzLord@lemm.ee
on 16 Oct 2024 20:36
nextcollapse
Cannot wait to see their solution.
Treczoks@lemmy.world
on 16 Oct 2024 22:51
collapse
That is a long shot at best. Games are hungry for power and resources, and adding an emulation layer, even a transpilation system between x86 code and ARM processor will not actually improve the situation.
WordBox@lemmy.world
on 16 Oct 2024 23:51
nextcollapse
Windows* runs on arm. Microsoft Surfaces user arm processors.
Even with windows apps being emulated… It’s basically enough already.
Windows* RT but Microsoft dropped the differentiator this time around.
IMALlama@lemmy.world
on 17 Oct 2024 01:25
collapse
It’s not just the surface devices anymore. In June this year, a fresh wave of ARM powered laptops from a verity of different OEs launched. There are offerings from Dell, HP, Lenovo, Samsung, Asus, etc.
IMALlama@lemmy.world
on 17 Oct 2024 01:35
collapse
I own a Lenovo Yoga slim 7x Gen 9, which is powered by a Snapdragon X. It certainly checks the “good enough” box. I use it primarily for photo culling/editing (I’m a holdout dedicated camera user). It is more than fit for purpose there, stays cool, is slim, and although I know the fan has come on a few times I wouldn’t have known if it wasn’t on my lap. When I bought mine, it was also one of the better deals - you could upgrade to 32 GB of memory and a SSD for under $125 in total. The SSD also isn’t soldered, but the memory is. The 3k OLED display is amazing, but if you want the ultimate battery sipper it’s probably not the best choice. I still get tons of runtime per charge, but am somewhat sad that I lose about 5% charge per day thanks to the laptop not really being off while asleep.
The biggest downside is linxu support is very hit and miss depending on the laptop in question, which means you’re tied to windows 11. I don’t have the time to tinker with it, so I haven’t looked much further into it than this.
01189998819991197253@infosec.pub
on 16 Oct 2024 18:37
nextcollapse
This means we must band together and wholeheartedly support the architecture. Clearly, it is a ship shaker, and industries don’t like ship shakers (even if it would result is a better industry or is better for the consumer, maybe especially if it was better for the consumer).
ARM is shaking the ship, so we must shake it too.
sugar_in_your_tea@sh.itjust.works
on 16 Oct 2024 20:54
collapse
Agreed. I’d really like RISC-V to be where ARM is, but having some competition is a good thing since it keeps that door open.
01189998819991197253@infosec.pub
on 17 Oct 2024 01:10
nextcollapse
RISC-V is where ARM was a few years ago. It’ll be a ship shaker, too, if it keeps developing at this rate. But, for now at least, we have ARM and clunky old (solid) x86 as the major players.
Croquette@sh.itjust.works
on 17 Oct 2024 12:14
collapse
Do you have good resources to read on risc-v. I hear about it a lot, but haven’t found meaningful resources (to me) on it.
Thanks
sugar_in_your_tea@sh.itjust.works
on 17 Oct 2024 14:31
collapse
I’m not sure what you’d like to read about, but risc-v is basically an open source CPU design. It’s not compatible with anything else, so developers will need to build software specifically for it, just like how x86 and ARM are completely different.
The nice thing about it being open source is that any company can tweak the design and send it off to a fab to get chips made, so you’re not really at the mercy of the big chip companies like AMD, Intel, or Qualcomm. So if someone like Samsung wants a super low-power device (e.g. for a watch), they can remove unnecessary stuff from the chip design and fab their own. Or they can go full-fat for a desktop chip. Or if they need a special instruction for their particular workload, they can implement it themselves and fab it.
At its current state, Linux largely works on it and there are hobby boards available and some other small devices, but it’s nowhere near ARM in performance or efficiency (hopefully getting there).
ARM isn’t the x86 solution people like it to be. It’s at least as proprietary as x86 and the fact that it’s more widely licenced today than x86 is a happy coincidence. ARM licensing can dry up with a change in corporate leadership or a takeover by one of a myriad large corporations. A solution worth cheering would be a good enough open RISC-V core.
MonkderVierte@lemmy.ml
on 17 Oct 2024 09:36
nextcollapse
And universal drivers are an issue in ARM, bad for Open Source.
shortwavesurfer@lemmy.zip
on 17 Oct 2024 12:02
collapse
threaded - newest
Really they need to work on power usage and temperature of x86 so the chips are easier to use in mobile devices without a fan and dying in 3 hours. Stationary devices seem to be chugging along with x86 comfortably, but the chips are currently impractical otherwise.
The new Intel chips already addressed that, at least for notebook class devices.
Realistically, there wasn’t really a reason for Intel and AMD to be super power efficient, simply because there wasn’t any competition for quite a while. It took Apple Silicon to show how powerful arm can be and how easy the transition could be.
Apple took all the old tricks Intel was always way too cheap to use, and turned them to 11.
Nothing magic, nothing special, just balls and the willingness to spend silicon.
You’re not going to see phones with x86. The architecture just isn’t going to scale down like that. Not if you want something faster than a Pentium III.
It actually can, the thing we learned is that the unpleasant bits of x86 scale well, so we spent 30% of the die doing uop decode, but that’s now just 1-2% because we blow so much on more registers and cache.
Also we can play games like soft-deprecating instructions and features so they exist, but are stupid slow in microcode.
We used to think only risc could run fast at low power, but our current cisc decoded to risc works fine, Intel just got stupid lazy.
Apple just took all the tradeoffs Intel was too cheap to spend silicon on and turned them to 11, we could have had them before but all the arm guys were basically buying ip and didn’t invest in physically optimized designs, but now that tsmc is the main game in town (fallback to gf was nice for price), there’s a lot more room to rely on their cell libraries.
Intel got so insanely arrogant, just like Boeing and all the other catastrophic American failures right now, we just need to correct for that and we can be decent again.
It’s hardly just Intel. There are two other x86 licenses out there. One gave up. The other is kicking ass, but Apple didn’t go with them, either.
Meanwhile, Intel themselves kept the 80486 alive until 2007 as an embedded processor. It outlasted the Pentium III by a few months. It was never as popular as PIC or ARM or z80 devices, but it found some kind of niche.
I’ll grant that in theory, it could be done. But why? There are millions of smartphones running fine with ARM, and they don’t have any backwards compatibility needs to x86. Why pick an ISA that can only legally be designed by three companies? Why pick an ISA that hasn’t been as well tested on mobile device OSes? ARM will hand a license to anyone who shows up with some cash, and if you want to take a plunge into a different ISA, then RISC-V is sitting right there. There doesn’t seem to be a single real benefit to x86 over what mobile device creators have now, and plenty of reasons not to.
No, it doesn’t make sense to do it.
I worked on platform enablement for armv8, bringing all the ecosystem to 64 bit arm. Was an everest, so much code was expecting x86, lots of secret asm and other assumptions like memory model.
But once it was done, we did it again for riscv in no time, all the work was done, it was basically setting defines, maybe adding tsc/rdcycle (now rdtime).
Architectures don’t really matter anymore, but also the overhead of architectures are pretty minor, riscv will probably win because it’s basically free and single thread performance isn’t as critical on client devices, lot of work goes to the GPU too, and servers do other heavy lifting. Qualcomm scared everybody too, and China is going their own way which means even more riscv.
Basically, nothing matters except cost now, we’ll figure out how to run things on a potato, we’ve gotten good at it.
ISA doesn’t matter as much as most people think it does. It’s all about how you implement it.
How do you hire people who can implement it right? There are three companies that can make x86. One is failing, one gave up years ago, and the third is kicking ass but seems uninterested in this part of the market. All the people who know how to do x86 well already work for one of them. That third company that nobody talks about gave up because by 2010, they lacked the ability to make a worthwhile product.
It’s an incredibly difficult ISA to work with, and all the talent is already busy. Due to its closed nature, there is little hope of significantly growing that talent base. Not unless you want the early 2000s version of x86-64, which is patent free.
Asus Zenfones used to use Intel Atom x86 processors.
And were they any good?
My car runs Android Automotive^1 on an Intel Atom and performance is trash. I would hate to have a phone on the same platform.
^1 As in, the car runs Android directly, not Android Auto running from a phone.
They were comparable to the rest of the phones at the time. Not great, not terrible. Compared to anything in 2024 they were obviously trash, but that’s mostly because we’ve made 10 years of progress since then.
The Samsung Galaxy is 15 years old, and it was excellent. First Android device I felt had decent performance. I think my first one was an Galaxy S III, which would have been 2012.
The Zenphone that’s out now uses a Snapdragon. There’s probably a good reason for that.
Yes, that there are no smartphone sized Intel Atom processors anymore.
The zenphone 2 performed similarly to much more expensive phones. www.anandtech.com/show/9251/…/4
I’m not going to be the person defending intel in 2024, but back in 2015, that atom was competitive.
It seems that they’re finally taking that seriously though so it’s good to see. They never really had any incentive to put too much effort in making x86 more efficient for consumer devices since their server chips have much, much higher profit margins.
Lunar Lake and AMD’s Z1 is a good start and it’s interesting to see where this goes.
It’s amazing what a modern process node and not cranking clock speeds to high hell will do.
You forgot about ditching more of the chipset etc. in favour of integrating everything into the CPU die.
I’m conflicted on ARM.
The additional competition is great, but it presents a great risk of PCs becoming more locked down. They don’t have an open, standardised BIOS/UEFI like x86 systems do.
Booting alternate OSes on ARM systems can be a nightmare. Usually it’s straight up not possible.
I don’t want PCs to be like smartphones. I don’t want locked bootloaders.
EDIT:
FFS people. I know there are some ARM devices that allow booting of non-official OSes. That’s why I said usually.
Even for those devices though, they typically have to use non-standardised firmware (you can’t just take an OS for device A and use it for device B in the same way you can take an .iso and install it on any x86 machine), and it requires the OEM to want the device to be open.
Your desire to go “umm ackshully…” and be technically correct over a point I never made in the first place is blinding you to the point I was actually making: x86 is fairly open, standardised, and modular by default. ARM isn’t. And all it takes is a look at the phone/tablet market to see that OEMs don’t want them to be.
I worry, and I don’t think unreasonably, that ARM becoming the standard could mean a further erosion of the openness of PCs.
I think that there’s a legitimate place for all-in-one “smartphone” SoC PCs. You can make them cheaper, smaller, and use less power.
It’s just not really what I want for myself in a PC. I want the modularity and third-parties competing to provide components.
But I am pretty sure that there are plenty of people who don’t care about that.
There has to be enough scale to support products like that, though. SoC systems might cannibalize enough to make scale hard.
sigh
Well, we’ll see where things go.
I mean, different use cases yeah? There’s certainly a big market for people that just do the basics on their devices, ie email, web browsing, documents/spreadsheets etc that don’t need a full blown powerhouse computer, nevermind that they have no clue what an operating system even is nor do they care, as long as they get their memes and cat videos in between work tasks.
I’ll bet there will always be an x86 segment of the market for gamers, power users, tinkerers, and the like. Though, that market may unfortunately shrink in the coming years that could lead to vendors abandoning the space, which could lead to fewer choices and higher component prices. On top of which, major venders might see it as an opportunity for lock-in and advertising, so yeah it’ll be interesting to see what happens.
The sliver of hope here is that the hacking community has always found ways around proprietary bullshit, and we can only hope and support that those efforts continue, lest we further our race in to a stupid corporate dystopia.
Standardized firmware isn’t something that’s specified in the ISA, is it? It’s just shitty phone manufacturers.
Asus had some x86 phones a few years back. I haven’t dug into them, but I doubt they had a full bios/efi.
pine64 arm devices have u-boot, while a bootloader does fullfil a subset of the uefi spec.
Tbh I really want to get my hands on a snapdragon X laptop at some point just to play around with it. The energy efficiency alone makes me very curious.
I was under the impression that most of the issues around getting Linux to work on them was around driver support. As in: people are absolutely able to install an arbitrary OS, but the functionality is just super janky in most cases. Is that not accurate?
You’re definitely right in terms of arbitrary OS installation, some folks have got Ubuntu running on Lenovo snapdragon laptops recently.
The lack of “portability” though is a bit troubling, it seems each device (tree) has to be manually added, developed, tested, and have an install image created for it, unless I’m missing something. And this will be arduous and potentially problematic for corner cases or small numbers of adopters of a particular machine model (so basically the same as right now I guess).
As well, software packages have to be ported to the new architecture, which in some cases is easier said than done. Sure basic Unix utilities are portable enough, but more advanced and complicated software might have some issues, unless an efficient compatibility layer could be developed.
Don’t be conflicted. RISC-V or GTFO.
Yep it keeps getting faster and faster.
“it keeps getting faster and faster” isn’t really saying much when it’s 1/10th the performance of a raspberry pi.
Yeah, but RISC-V also costs 1/10th the price of a Pi.
To license the arcitecture it costs a whole lot less, but when it comes to getting an actual usable computer they cost the same or more as an ARM machine, and perform worse.
For micro controllers (currently) it’s great.
Oh, I absolutely agree. Licensing is where the big difference is at, but that makes sense though, as ARM and RISC-V are both RISC based processors.
It’s loosely akin to comparing AMD vs Intel. Of course, you cannot pop-out an RISC-V and replace it with an ARM. However, the PCB’s should contain all the same parts, meaning they’ll have both have a similar price.
Unlike Intel/AMD, which you’d need extra capacitor, heat sinks, whatever - to help it handle all that extra power those CISC processors need (which results in heat).
The pi foundation is starting to roll out chips with risc v processors built in. It’s extremely cheap and open spec.
One example is RP2350.
More options in computing is better for everyone.
Someone just needs to work on some open-source GPU for it, otherwise it’ll still have some of the usual shortcomings of many ARM SoCs.
“RISC is good.”
Not all ARM chips are in phones, nor are they all locked down like one. There are several ARM devices and SBCs now where switching OSes is as easy as swapping out an SD card. Most do use uboot as a standard and some are even capable of utilizing UEFI.
I was gonna say I thought Raspi was ARM.
But it’s not standard.
What made PCs take off was the BIOS war, which occurred because manufacturers were dependent on 3rd party OS’s, which were still competing for dominance.
Where can I watch or read about this? Like what can I Google for more?
Some SBCs only boot from said SD card though, while some do support more robust media. However, too many images are presuming you boot from SD which is a pita.
With or without Das Uboot, they still rely on board specific firmware (even Uboot is customised for many boards to make it work). OSes that state they do support aarch64, often require to have UEFI on your system so no way they are gonna boot on e.g. your Raspberry Pi.
Add to that, that is unlikely that browsers compiled for arm64 will have feature parity with their x86-64 counterparts. Goodbye Digitale Rights Management, and with that goodbye services like Tidal or Spotify (unless you run an OS that is still supported by their apps).
This! Exactly!
Raspberry Pi owners: 🙄
“usually”
I am a raspberry pi owner (actually I’ve owned like a dozen and still own 8). And a pinephone owner.
And even on those very open devices, booting alternate OSes is less straightforward than on x86.
As a simple user of Linux, I totally understand what it means for me…less choice, more google-android-like shit hardware. No thanks 👍.
I’m sorry to burst your bubble, but since Microsoft made TPM mandatory for Windows 11+, locked down bootloader are on their way.
Basically, TPM allows (Windows) software to validate/verify the integrity of the OS and hardware. This also (could) include the bootloader/bios if Microsoft chooses to do so.
TPM is the equivalent of attestation on Android, which is the exact reason why your Banking App won’t work on your rooted/custom Android Phone.
That being said, we should embrace ARM. X86/AMD has 30+ years worth of “history” baked into each ( CISC) chip. This complexity is why your PC draws soooo much power and generates soooo much heat.
I would like RISC-V to also get ‘good enough’, considering that it seems to be a shitton more open than ARM is.
UEFI exists on arm and windows on arm devices can boot other OSes through it just like on x86.
LOL this is the biggest sign out there that ARM is making a superior product. Once people start going protectionist, it’s time to move on to the new thing.
This is a sign of ARM approaching the “enough” level. I remember the times when it was actually important to buy the latest PC at least every other year to have enough power to run a basic office suite or similar programs with acceptable speed.
Nowadays, you can staff offices with about any PC off the shelf - it is powerful and big enough to fulfill the needs of the majority of users. Of course there are servers, there are power users, engineers running simulations, and of course gamers who need more power, and who still fuel the cutting edge of PC building. But the masses don’t need to be cutting edge anymore. A rather basic machine is enough.
Here comes the ARM: For many years, ARM-based chips were used as SOCs, running anything from washing machines to mobile phones. But they have grown bigger and faster, and I can see them approaching the point that they can cover the basic needs of the average office and home user - which would be a damn big chunk of the market. It would be enough for those needs, but it would be cheaper and in many aspects less troublesome than Intel and AMD. Take for example power consumption in relation to computational power, where ARM is way better than the old and crusty x86 architecture. And less power leads to less cooling requirements, making the machines smaller, more energy efficient, and less noisy.
I can see ARM-based systems approaching this enough level, and I can see that Intel and ARM are deadly afraid of that scenario.
This!
I have wondered for a long time when we’ll hit that ceiling (ssd size, cpu power, ram, …) and I think it’s about right now. There are not many exciting PC hardware news nowadays is another sign IMO.
I also windered for a long time why I shouldn’t have a mobile phone PC, or more like “where are they?”, I have an old Xiaomi redmi note pro 9, 4+4 core with 6+2GB RAM (Whatever that +2 means), 128GB storage and, well, graphics. For not expencive.
It could be an OK home computer.
A little bit of interesting times ahead!
I think it will not take long until there is a cell phone/PC hybrid: you plug your cell phone into a base and can use it with a normal desktop interface on a screen with mouse and keyboard. A bit like the Nintendo switch.
Samsung DeX? My Galaxy Note 9 already has it.
The Steam Deck is a similar concept
But, my prediction for the majority of users is that the device will just connect to a vdi infrastructure that you pay monthly for.
I mean, ARM chips have been at that level of performance for at least a decade by now. Normal people’s most strenuous activity is watching Youtube, which every cellphone since what? 2005? could do.
The thing is that’s very much not the actual situation for most people.
Only Apple really has high performance, very low power ARM chips you can’t really outclass.
Qualcomm’s stuff is within single-digit percentage points of the current-gen AMD and Intel chips both in power usage, performance, and battery life.
I mean, that’s a FANTASTIC achievement for a 1st gen product, but like, it’s not nearly as good as it should be.
The problem is that the current tradeoff is that huge amounts of the software you’ve been using just does not work, and a huge portion of it might NEVER work, because nobody is going to invest time in making it behave.
(Edit: assuming the software you need doesn’t work in the emulation layer, of course.) You might get Photoshop, but you won’t get that version of CS3 you actually own updated. You might get new games, but you probably won’t get that 10 year old one you like playing twice a year. And so on.
The future might be ARM, but only Apple has a real hat in the ring, still.
(Please someone make better ARM chips than Apple, thanks.)___
Back in June, the new Snapdragon X processors were a lot more efficient than their x86 based counterparts. I can personally attest to much lower levels of heat generation.
I agree with the sentiment, but IMO this is a PC and Windows problem. I would also extend this beyond pure comparability. I say this for a few reasons
All that said, I’ve had zero issues with emulation so far. I never personally used a M1 max when they launched, but from reports of that era the current Windows experience is at least as good as that.
It always drives me insane when I have to spec out a $4k system for execs that use it mostly to browse Facebook and LinkedIn. At least the devs get the same systems.
I’ve seen worse. A group at the university was using the IBM mainframe for basically everything from their terminals. To reduce load on the mainframe, the university spent a load of money to buy a cluster of workstations with crazy specs and software, each one more expensive than a big new car back then.
I visited them shortly after they got those killer machines. For comparison: in our university department, we had green serial terminals connected to an old VAX 11/780. They had those shiny new workstations with GUI on high-resolution (for that time) color monitors. My friend there logged in - and his autostart just opened two terminal sessions on the IBM mainframe, where he did all his work just like before. He was happy that he had the terminals in a windowed environment, though, so he could easily open and handle several sessions on the machine at the same time.
Jira is the new reason we need to constantly upgrade
And edr solutions
Valve is currently working on some tech to let x86 games work on arm64.
Cannot wait to see their solution.
That is a long shot at best. Games are hungry for power and resources, and adding an emulation layer, even a transpilation system between x86 code and ARM processor will not actually improve the situation.
Windows* runs on arm. Microsoft Surfaces user arm processors. Even with windows apps being emulated… It’s basically enough already.
Windows* RT but Microsoft dropped the differentiator this time around.
It’s not just the surface devices anymore. In June this year, a fresh wave of ARM powered laptops from a verity of different OEs launched. There are offerings from Dell, HP, Lenovo, Samsung, Asus, etc.
I own a Lenovo Yoga slim 7x Gen 9, which is powered by a Snapdragon X. It certainly checks the “good enough” box. I use it primarily for photo culling/editing (I’m a holdout dedicated camera user). It is more than fit for purpose there, stays cool, is slim, and although I know the fan has come on a few times I wouldn’t have known if it wasn’t on my lap. When I bought mine, it was also one of the better deals - you could upgrade to 32 GB of memory and a SSD for under $125 in total. The SSD also isn’t soldered, but the memory is. The 3k OLED display is amazing, but if you want the ultimate battery sipper it’s probably not the best choice. I still get tons of runtime per charge, but am somewhat sad that I lose about 5% charge per day thanks to the laptop not really being off while asleep.
The biggest downside is linxu support is very hit and miss depending on the laptop in question, which means you’re tied to windows 11. I don’t have the time to tinker with it, so I haven’t looked much further into it than this.
This means we must band together and wholeheartedly support the architecture. Clearly, it is a ship shaker, and industries don’t like ship shakers (even if it would result is a better industry or is better for the consumer, maybe especially if it was better for the consumer).
ARM is shaking the ship, so we must shake it too.
Agreed. I’d really like RISC-V to be where ARM is, but having some competition is a good thing since it keeps that door open.
RISC-V is where ARM was a few years ago. It’ll be a ship shaker, too, if it keeps developing at this rate. But, for now at least, we have ARM and clunky old (solid) x86 as the major players.
Do you have good resources to read on risc-v. I hear about it a lot, but haven’t found meaningful resources (to me) on it.
Thanks
I’m not sure what you’d like to read about, but risc-v is basically an open source CPU design. It’s not compatible with anything else, so developers will need to build software specifically for it, just like how x86 and ARM are completely different.
The nice thing about it being open source is that any company can tweak the design and send it off to a fab to get chips made, so you’re not really at the mercy of the big chip companies like AMD, Intel, or Qualcomm. So if someone like Samsung wants a super low-power device (e.g. for a watch), they can remove unnecessary stuff from the chip design and fab their own. Or they can go full-fat for a desktop chip. Or if they need a special instruction for their particular workload, they can implement it themselves and fab it.
At its current state, Linux largely works on it and there are hobby boards available and some other small devices, but it’s nowhere near ARM in performance or efficiency (hopefully getting there).
ARM isn’t the x86 solution people like it to be. It’s at least as proprietary as x86 and the fact that it’s more widely licenced today than x86 is a happy coincidence. ARM licensing can dry up with a change in corporate leadership or a takeover by one of a myriad large corporations. A solution worth cheering would be a good enough open RISC-V core.
And universal drivers are an issue in ARM, bad for Open Source.
Tisc-v ftw