Despite some if the thinking here, the article does express some of what I’ve been reading elsewhere that shows impressive possibilities. Never count Intel out.
No one has ever doubted Intel’s ability to demonstrate. What they doubt is Intel’s ability to stick to a timeline and deliver on it.
It seems not so shocking piecemeal improvements of things (except maybe 3D stacking chip technology) already implemented in the A chip range. Apple can make a big move and make all its (MAC) products $200 to $300 cheaper. Even MS seems to port it’s OS and presumably a lot of its legacy software to ARM which removes the last ‘but but but ...’ from the list.
Mmmm.
Okay, but I wouldn’t assume that using their own processor is going to make their devices cheaper. They’ll be looking to recoup development costs for one thing.
Sunny Cove is a fundamentally wider core than most every Intel design since Sandy Bridge, as well, expanding from four-wide issue to five-wide and increasing the number of execution ports from eight to 10. Each of those execution units, in turn, is more capable than those of Skylake. Intel added a dedicated integer divider on port 1 to reduce latency for those operations.
Interesting to contrast
Monsoon (A11) was a major microarchitectural update in terms of the mid-core and backend. It’s there that Apple had shifted the microarchitecture in Hurricane (A10) from a 6-wide decode from to a 7-wide decode. The most significant change in the backend here was the addition of two integer ALU units, upping them from 4 to 6 units.
Monsoon (A11) and Vortex (A12) are extremely wide machines – with 6 integer execution pipelines among which two are complex units, two load/store units, two branch ports, and three FP/vector pipelines this gives an estimated 13 execution ports, far wider than Arm’s upcoming Cortex A76 and also wider than Samsung’s M3. In fact, assuming we're not looking at an atypical shared port situation, Apple’s microarchitecture seems to far surpass anything else in terms of width, including desktop CPUs.
-AT
Intel going wider at last is great, it's crazy that the thing in my pants still beats it on width.
The one question I have is if Apple is going to continue to rely upon Intel's roadmap, or have they had enough and go and make their own processors for their Mac lineup? My thinking is that this is too little, too late and Apple will just make their own chips.
I think it would be ignorant to believe that Apple does not have a non-Intel solution in the works to replace Intel. Too many people think that ARM's capabilities are on-par with Intel's which is flat-out wrong.
I myself hope that whatever Apple does, it does license x86 like what AMD was able to pull off. x86 still rules the world and I do end up using Windows on my Mac quite often.
x86 is dead, MS does a transition of Windows to ARM
It seems not so shocking piecemeal improvements of things (except maybe 3D stacking chip technology) already implemented in the A chip range. Apple can make a big move and make all its (MAC) products $200 to $300 cheaper. Even MS seems to port it’s OS and presumably a lot of its legacy software to ARM which removes the last ‘but but but ...’ from the list.
Apple's big move would be Zen 2 and Vega 2 platform wide and see their CPUs cost between 150-300% less than Intel.
That would be a very bad idea, considering the A12.
I’m more interested in if the A series will move to 3D stacking.
For a while now, chip design takes advantage of turning off & on certain parts of the chips to maximize performance.
For example, turbo charging for single/dual core then turning off the the other cores to reduce heat. It will be interesting to see if 3D stacking makes a difference. My guess is it improves multi core performances/tasks. That sounds like significant graphs performance improvements...
The A Series will never compete with chip designs from AMD or Intel. It's just not ever going to happen.
Not such a smart remark since the A12 is at the moment clearly superior to anything else. A desktop version of the chip the A12D will be awe inspiring.
Apple can make a big move and make all its (MAC) products $200 to $300 cheaper.
ha ha ha ha ha. There is no way Apple would pass any savings on to the user. Cheaper for them to build, maybe. They’d keep the extra profit. That’s the mission: all profit, all the time. No public company serves the consumers. Their real client is shareholders & “investors”.
You never know when this comes in handy. Maybe when import tariffs go up? Trump?
... but if you stack the processors, the heat from one processor will affect the performance of the processors above or below it. AMD places their processors next to each other so they can cool all of them at the same time (and also run all of them at the same time).
Interesting they’re starting with Xeons. Traditionally don’t they always start with low power chips and move to desktops and midrange mobiles and end with Xeons a year or more later?
No. They used to start with big, expensive chips to help quickly recoup their R&D and horrid expenses of new process plants. But as things got smaller, and new processes were pushed back, with lower yields, they began to make small chips first, because the yields were higher, and then moved to larger chips once the process was stabilized.
Despite some if the thinking here, the article does express some of what I’ve been reading elsewhere that shows impressive possibilities. Never count Intel out.
No one has ever doubted Intel’s ability to demonstrate. What they doubt is Intel’s ability to stick to a timeline and deliver on it.
Well, tht]Arstechnica a jaundiced view. It reminds me of the views that those who do t like Apple tend to hold about them. When all was thought to be lost against and back in the early 2000’s, Intel roared back, almost destroying and in the process, as and blundered around with Bulldozer. Are people already forgetting about that?
Intel is stuck possibly because they are trying to make a proper 10nm process with transistor density that actually matches 10nm, and. Not 14 or 16 as their competitors have been doing. It’s obviously much harder to do, and so Intel seems to be behind. But if they get their process working in late 2o19, they will have lost a year, competitively. The point is that it doesn’t matter as much for them. It’s interesting to note that AMD has recently come out with a 7nm chip design, which seems to be about at the same level, core for core, to what Intel has been doing on their 14+++ process for over a year now. What does that tell you?
its the mobile market that’s grabbing 7nm now, really equivalent to 10-12nm from Intel. Intel isn’t doing mobile chips like that, even with the M series, so they’re less affected.
La seule question que je pose est de savoir si Apple va continuer à utiliser la feuille de route ou si elles ont assez et veulent créer leurs propres processeurs pour leur gamme Mac. Ma pensée est que c'est trop peu, trop tard et Apple vont juste faire leurs propres puces.
En cas de succès, the power engine, on a lui-même, un avantage majeur pour les fabricants d'appareils. A titre d’exemple, a été petit processeur aider à réduire une carte mère, permettant ainsi de réduire la taille, la taille et la masse tout en augmentant la capacité de la batterie.
I’m more interested in if the A series will move to 3D stacking.
For a while now, chip design takes advantage of turning off & on certain parts of the chips to maximize performance.
For example, turbo charging for single/dual core then turning off the the other cores to reduce heat. It will be interesting to see if 3D stacking makes a difference. My guess is it improves multi core performances/tasks. That sounds like significant graphs performance improvements...
Apple has been using multi-chip stacking (not a new idea) on A-series packages for about 4 generations now. A typical A-series package includes the SoC (CPU, GPU, ISP, etc on a single die) and the DRAM stacked on top of the SoC in the package.
If they are going to stack something else in the package for a third layer, my bet would be NAND chips.
For your latter comments about about turning off CPU cores, that can already be done at the chip level. Intel already does this and Apple probably does this to varying degrees.
Intel is saying that their micro solder bumps is the big innovation, enabling higher I/O between chips in the package.
Apple can make a big move and make all its (MAC) products $200 to $300 cheaper.
ha ha ha ha ha. There is no way Apple would pass any savings on to the user. Cheaper for them to build, maybe. They’d keep the extra profit. That’s the mission: all profit, all the time. No public company serves the consumers. Their real client is shareholders & “investors”.
You never know when this comes in handy. Maybe when import tariffs go up? Trump?
This guy you’re replying to is a student troll. He believes all corporations are the same. He hasn’t read essays about management teams that focus on “delighting the customer” firstly, vs “Maximizing Shareholder Value: The Dumbest Idea in the World”. He doesn’t even realize there’s a difference in corporate philosophies.
Apple can make a big move and make all its (MAC) products $200 to $300 cheaper.
ha ha ha ha ha. There is no way Apple would pass any savings on to the user. Cheaper for them to build, maybe. They’d keep the extra profit. That’s the mission: all profit, all the time. No public company serves the consumers. Their realm client is shareholders & “investors”.
What about how cheap Apple made the iPhoneX when they stopped using expensive QualComm modems?
Despite some if the thinking here, the article does express some of what I’ve been reading elsewhere that shows impressive possibilities. Never count Intel out.
No one has ever doubted Intel’s ability to demonstrate. What they doubt is Intel’s ability to stick to a timeline and deliver on it.
Well, tht]Arstechnica a jaundiced view. It reminds me of the views that those who do t like Apple tend to hold about them. When all was thought to be lost against and back in the early 2000’s, Intel roared back, almost destroying and in the process, as and blundered around with Bulldozer. Are people already forgetting about that?
Intel is stuck possibly because they are trying to make a proper 10nm process with transistor density that actually matches 10nm, and. Not 14 or 16 as their competitors have been doing. It’s obviously much harder to do, and so Intel seems to be behind. But if they get their process working in late 2o19, they will have lost a year, competitively. The point is that it doesn’t matter as much for them. It’s interesting to note that AMD has recently come out with a 7nm chip design, which seems to be about at the same level, core for core, to what Intel has been doing on their 14+++ process for over a year now. What does that tell you?
its the mobile market that’s grabbing 7nm now, really equivalent to 10-12nm from Intel. Intel isn’t doing mobile chips like that, even with the M series, so they’re less affected.
Despite some if the thinking here, the article does express some of what I’ve been reading elsewhere that shows impressive possibilities. Never count Intel out.
No one has ever doubted Intel’s ability to demonstrate. What they doubt is Intel’s ability to stick to a timeline and deliver on it.
Well, tht]Arstechnica a jaundiced view. It reminds me of the views that those who do t like Apple tend to hold about them. When all was thought to be lost against and back in the early 2000’s, Intel roared back, almost destroying and in the process, as and blundered around with Bulldozer. Are people already forgetting about that?
Intel is stuck possibly because they are trying to make a proper 10nm process with transistor density that actually matches 10nm, and. Not 14 or 16 as their competitors have been doing. It’s obviously much harder to do, and so Intel seems to be behind. But if they get their process working in late 2o19, they will have lost a year, competitively. The point is that it doesn’t matter as much for them. It’s interesting to note that AMD has recently come out with a 7nm chip design, which seems to be about at the same level, core for core, to what Intel has been doing on their 14+++ process for over a year now. What does that tell you?
its the mobile market that’s grabbing 7nm now, really equivalent to 10-12nm from Intel. Intel isn’t doing mobile chips like that, even with the M series, so they’re less affected.
Then maybe intel should just try to get to 12nm?
they probably should have done that a year ago.
The equipment is made for a specific process size. In order to change that, a lot of reinvestment needs to be made, and it takes a lot of time as well. It would make no sense to me.
Comments
Mmmm.
Okay, but I wouldn’t assume that using their own processor is going to make their devices cheaper. They’ll be looking to recoup development costs for one thing.
Interesting to contrast
-AT
Intel going wider at last is great, it's crazy that the thing in my pants still beats it on width.
Not such a smart remark since the A12 is at the moment clearly superior to anything else.
A desktop version of the chip the A12D will be awe inspiring.
hopefull, this is a good sign.
Intel is stuck possibly because they are trying to make a proper 10nm process with transistor density that actually matches 10nm, and. Not 14 or 16 as their competitors have been doing. It’s obviously much harder to do, and so Intel seems to be behind. But if they get their process working in late 2o19, they will have lost a year, competitively. The point is that it doesn’t matter as much for them. It’s interesting to note that AMD has recently come out with a 7nm chip design, which seems to be about at the same level, core for core, to what Intel has been doing on their 14+++ process for over a year now. What does that tell you?
its the mobile market that’s grabbing 7nm now, really equivalent to 10-12nm from Intel. Intel isn’t doing mobile chips like that, even with the M series, so they’re less affected.
If they are going to stack something else in the package for a third layer, my bet would be NAND chips.
For your latter comments about about turning off CPU cores, that can already be done at the chip level. Intel already does this and Apple probably does this to varying degrees.
Intel is saying that their micro solder bumps is the big innovation, enabling higher I/O between chips in the package.
Blocked and forgotten long ago.
Never mind that never happened.
they probably should have done that a year ago.