I don’t think you guys understand the “/s” and “in all seriousness” part of the post. Clearly the first statement was sarcasm. Or maybe not so clearly …Barely outrun an M1 MBA, now that's funny. Even my medium end i9 outruns my M1 MBA easily, and this new chip blows my i9 away. (The i9 machine also costs about $24000 less.![]()
Who cares about single core performance when running a general purpose computer -- I don't. I don't even run VM's with a single core!No Intel i9 can outruns M1 MBA on single core performance, which was what Intel loved to marketing when they have lead in that field.
Definitely not clearly, sorry.I don’t think you guys understand the “/s” and “in all seriousness” part of the post. Clearly the first statement was sarcasm. Or maybe not so clearly …
They will sell four scores of these things.They will sell dozens of these things.
Probably not. Intel is still missing key hardware to power the AI stuff in macOS.
I guess this will be a special option dedicated to those movie studios just like how dual G5 was.
Please describe your idea of a ”real” computer? Pretty sure I know what you think that is, but just for the audience members who may be lost in your heady rhetoric.Didn't think apple will release a real computer ever again.
Who would have thought.
Apple is facing a much larger problem than GPU integration or hardware in general. Software support is the problem. While I love the AS SoCs, for most of the scientific world, macOS is almost dead at this point (Intel or AS). When Apple dropped Nvidia we had to start working around problems, with OpenCL. etc. The state of the software world is catastrophic and simply not worth to bother with. Nvidia software won't work, no problem (well, not really but expected and acceptable). Things like Tensorflow, Pytorch, Carla, AirSim, other simulators and frameworks either don't fully work or are a major pain to work with unless one is ok with CPU only support. Apple did a Tensorflow fork a while back. It's bugged, doesn't fully work, it's in a useless state. But what should we expect when Apple is using Linux and Nvidia to train their own AI? Once the research is done, it's easy to convert models to support metal including GPU and Neural Engine for inference. But how does that help the people who have to do research?
At this point (and of course that can change in the future), Apple is really good for reading/writing, browsing the web and emails. In addition, musicians can use it, so can photographers and YouTubers to cut their videos. I do a bit of photo image processing and lightweight video work, but otherwise my Macs have become a tool for lightweight work I could easily do on a MBA/MacMini while I have to use Linux on different hardware to get the heavier workloads done.
I think that the ARM version they'll come up with will be faster than the Intel one. They will just update the intel because of legacy app support.They will, but it'll still take Apple more time to develop a home-grown solution that can match a 32-core Xeon with 128GB of RAM and four dedicated desktop GPUs. I would wager that the next Intel Mac Pro revision will be the last Intel Mac Pro.
Well, no 4 core intel matches the multicore performance of M1 either. Only 8 core Intels can beat it, that's basically twice the amount of performance cores to beat 4 performance cores.Who cares about single core performance when running a general purpose computer -- I don't. I don't even run VM's with a single core!
We do not care about a computer with only 1 core. But single core performance is still and will still be really important as a lot of workloads are not mathematically possible to be paralleled.Who cares about single core performance when running a general purpose computer -- I don't. I don't even run VM's with a single core!
Because businesses are still owning a lot of x86 Mac software. The transition is not meant to be a immediate end to what Apple sells to companies as solutions. Common sense requires high performance ARM workstation/server products to be introduced that sways enterprise customers to migrating, not heaving their intel based business solutions overboard.![]()
Just because it's likely doesn't make it less disappointing.You knew Apple is likely to continue to sell intel based Mac Pro's per WWDC 2020 statements...
This is being done so that business customers who can't transition to m1 computers yet, still have an option if they need to buy a mac. I suspect that a maxed out 16" Macbook Pro will exist for a while longer, too.Even if the tech is amazing, buying an Intel Mac in 2022 seems odd.
Who owns a 4 core intel machine for serious work these days? My i9 desktop is 10 core + HT, 128G RAM, my home Windows laptop is 6 core +HT 32G RAM. My Intel Mac Mini has 6 cores +HT and 64G RAM. And I don't even own a high end machine.Well, no 4 core intel matches the multicore performance of M1 either.
The important thing is they beat it. It would be good if there was an M processor in their league, but there isn't *yet* unfortunately. (I'd own it at home if there were!)Only 8 core Intels can beat it, that's basically twice the amount of performance cores to beat 4 performance cores.
You can take out the we there, I don't agree with you at all. And I'm not talking about parallel processing, just general purpose computing. I don't do much video, but I think you're way underestimating just what more cores can give you when doing video -- you don't have to be only processing just 1 frame/stream at a time...We do not care about a computer with only 1 core. But single core performance is still and will still be really important as a lot of workloads are not mathematically possible to be paralleled.
Even video editing is still heavily single core bond in most scenario.
but with thing like big sets of raw data you may need an 10G uncapped fiber line to use the clouldYou can scale x86 as much as any other platform, just not in a small low power box (https://home.cern/science/computing/data-centre). There is however a point where it makes much more sense to offload work to the cloud or local data center instead of having that power in a single box under the desk.