alucard0712_rus
Banned
It's absolutely amazing but why do you need such power in a phone? Maybe for VR or some other things but not for a fucking android phone.
Dude read the first sentence of OPIt's absolutely amazing but why do you need such power in a phone? Maybe for VR or some other things but not for a fucking android phone.
True, but Rosseta translation layer is pretty good. Hopefully MS have something similar.
Most important is for MS to release SDK, so you wouldn't care, if you are compiling to x86 or ARM64, as it is on Apple platforms
It's absolutely amazing but why do you need such power in a phone? Maybe for VR or some other things but not for a fucking android phone.
Is it x64 compatible?Dude read the first sentence of OP
Is it x64 compatible?
Apple popularized it to an extent, but tech sites have been talking about the coming of ARM for at least a decade now. I doubt dying quickly in the cards since this is really still in the early stages of ramping up.I like that we are giving up the more power without limit. Apple open the door and now competition is building up.
let see if the market follow or all this arm stuff in the computer space will die quickly.
It has to be a when more than an if I’d think.What the chance these arm chips start being the main thing for pc and pc gaming? Isn’t the gpu side still pretty weak when it comes to that for now?
This is only for laptops, power budget is way too high for mobile. New mobile chip announced is the snapdragon 8 gen 3.I just bought Pixel 8 Pro but if this is in a new Galaxy then I may jump back on the cooter train.
the next battle for mobile is local AI assistant, so phones will become more powerful from now on, specially in neural cores and memory bandwidth/capacity.It's absolutely amazing but why do you need such power in a phone? Maybe for VR or some other things but not for a fucking android phone.
Is it x64 compatible?
Who gives a shit about readingLiterally says in the first sentence that it’s for Windows PCs
Apple already demonstrated than you can write an extremely efficient translation layer for x86->arm. I don't think this will be a problem.No wonder Xbox is looking at ARM for the next console. Would be huge but back compat may be impacted.
Apple already demonstrated than you can write an extremely efficient translation layer for x86->arm. I don't think this will be a problem.
They might need something different for 360 / OG Xbox, indeed. MS has some of the worlds best developers. If they really wanted it, I'm sure they could do it.Will be as soon as you start to emulate emulation.
RDNA 3 is an architecture and has nothing to do with Tflops. Power and architecture are 2 completely different things. Ideally you want the best of both worlds of courseHow can the GPU be faster than RDNA 3 if it’s only 4.6 TFlops? I know TFlops aren’t everything but that’s a stretch.
They might need something different for 360 / OG Xbox, indeed. MS has some of the worlds best developers. If they really wanted it, I'm sure they could do it.
The Apple M-chips have special instructions and additional support logic to speed up emulation of x86 code. I've not seen any of that in non-Apple ARM chips.True, but Rosseta translation layer is pretty good. Hopefully MS have something similar.
They have had ARM support for Windows for years with the compilers to support it. Finding hardware that can be used to actually test ARM on Windows is more of a challenge, not to mention finding users actually using it.Most important is for MS to release SDK, so you wouldn't care, if you are compiling to x86 or ARM64, as it is on Apple platforms
Eh, article clearly says it's for PC. Even the title says it.It's absolutely amazing but why do you need such power in a phone? Maybe for VR or some other things but not for a fucking android phone.
Eh, article clearly says it's for PC. Even the title says it.
One step closer to phones becoming docked consoles. Not to mention steam consoles and more portables at better prices......
Nintendo phone in 2035 with 25 tf of performance......
1) Yeah true, but it can't be that complex, because anything remotely x86 has to come from either AMD or Intel and it cannot be licencedThe Apple M-chips have special instructions and additional support logic to speed up emulation of x86 code. I've not seen any of that in non-Apple ARM chips.
They have had ARM support for Windows for years with the compilers to support it. Finding hardware that can be used to actually test ARM on Windows is more of a challenge, not to mention finding users actually using it.
Qualcomm bought a company started by ex-Apple chip engineers. Gee, imagine that!This is the technology acquired when Qualcomm acquired the startup company Nuvia.
The founders of Nuvia are all ex-Apple chip design people, and presumably worked on Apple Silicon before jumping ship to found Nuvia.
What I'm saying is this is a big deal. x86 is decrepit and should have been replaced a decade ago or earlier. Getting PC's off x86 is good for everyone, except Intel, and honestly fuck Intel.
You can already run windows games and apps with wine and dxvk on phones that have qualcomm soc with compatible mesa drivers.
If this thing has decent driver support it should run.
129 million Nintendo Switch don't exist?I hope this could bring Nvidia to console space again...
How can the GPU be faster than RDNA 3 if it’s only 4.6 TFlops? I know TFlops aren’t everything but that’s a stretch.
You write the code in any of the portable languages, most common being C++. Then you just cross compile it to the architecture you want. Heck, you can cross compile it on Linux and target Windows x86 or ARM easily if you want, using LLVM aka clang or similar.2) Sure, however you can't compile any code, you have to write it for ARM, that's the difference, besides many libraries are distributed as a binary file, not code, which can be compiled to to ARM
All in all there are challenges, to overcome, besides it is harder if you don't control the whole ecosystem like Apple does
Could be interesting. But AFAIK the reason Apple M-series chips are so performant is not so much the ARM CPU but the conglomeration of specialized chips and transistor/cache arrangement on what is a huge SoC. Granted, a still very power-efficient one.
Also, x86-64 has a lot of instructions ARM doesn't, and from what I gather, while there's a lot of nonsense in the former, what it has would require the latter to go through more instructions to replicate. That may have changed somewhat for all I know.
Also, Intel has yet to really push out its chiplet strategy. (Not that I'm expecting it to compete directly with Apple and other ARM based macro SoC solutions, but we'll see.) And just in case they decide to cut all that legacy baggage, they have x86-S (S for simple), a pure 64-bit instruction set.
Microsoft will 100% experiment with probably something like Proton but for ARM at least. Windows need close 100% BC as one of the strongest points (up to DOS). As far as I recall, Windows for ARM did not succeed due to BC.True, but Rosseta translation layer is pretty good. Hopefully MS have something similar.
Windows on ARM has had x86 emulation for a while, and added x86-64 as well
Microsoft Adds 64-bit x86 Emulation to Windows on ARM
Microsoft announced today that the expected support for 64-bit x86 emulation on Windows on ARM devices has arrived, provided you are running Build 21277. You'll need to be part of Microsoft's Windows Insider program to test the build.www.extremetech.com
Could be interesting. But AFAIK the reason Apple M-series chips are so performant is not so much the ARM CPU but the conglomeration of specialized chips and transistor/cache arrangement on what is a huge SoC. Granted, a still very power-efficient one.
-There are licensing fees for using ARM, so will RISC-V be gaining momentum soon since its open source?
Yes, so what I meant is that many libraries or hell most of them aren't compatible with Win ARM64. If you would have thsoe libraries given to you with cpp/h files, that would obviously would be less of an issue, however open-source is fine, but MS don't really like that.You write the code in any of the portable languages, most common being C++. Then you just cross compile it to the architecture you want. Heck, you can cross compile it on Linux and target Windows x86 or ARM easily if you want, using LLVM aka clang or similar.
All the standard libraries on Windows also support ARM. STL is not an issue at all as it's always in source form. I assume you're talking about some third party DLLs in binary-only form, but those are not very common shouldn't influence this at all.Yes, so what I meant is that many libraries or hell most of them aren't compatible with Win ARM64. If you would have thsoe libraries given to you with cpp/h files, that would obviously would be less of an issue, however open-source is fine, but MS don't really like that.
vcpkg already exists and it's awesome. Microsoft open source project, used everywhere and it supports linux, macOS and Windows.Also I hope that sooner or later they will introduce some competent package manager
Time move fast then, will take a look at itAll the standard libraries on Windows also support ARM. STL is not an issue at all as it's always in source form. I assume you're talking about some third party DLLs in binary-only form, but those are not very common shouldn't influence this at all.
vcpkg already exists and it's awesome. Microsoft open source project, used everywhere and it supports linux, macOS and Windows.
They fall flat when running native ARM software?The benchmarks don't really cover the differences very well, in general usage this could be slower than x86 chips, I'd love if it were faster but these ARM claims always fall flat in real world usage.
Obviously an integrated GPU will not be as fast as a dedicated GPU. But I think ARM chips still support discrete GPUs. So that shouldn't be an issue.What the chance these arm chips start being the main thing for pc and pc gaming? Isn’t the gpu side still pretty weak when it comes to that for now?