More power with less: Apple's A13 Bionic is faster and more power efficient

2

Comments

  • Reply 21 of 46
    What I don’t understand is why the Pros are gaining so much hours of extra battery life, while the iPhone 11 is only gaining 1 hour. It’s the same A13 SOC right?
    Or is the smartness of regulating power requirements a ‘pro’ feature that’s not built into the non-pro iPhone 11?
    Or are there other reasons for the Pros gaining more hours of battery life? Like an optimised screen (like smart refresh rates), that Apple didn’t talk about?
    More place for bigger battery?
    It indeed seems like the reason. According to these articles: http://bit.ly/2I4JN9w and bit.ly/2UMrbQY the Pros are quite larger and heavier than the XS of last year. While the XR was already quite heavier and larger last year. 
    edited September 2019 watto_cobra
  • Reply 22 of 46
    What I don’t understand is why the Pros are gaining so much hours of extra battery life, while the iPhone 11 is only gaining 1 hour. It’s the same A13 SOC right?
    Or is the smartness of regulating power requirements a ‘pro’ feature that’s not built into the non-pro iPhone 11?
    Or are there other reasons for the Pros gaining more hours of battery life? Like an optimised screen (like smart refresh rates), that Apple didn’t talk about?
    The biggest power draw on any phone is the display. Last year's flagship phones were using OLED displays for the first time. This year they've figured out a number of ways to minimise the power required for the OLED display.

    Was it due to Apple's software and hardware teams? Was it improvements in the component from the supplier (Samsung)? Most likely a mixture of both.

    But the improvements in the chip's energy efficiency are pretty much 100% Apple.
    edited September 2019 muthuk_vanalingamwatto_cobra
  • Reply 23 of 46
    Anandtech mentions that the new OLED screens are 15% more efficient that the XS/XS Max versions; that accounts for quite a bit of the battery life increase,
    PickUrPoisonwatto_cobra
  • Reply 24 of 46
    mobird said:
    it seems as though Apple is moving further from announcing anything technical at all. I don’t like that direction. They didn’t go over the last year, as usual, because Cook said that they had so much to talk about, but still ended short of the usual two hours, and well short of last year’s.
    I find this observation very interesting and would like to hear others from the AI community provide some insight to why this is occurring as well. Let's hear your thoughts please.
    Lopping off the "updates" section at the beginning is a welcome change to me. I didn't really notice less "tech" talk per se, especially since they had a member of the chip engineering team come out and talk about the A13 in a fair amount of detail. 
    radarthekatwatto_cobra
  • Reply 25 of 46
    Really impressive especially the power savings and switching of active areas, Kudos to the design team.
    The new Ad13 iMac will be a blast.
    But Teslas FSD chip is currently 12 times faster (73 TOPS vs A13 6 TOPS?) which makes clear Tesla has a design team (only a few people I understood) that can easily match Apples.
    So I expect much room for improvement for the A14 and its desktop version the Ad14 next year.
    Exciting times, it must be difficult working at Intel now.

    Edit: note the TOPS (not TFLOPS)
    edited September 2019
  • Reply 26 of 46
    melgrossmelgross Posts: 33,510member
    jtguy said:
    Anyone know how much RAM the new phones have? I know Apple doesn’t specify.
    I read that the 11 has 4, and the others 6.
    watto_cobra
  • Reply 27 of 46
    melgrossmelgross Posts: 33,510member

    tmay said:

    I was always under the impression that any CPU sent DC power (which I presume is defined as a 1 bit, or "power on") only to circuits that were actually intended to do anything. But I guess that's not true. Kinda stupid of chip designers to have power running through circuits when its not necessary. I guess they should have hired me decades ago, cause it's un-possible for me to be stupid.

    You're not appreciating the true complexity of these chips. It's not anywhere near as simple or as straightforward as you're describing it to be. There's always the possibility that they had the 'idea', but were unable to implement it, and each year they learn new things and better ways of doing things.

    It's interesting that Huawei is quoting 10.3 Billion transistors for their Kirin 990. but I'm thinking, that's a lot of brute force that isn't going to get them anywhere near the A13.
    That’s the version of the chip with the 5G modem built in. The other has a 4G modem, and fewer transistors. These modems are very complex. The 5G could easily have a few billion transistors on its own. Qualcomm’s is almost as big as the 855 SoC when used alone. In fact, qualcomm hasn’t yet integrated it into an SoC. huawei is the first with that.
    edited September 2019
  • Reply 28 of 46
    melgrossmelgross Posts: 33,510member

    Designing their own silicone was one of the best business decisions Apple made. And while true in the first gen iPhone it is amazing that it is at least as true today 
    The first chip to have some Apple IP was the A4.
    watto_cobra
  • Reply 29 of 46
    melgrossmelgross Posts: 33,510member
    What I don’t understand is why the Pros are gaining so much hours of extra battery life, while the iPhone 11 is only gaining 1 hour. It’s the same A13 SOC right?
    Or is the smartness of regulating power requirements a ‘pro’ feature that’s not built into the non-pro iPhone 11?
    Or are there other reasons for the Pros gaining more hours of battery life? Like an optimised screen (like smart refresh rates), that Apple didn’t talk about?
    The biggest power draw on any phone is the display. Last year's flagship phones were using OLED displays for the first time. This year they've figured out a number of ways to minimise the power required for the OLED display.

    Was it due to Apple's software and hardware teams? Was it improvements in the component from the supplier (Samsung)? Most likely a mixture of both.

    But the improvements in the chip's energy efficiency are pretty much 100% Apple.
    Apple has been claiming that the OLED displays use their own technology. I assume that’s in addition to what Samsung uses in producing the displays.
    watto_cobra
  • Reply 30 of 46
    gatorguygatorguy Posts: 24,176member
    What I don’t understand is why the Pros are gaining so much hours of extra battery life, while the iPhone 11 is only gaining 1 hour. It’s the same A13 SOC right?
    Or is the smartness of regulating power requirements a ‘pro’ feature that’s not built into the non-pro iPhone 11?
    Or are there other reasons for the Pros gaining more hours of battery life? Like an optimised screen (like smart refresh rates), that Apple didn’t talk about?
    I suspect reversing charging was originally in the plans for the Pro series 11's and dropped somewhat recently as not quite ready.  That would mesh with the rumors of it's there then it's not. The battery was already in production so now you get the benefit of much longer run times. 
    melgross
  • Reply 31 of 46
    melgrossmelgross Posts: 33,510member
    gatorguy said:
    What I don’t understand is why the Pros are gaining so much hours of extra battery life, while the iPhone 11 is only gaining 1 hour. It’s the same A13 SOC right?
    Or is the smartness of regulating power requirements a ‘pro’ feature that’s not built into the non-pro iPhone 11?
    Or are there other reasons for the Pros gaining more hours of battery life? Like an optimised screen (like smart refresh rates), that Apple didn’t talk about?
    I suspect reversing charging was originally in the plans for the Pro series 11's and dropped somewhat recently as not quite ready.  That would mesh with the rumors of it's there then it's not. The battery was already in production so now you get the benefit of much longer run times. 
    That’s an interesting take that I hadn’t thought about. It’s very possible that the extra battery life was at least partly aimed at the reverse charging feature that didn’t show up.
    watto_cobra
  • Reply 32 of 46
    knowitall said:
    Really impressive especially the power savings and switching of active areas, Kudos to the design team.
    The new Ad13 iMac will be a blast.
    But Teslas FSD chip is currently 12 times faster (73 TOPS vs A13 6 TOPS?) which makes clear Tesla has a design team (only a few people I understood) that can easily match Apples.
    So I expect much room for improvement for the A14 and its desktop version the Ad14 next year.
    Exciting times, it must be difficult working at Intel now.

    Edit: note the TOPS (not TFLOPS)
    Tesla's chip is much faster than Apple's at running, say, resnet 50. But it's MUCH slower than Apple's at running any normal app.

    You're failing to distinguish between different types of TOPS. Tesla's achievement is notable but not nearly in the same category as Apple's. In principle, it's not all that hard to add more TOPS, if you're talking about tensor/matrix ops, vectors, NNI, GPU, etc. Those ops (and OPS) are all easily parallelized. Further, you fail to recognize the hard limits placed by the power and cooling budgets for each chip. Tesla is entirely focused on video image recognition, so they need massive NN processing. They have the power budget of a car - not unlimited, by a long shot, but still... the battery in a Tesla is a little bigger than the battery in an iPhone! Whereas Apple is building a much more general-purpose chip, and specifically one with extraordinary traditional integer (and FP) OPS. That's a MUCH MUCH harder problem to solve, as it's extremely difficult to extract parallelism from conventional software (that is, pretty much every app that isn't doing AR or a few other very specific things).

    So far, Apple has in the last couple of years kicked *everyone's* ass at that, inside their domain (low-power chips). Nobody even comes close. And if you look carefully at what they've done, you can build a fairly convincing case that they've already built every part necessary to beat Intel at their own game (fast high-power multicore chips), they just don't want to sell those yet.

    The biggest open question is this: Can Apple build a ring/mesh/whatever connecting 8-16 high-performance cores in a reasonable power budget? As I've written previously about the A12X, they've *already done this*. So they can, right now today, build something competitive with Intel's best mainstream desktop CPU (the 8-core i9-9900). Whether or not they can actually bet it will depend on whether or not they can clock up. And we know more about that than we did a few months ago, as we can see AMD pushing the same process to around 4.3 GHz tops, about 4.1 comfortably. We still don't know if the A13's pipeline is long enough to sustain this sort of speed, or how easy it would be for Apple to change it enough, but the performance crown there seems easily within their grasp.

    Going to more cores is the biggest question mark. The ring or mesh good enough to handle 8 cores really well may not be enough to handle 12 or 16 cores. But the only machines where that would matter is the iMac Pro and the desktop Mac Pro. And I don't think anyone expects those to transition to ARM as early as the laptops.
    kevin keewatto_cobra
  • Reply 33 of 46
    melgross said:
    It’s interesting that Apple chose to save a greater percentage of power than they did to increase performance. It’s a choice. Next year they may decide to go the other way.
    Apple’s decision makes a lot of sense when you think about sustained performance. Anandtech’s review shows the A12 based phones (like many Android devices) struggle with that. Short bursts of peak performance are fine for bragging rights on short benchmarks, but higher sustained performance brings real usable gains.
    watto_cobra
  • Reply 34 of 46
    silvergold84silvergold84 Posts: 107unconfirmed, member
    Lime I wrote Apple is pure innovation and excellence 
    watto_cobra
  • Reply 35 of 46
    melgrossmelgross Posts: 33,510member
    knowitall said:
    Really impressive especially the power savings and switching of active areas, Kudos to the design team.
    The new Ad13 iMac will be a blast.
    But Teslas FSD chip is currently 12 times faster (73 TOPS vs A13 6 TOPS?) which makes clear Tesla has a design team (only a few people I understood) that can easily match Apples.
    So I expect much room for improvement for the A14 and its desktop version the Ad14 next year.
    Exciting times, it must be difficult working at Intel now.

    Edit: note the TOPS (not TFLOPS)
    Tesla's chip is much faster than Apple's at running, say, resnet 50. But it's MUCH slower than Apple's at running any normal app.

    You're failing to distinguish between different types of TOPS. Tesla's achievement is notable but not nearly in the same category as Apple's. In principle, it's not all that hard to add more TOPS, if you're talking about tensor/matrix ops, vectors, NNI, GPU, etc. Those ops (and OPS) are all easily parallelized. Further, you fail to recognize the hard limits placed by the power and cooling budgets for each chip. Tesla is entirely focused on video image recognition, so they need massive NN processing. They have the power budget of a car - not unlimited, by a long shot, but still... the battery in a Tesla is a little bigger than the battery in an iPhone! Whereas Apple is building a much more general-purpose chip, and specifically one with extraordinary traditional integer (and FP) OPS. That's a MUCH MUCH harder problem to solve, as it's extremely difficult to extract parallelism from conventional software (that is, pretty much every app that isn't doing AR or a few other very specific things).

    So far, Apple has in the last couple of years kicked *everyone's* ass at that, inside their domain (low-power chips). Nobody even comes close. And if you look carefully at what they've done, you can build a fairly convincing case that they've already built every part necessary to beat Intel at their own game (fast high-power multicore chips), they just don't want to sell those yet.

    The biggest open question is this: Can Apple build a ring/mesh/whatever connecting 8-16 high-performance cores in a reasonable power budget? As I've written previously about the A12X, they've *already done this*. So they can, right now today, build something competitive with Intel's best mainstream desktop CPU (the 8-core i9-9900). Whether or not they can actually bet it will depend on whether or not they can clock up. And we know more about that than we did a few months ago, as we can see AMD pushing the same process to around 4.3 GHz tops, about 4.1 comfortably. We still don't know if the A13's pipeline is long enough to sustain this sort of speed, or how easy it would be for Apple to change it enough, but the performance crown there seems easily within their grasp.

    Going to more cores is the biggest question mark. The ring or mesh good enough to handle 8 cores really well may not be enough to handle 12 or 16 cores. But the only machines where that would matter is the iMac Pro and the desktop Mac Pro. And I don't think anyone expects those to transition to ARM as early as the laptops.
    Apple has already gone to 4 cores in the A12x, 5 cores when counting the efficiency cores together. I don’t see why they can’t remove some unneeded sections from the chips that duplicate functions that don’t need to be duplicated, and run two of these, I suppose now, A13x chips together. Apple has the ability to do it however they think best, as they control the IP.
    watto_cobra
  • Reply 36 of 46
    melgrossmelgross Posts: 33,510member

    techconc said:
    melgross said:
    It’s interesting that Apple chose to save a greater percentage of power than they did to increase performance. It’s a choice. Next year they may decide to go the other way.
    Apple’s decision makes a lot of sense when you think about sustained performance. Anandtech’s review shows the A12 based phones (like many Android devices) struggle with that. Short bursts of peak performance are fine for bragging rights on short benchmarks, but higher sustained performance brings real usable gains.
    What’s interesting about the A series performance during “throttling” is that their performance is still better than the unthrottled performance of their competitors. This is something you don’t read about when their throttling is mentioned, often, gleefully. So the Kirin 980 and the Qualcomm 855 don’t throttle as much, percentage wise, but Apple’s chip performs better even when it does. Anandtech’s testing shows that quite clearly.  It Apple seems to want to eliminate this “perceived” inadequacy, even though it doesn’t really exist.

    and I’ve pointed out here, and in other places, that I find it funny that both AMD’s chips and Intel’s chips throttle too, but then it’s called “turbo mode”. It’s all marketing. 
    watto_cobra
  • Reply 37 of 46
    wizard69wizard69 Posts: 13,377member
    knowitall said:
    Really impressive especially the power savings and switching of active areas, Kudos to the design team.
    The new Ad13 iMac will be a blast.
    But Teslas FSD chip is currently 12 times faster (73 TOPS vs A13 6 TOPS?) which makes clear Tesla has a design team (only a few people I understood) that can easily match Apples.
    So I expect much room for improvement for the A14 and its desktop version the Ad14 next year.
    Exciting times, it must be difficult working at Intel now.

    Edit: note the TOPS (not TFLOPS)
    I know some want to dismiss it but this is innovation.  

    As for TOPS the two chips are not really comparable.  Apple will put improvements where it makes sense for the direction they are going in.    For a desktop chip I would expect them to focus more silicon on the GPU, much like the iPad.   There would likely be a secondary focus on cache and ARM cores.  
    watto_cobra
  • Reply 38 of 46
    wizard69wizard69 Posts: 13,377member
    melgross said:

    techconc said:
    melgross said:
    It’s interesting that Apple chose to save a greater percentage of power than they did to increase performance. It’s a choice. Next year they may decide to go the other way.
    Apple’s decision makes a lot of sense when you think about sustained performance. Anandtech’s review shows the A12 based phones (like many Android devices) struggle with that. Short bursts of peak performance are fine for bragging rights on short benchmarks, but higher sustained performance brings real usable gains.
    What’s interesting about the A series performance during “throttling” is that their performance is still better than the unthrottled performance of their competitors. This is something you don’t read about when their throttling is mentioned, often, gleefully. So the Kirin 980 and the Qualcomm 855 don’t throttle as much, percentage wise, but Apple’s chip performs better even when it does. Anandtech’s testing shows that quite clearly.  It Apple seems to want to eliminate this “perceived” inadequacy, even though it doesn’t really exist.

    and I’ve pointed out here, and in other places, that I find it funny that both AMD’s chips and Intel’s chips throttle too, but then it’s called “turbo mode”. It’s all marketing. 
    Beyond that Intel and AMD have significant power issues so one isn’t likely to see their chips in a watch or cell phone.  I’ve come to believe that Intel is very misleading with their power numbers. If one designs around their TDP you will never see good performance above base clock rate.  
    watto_cobra
  • Reply 39 of 46
    melgross said:
    knowitall said:
    Really impressive especially the power savings and switching of active areas, Kudos to the design team.
    The new Ad13 iMac will be a blast.
    But Teslas FSD chip is currently 12 times faster (73 TOPS vs A13 6 TOPS?) which makes clear Tesla has a design team (only a few people I understood) that can easily match Apples.
    So I expect much room for improvement for the A14 and its desktop version the Ad14 next year.
    Exciting times, it must be difficult working at Intel now.

    Edit: note the TOPS (not TFLOPS)
    Tesla's chip is much faster than Apple's at running, say, resnet 50. But it's MUCH slower than Apple's at running any normal app.

    You're failing to distinguish between different types of TOPS. Tesla's achievement is notable but not nearly in the same category as Apple's. In principle, it's not all that hard to add more TOPS, if you're talking about tensor/matrix ops, vectors, NNI, GPU, etc. Those ops (and OPS) are all easily parallelized. Further, you fail to recognize the hard limits placed by the power and cooling budgets for each chip. Tesla is entirely focused on video image recognition, so they need massive NN processing. They have the power budget of a car - not unlimited, by a long shot, but still... the battery in a Tesla is a little bigger than the battery in an iPhone! Whereas Apple is building a much more general-purpose chip, and specifically one with extraordinary traditional integer (and FP) OPS. That's a MUCH MUCH harder problem to solve, as it's extremely difficult to extract parallelism from conventional software (that is, pretty much every app that isn't doing AR or a few other very specific things).

    So far, Apple has in the last couple of years kicked *everyone's* ass at that, inside their domain (low-power chips). Nobody even comes close. And if you look carefully at what they've done, you can build a fairly convincing case that they've already built every part necessary to beat Intel at their own game (fast high-power multicore chips), they just don't want to sell those yet.

    The biggest open question is this: Can Apple build a ring/mesh/whatever connecting 8-16 high-performance cores in a reasonable power budget? As I've written previously about the A12X, they've *already done this*. So they can, right now today, build something competitive with Intel's best mainstream desktop CPU (the 8-core i9-9900). Whether or not they can actually bet it will depend on whether or not they can clock up. And we know more about that than we did a few months ago, as we can see AMD pushing the same process to around 4.3 GHz tops, about 4.1 comfortably. We still don't know if the A13's pipeline is long enough to sustain this sort of speed, or how easy it would be for Apple to change it enough, but the performance crown there seems easily within their grasp.

    Going to more cores is the biggest question mark. The ring or mesh good enough to handle 8 cores really well may not be enough to handle 12 or 16 cores. But the only machines where that would matter is the iMac Pro and the desktop Mac Pro. And I don't think anyone expects those to transition to ARM as early as the laptops.
    Apple has already gone to 4 cores in the A12x, 5 cores when counting the efficiency cores together. I don’t see why they can’t remove some unneeded sections from the chips that duplicate functions that don’t need to be duplicated, and run two of these, I suppose now, A13x chips together. Apple has the ability to do it however they think best, as they control the IP.
    That's not how it works. You don't see why, because you don't design chips for a living. The story is both better and worse than you think.

    About core count: Since the efficiency and fast cores can (and do!) all run simultaneously, Apple has with the A12X demonstrated that whatever they're using to connect all those cores (almost certainly a ring bus, but just possibly some sort of more complex mesh) is capable of handling not just 4 or 5, but 8 cores. "Efficiency" or not, the bus has to handle the same kind of work- all cores have to have cache coherency, equal access to main RAM, etc. Doing this at low enough power, with that many cores, is the big trick that will be key to winning on the desktop - and they've managed to do it well enough to work in an iPad. That's very impressive. We don't know if that architecture will extend to more cores than that, and that's an open question. It may be completely inappropriate for more than 8 cores. But still, 8 cores will get you a VERY long way today.

    Now, you talked about trimming "unneeded sections" from the chips. That's not likely to help very much. Most of those "unneeded sections" probably do not participate on the bus/mesh on an equal basis with the scalar cores, because why would they bother? However, nobody knows for sure, because Apple doesn't tell, and Andrei over at AT (the only person I know of who's gained deep insight into the chips and published about them) only has so much time to go poking at the innards with clever software and more clever analysis. So it's vaguely possible that they already have a big-time mesh architecture or multi-ring-bus (like newer and older XCC Xeons, respectively) already, which would be amazing. But it's very unlikely, as the power draw would be incredibly difficult to deal with.

    Lastly, it's not at all simple to "run two chips together". If you're thinking about 2S systems like typical Xeons... then you need significant logic to get them to play nicely together and with RAM and the rest of the system. And you'd need to do really major surgery on the A12/13/whatever. On the other hand, if you're thinking about a chiplet setup like AMD's... then it's the same deal with slightly different details. In both cases, you don't know that the secret sauce Apple's using will carry over well. For example, one of the biggest factors in the massive speedups seen in the A12 is apparently the cache architecture and the large L3. If you took that out and stuck it in a central chiplet (like AMD ZEN) you'd probably take a massive perf hit. This is all moderately wild speculation, but the point is, it's not a slam dunk. I personally believe that if the decide to do it, they will embarrass the crap out of everybody else. But I'm skeptical that they'll bother any time soon.

    tl;dr: As I've said before, they *already* have shown the ability to go neck-and-neck with Intel's top mainstream chips. If they wanted to fight over the HEDT and Xeons, they could probably do a great job, but it seems unlikely that they'll bother in the near future. They've already got everything they need for every laptop segment, excepting only people who need x64 Windows (or Linux) virtualization at native speed.

    watto_cobra
  • Reply 40 of 46
    knowitall said:
    Really impressive especially the power savings and switching of active areas, Kudos to the design team.
    The new Ad13 iMac will be a blast.
    But Teslas FSD chip is currently 12 times faster (73 TOPS vs A13 6 TOPS?) which makes clear Tesla has a design team (only a few people I understood) that can easily match Apples.
    So I expect much room for improvement for the A14 and its desktop version the Ad14 next year.
    Exciting times, it must be difficult working at Intel now.

    Edit: note the TOPS (not TFLOPS)
    Tesla's chip is much faster than Apple's at running, say, resnet 50. But it's MUCH slower than Apple's at running any normal app.

    You're failing to distinguish between different types of TOPS. Tesla's achievement is notable but not nearly in the same category as Apple's. In principle, it's not all that hard to add more TOPS, if you're talking about tensor/matrix ops, vectors, NNI, GPU, etc. Those ops (and OPS) are all easily parallelized. Further, you fail to recognize the hard limits placed by the power and cooling budgets for each chip. Tesla is entirely focused on video image recognition, so they need massive NN processing. They have the power budget of a car - not unlimited, by a long shot, but still... the battery in a Tesla is a little bigger than the battery in an iPhone! Whereas Apple is building a much more general-purpose chip, and specifically one with extraordinary traditional integer (and FP) OPS. That's a MUCH MUCH harder problem to solve, as it's extremely difficult to extract parallelism from conventional software (that is, pretty much every app that isn't doing AR or a few other very specific things).

    So far, Apple has in the last couple of years kicked *everyone's* ass at that, inside their domain (low-power chips). Nobody even comes close. And if you look carefully at what they've done, you can build a fairly convincing case that they've already built every part necessary to beat Intel at their own game (fast high-power multicore chips), they just don't want to sell those yet.

    The biggest open question is this: Can Apple build a ring/mesh/whatever connecting 8-16 high-performance cores in a reasonable power budget? As I've written previously about the A12X, they've *already done this*. So they can, right now today, build something competitive with Intel's best mainstream desktop CPU (the 8-core i9-9900). Whether or not they can actually bet it will depend on whether or not they can clock up. And we know more about that than we did a few months ago, as we can see AMD pushing the same process to around 4.3 GHz tops, about 4.1 comfortably. We still don't know if the A13's pipeline is long enough to sustain this sort of speed, or how easy it would be for Apple to change it enough, but the performance crown there seems easily within their grasp.

    Going to more cores is the biggest question mark. The ring or mesh good enough to handle 8 cores really well may not be enough to handle 12 or 16 cores. But the only machines where that would matter is the iMac Pro and the desktop Mac Pro. And I don't think anyone expects those to transition to ARM as early as the laptops.
    Thanks for pointing that out (good info!).
    I must admit my posting was a bit of a click bait.
    I hinted at that with my remark about TOPS vs TFLOPS (not all OPS are created equally)
    So I beg to differ about ‘failing to point out’ in any way.

    I agree with most remarks you make.
    Only, feature size of the Tesla FSD chip is only 14nm, whiles A13 is build on 7nm.
    This means that on the same nm scale FSD would be 2 to 4 times as fast and with the same number of transistors almost 8 times. This is almost an order of magnitude faster still, incredible.

    Your remark about connecting cores is interesting, I do believe that Apple is quite capable in that regard, they have a history creating such things. I do believe that the current setup scales to a lot of cores.
    I also think having a less elaborate pipeline with less depth can be an advantage instead of a problem. Sustaining the processors instruction speed can also be improved by the way the software is compiled, for example by replicating code at some places so pre fetching works out all the time. Also having ultra fast (code) memory is an important factor.

    Parallelization of code is difficult at low (machine language) level but a good processor design should do this with its pipeline and instruction level parallelism laid out in its core. I don't see any problems in this regard for Apples Arm cores. Higher level parallelism which gives the most speedup, if the algorithm implemented is inherently parallel that is, can be created by using Apples GCD library which enables the programmer to be focused on the parallel abstraction while the scheduling of the jobs and the number cores (threads) used is done automatically.
    So I think Apple is quite ahead in this respect and that makes having a high number of cores for desktop processing a sensible thing to do.
    So a 16 or even 32 core Ad14 iMac pro isn't out of the question I think.

    As I stated before, kudos to the Apple processor design team. They are very capable and I expect great things in the future.


    watto_cobra
Sign In or Register to comment.