raw power < effiencymake the hardware work without it running like a steam enginedont know why this is so uncontroversial opinion among general users when this is obvious shit in industry
>>107264624being able to run whatever the fuck you want on your hardware > armkys jew
>>107264652thats not ARM issue specifically by design
>>107264656It kinda is, due to leaving some things nonstandardized.
>>107264624Lunar Lake has obliterated x86 efficiency records and at 3 watt average power consumption for web browsing it's plausible that x86 smartphones will become a thing again.IMHO it's not that ARM is more energy efficient than x86, it's cheaper. That's really it.
>>107264743that will sort out itself
>>107264624Same, I don't want loud, big and hot PCs anymore. A fucking GPU with 300w + power draw on top of all the other hardware. I'd love to get the Ryzen AI 395+
>>107264791... in 2050
>>107264777>IMHO it's not that ARM is more energy efficient than x86, it's cheaper.cheaper as in the cost in currency? like the hardware?
>>107264905once ARM comes into general computing it will be a thingand looking at the current way things are going...I am gonna guess its gonna take 15 years max
during the space race they could put a satellite in orbit with a computer as powerful as a calculator
>>107264777Sit down bitch
>>107265006Looking at this chart makes me want to go with an Apple M5 system at some point in the future.
>>107264777>>107265006
I agree. I switched my Proxmox hosts at home with a couple ARM SBCs and I barely consume power now, don't generate any heat and my rack is fully silent. All the services I need run on ARM, and Docker makes it all pain-free and the performance is more than sufficient.I've tried to go back to x86 a few times but I really see no reason to do it. My complaint is that ARM is too closed - there's nothing like BIOS/UEFI that lets me install the OS I want instead of having to deal with vendor specific shit.As I plan to build a solar powered rack, I tried many times to get my x86 hosts to use less power but C-states and ASPM runs like garbage when you can make it to work (most of the time you can't). I wasted a lot of time and money but I just had to move to ARM.
>>107265032Go for it. Tahoe is a shit show though. Sold my macbook cause it's full of bugs cause of that update.If I were you, I'd skip m5 and wait for the next year's MacBook where they will have a tandem OLED display.
>>107265058pretty sure snapdragons laptops boot anything that supports itits driver issues mainly with linux and new hardware like usual
>>107264938In 15 years we're not gonna be making computers anymore
>>107264777Lunar lake 258V at 15W+8-10W for the platform and rest of the system is literally slower than the Snapdragon Elite G at 10-12W for an entire smartphone.
>>107264791arm is almost as old as x86shit ain't happening bro
>>107264908Everything. Apple for example stopped buying x86 CPUs from intel and made their own custom ARM designs that TSMC printed out. They ultimately ended up paying a fraction of what they payed intel and they got full control of their hardware which resulted in more cost cutting.>>107265006>>107265033>>107265340I don't think you understand how serious of a threat x86 having an average of 3 watts of power consumption for web browsing is to ARM. ARM for like the past decade or so has been UNTOUCHABLE below 5 watts. The only exceptions have been very anemic atoms/celerons that lacked any L3 cache.The Ultra 7 258V isn't just clocking a meager 3 watts during web browsing, it's also one of the first 8-core X86-64-v3 CPUs with 12MB of L3 cache to do so. You could argue that they're "cheating" by using only 4 performance cores and 4 efficiency cores but all of them are officially X86-64-v3 compliant AFAIK.https://hackaday.com/2024/02/25/what-is-x86-64-v3/>The 258V has a very healthy 12 MB of level 3 cachehttps://www.notebookcheck.net/Intel-Core-Ultra-7-258V-Processor-Benchmarks-and-Specs.892883.0.htmlA phone manufacturer that isn't a coward could unironically release a smartphone with this x86 CPU and achieve like 6 hours of web browsing battery life with a modern 5,000 mAh battery. It would be destroyed in battery life by other ARM phones but it would officially be the first X86-64-v3 phone which would be based IMHO.
>>107265459There are consortiums to agree on platforms now that Arm is used a lot in data centers.
>>107265471>AIIIE MY CPU IS USING 5 WATTS INSTEAD OF 3 AND I HAVE TO PAY AN EXTRA 3 CENTS EVERY DECADE ON MY ELECTRICITY BILL IT'S THE END OF THE WORLD
>>107265471That would be impressive, but current Arm chips can browse on cellular data while using 1W for the entire smartphone. The S25 Ultra has a 5000 mAh battery rated at 3.88V. That's around 19 Whr and it can browse the web on cell data for 17 hours. That's 1.1W. It's probably sub-W on wifi. And this chip is faster than the 258V unless you push the 258V to like 35W. Both on 3nm.https://www.tomsguide.com/phones/samsung-phones/galaxy-s25-edge-battery-life-tested-its-even-worse-than-we-thought
>>107264624>ARM>efficiencyIt's a meme. Zen whatever we're on now cores consume just as little, it's just that AMD doesn't give a shit about optimizing the rest of their SOCs. Intel tried, Intel did a good job of it and then Intel gave up because of some corpo kikery with RAM procurement.
>>107265515Except on mobile devices it's not the electricity bill you have to worry about, it's the battery with a finite number of times you can charge/discharge it. That said a 5,000 mAh phone with an Ultra 7 258V x86 CPU at 6 hours of web browsing battery life is actually usable desu. You'd have to replace the battery in like 2-3 years though. Annoying but still usable.>>107265567Yeah, that's expected and a phone like this won't need a battery replacement for 8-10 years but again sub-5 W was ARM dominated for the past 10 years or so and now it isn't which is exciting to say the least. If intel has truly learned from history then hopefully Lunar Lake is the tip of the x86 phone future iceberg.
>>107265471>3 watts during web browsingMy 7nm mobile Ryzen 5 does this. As in the whole SOC. It's not impressive. Are you sure it's not the whole laptop you are talking about?
>>107264624I don't care what's voodoo shit I don't understand in my PCs makes it work. I have an MacBook on ARM. It's great. Runs fast, feels snappy, has huge battery life. I have a desktop gaming-aislop PC. It runs great, draws 700 watts at peak. It good besides the time when it starts to get hot and spin fans for apparently no reason.
>>107265645Yes, it makes sense for mobile devices that you only use to browse websites, but I'm not going to use my mobile phone to render out complex 3d scenes, so I absolutely do not give a shit about hitting 300W for maximum performance
>>107265006>armbenchkys
>>107265058>there's nothing like BIOS/UEFI that lets me install the OS I want instead of having to deal with vendor specific shit.this seems to be the problem with risc-v as well. if you want widescale consumer adoption as an industry standard you need an industry standard boot system. i don't understand why so many seem to miss this.
>>107265651Everything and the delta from idle minimum is actually 0.3W as shown in >>107264777 BUTT it's just easier for people to understand the average better, no long explanation about SoC power required and all that.Still 8-cores, 12MB L3 cache, X86-64-v3 are a combination you would never expect at 3W average power consumption for web browsing.
>>107265665Ok retard
>>107265739Now you have to dig out the GPU benchmarks for comparing CPUs instruction sets? And you expect me to believe this?
>>107265759CES is soon. Just wait till then. And PM me your response faggot :)
i wonder where data centers fall on this trade off. It generally seems retarded to me to double power consumption just to increase performance by 10%
>>107264624x86/64 will continue to dominate for as long as arm/risc machines dont fucking standardize shitnobody is going to deal with having hyper-specific bullshit just to boot the goddamn machine i'm sorry
