wizard69
About
- Banned
- Username
- wizard69
- Joined
- Visits
- 154
- Last Active
- Roles
- member
- Points
- 2,255
- Badges
- 1
- Posts
- 13,377
Reactions
-
Apple's claims about M1 Mac speed 'shocking,' but 'extremely plausible'
Apple has never been successful selling servers. In fact I'm not even sure they understand the market for small systems.kkqd1337 said:I'm dubious of their claims. But i'm all for it if its true. I'll probably wait for M2 tho as I dont need anything right now.
One thought I had....
if these chips are so amazing.... why are they not making servers? using them in data centres, thats where power per watt really matters isnt it?
As for power per watt that is exactly where Apple has been most competitive. Performance per watt is what makes the iPad great, they have great functionality in a thin device with performance nobody can match. This is even more obvious when you consider that the MBA is now fan free. Performance per watt is far more importnat in the portable world than in the data center. In the data center it is about the electrical bill, in the portable world it is about getting work done without the battery dying.
-
How to watch the Apple Silicon Mac 'one more thing' event later today
This should be the most interesting Apple event in years. I hope it isn’t watered down to grade 5 level.
-
How Apple Silicon Macs can supercharge computing in the 2020s
Nobody cares any more. Back in 2008 I actually thought that that was important in my decision to switch to a MBP. Come to find out it wasn't a capability I really used. The only other OS I ran was Linux and Linux has been on ARM for some time.jcc said:This article paints a too rosy picture of the transition. The fact of the matter is that moving away from x86 will end Mac’s “best of both worlds” status. That means no more running Windows software.
Even more obvious is that ARM has never hurt software on iPhone nor iPad. Basically x86 is not a concern any more.
-
A14X Bionic allegedly benchmarked days before Apple Silicon Mac event
While certainly possible I don’t see them building in encoder logic for every need out there. Given that the current hardware codecs will be enhanced.anonconformist said:
If Apple wants to get max video encoding/decoding power in a machine as an important optimization, it makes far more sense to add specialized hardware to accelerate that, rather than adding more general purpose CPU cores: dedicated logic doesn’t need to have the overhead of Out-of-Order instruction execution and branch prediction logic the CPU would have, and this can make it faster (less distance in circuitry to travel) as well as use less power (speculative execution eats up several times as much power as strictly in-order execution).wizard69 said:
The top tier requires more cores. This processor would be great in a Mac Book, Mac Book Air and maybe even a 13" MBP. However the types of professionals Apple caters to can always use more cores, so I can't see Apple going with this specific processor in the MBP 16" and even the 13 inch is a maybe. Video production and compression/encoding can almost always use more cores and that is exactly what pros will want in the large Mac Book Pros. If Apple has any sense at all they will put as many cores in their chips as they can while keeping under the 45 and 35 watt power profiles.jumpingcoco said:this result is something that I would never expected early this year when leakers said there's gonna be new Apple Silicon-based Macs, I thought it would be MacBook Air or 13 MBP the most, but looks like they could easily make top-tier 16-inch MBP with this horsepower. Absolutely amazing!
In any event this is why I can see this particular chip going into a number of laptop Macs but not the high end laptops. Embarrassing as it is, I have to admit I'm excited about the possibilities with Apple Silicon. The combination of Neural Engine, high performance cores, high density 5nm chips and far better GPU's, will really lead to a new generation of Macs. In fact it is the lack of leaks about Neural Engine that have me worried a bit because AI is the wave of the future and Apple really needs that as a default feature in Mac OS. A boost over current Neural Engine performance could make these Macs the goto for people doing machine learning. I'd like to see Apple own this niche like they have owned content creation.
I believe Apple’s special card for MacPros is designed to do exactly this, and I’ve forgotten how old it is. 5 nm process node didn’t exist when it first came out. They’ve likely got a more optimized revision. Of course, it’d also make sense for Apple to expect people to buy MacPros instead to get that hardware: it has a cost to add, and chips can’t be infinitely large.This actually is a very interesting question that was brought up on another site, which amounts to “what does Apple do with all of that potential die space. I’m thinking that they will go much deeper into the world of die space for ML/AI. However putting in a full blown hardware encoder, replicating the Mac Pro option certainly is possible. The problem here is that they are then trying to hit a moving target.Which brings us back to lots of cores. Lots of cores leaves software selection up to the user. The user doesn’t have to use Apple favored solution. While maybe not as obvious there are many other uses where cores are the right solution, developers are just one example. Beyond that on the desktop they need to compete with the AMD solutions that are kicking every bodies but. AMD currently supplies 32 threads on low end and 128 for the Thread ripper series. Those systems blow everything away for cost and performance in highly parallel workloads. Due to Apples volume demands and the need to fix the Mac Pro I can see Apple going the chiplet route mirroring AMD. Well the concept anyways, Apple could really tailor an I/O die to their specific needs putting slower changing logic there. ARM and 5nm easily allows for 12 cores on a chiplet and a huge cache.Doing something like chiplets would allow for a 12 core 13” MBP with the larger MBP jumping to 24 cores. The low power cores could stay on the “I/O die” for an advertised 16 and 26 core option.Lots of speculation but that is what makes Apple Silicon so interesting. There are so many directions that Apple can go with here. It just keeps the imagination boiling over with thought. -
A14X Bionic allegedly benchmarked days before Apple Silicon Mac event
Looking at CPU benchmarks, of the traditional type, is likely to become a thing of the past as so much will be done outside of the traditional path through the CPU. My personal hope is that Apple is going all in on Machine Learning & AI technologies. It is interesting that there has been almost no leaks with respect to the hardware that supports ML. What is cool in my mind is that Macs will now get the same dedicated hardware for things like cameras that iOS gets. Maybe not huge but again it takes a load off the CPU cores. It looks like we are getting an expanded video Codec unit too.d_2 said:The semi obvious thing that isn’t being said here is that Apple has *complete control* of the processor design and OS for the new Macs, like they have had for iOS devices... thus, besides the usual compute logic within the CPU, there are all of the other optimizations that they are building into the silicon for Big Sur to take advantage of - many of which are known / have been explained... but, what else might Apple be planning?
So what else is Apple planning - I have no idea (can't read minds). I would not be surprised at all if they adopt the ARM / Fujitsu vector processing capabilities. That would be really impressive if they fully implemented, I'm just not sure if they could do that in a low power chip effectively. In any event I totally agree Apple doesn't care so much about the ARM cores per say, but rather access to that silicon. This years Apple Silicon might not even reveal their long term intentions.
