4 ms·

Regarding the Process–architecture–optimization model from intel, what do each of those upgrades mean to the user? I believe Process has the most to deal with

Regarding the Process–architecture–optimization model from intel, what do each of those upgrades mean to the user? I believe Process has the most to deal with energy efficiency due to the shrinkage, which is why M1 was so energy efficient when it was launched. M2 seems like an architecture/optimization change, seems like they are just able to cram more stuff which is why it's faster without increasing battery life. For those familiar with Intel, what does the consumer mainly gain out of optimization product launches?

msbarnett 4y ago
> I believe Process has the most to deal with energy efficiency due to the shrinkage, which is why M1 was so energy efficient when it was launched. Eh, that's part of it, but a lot of it has to do with the M1 having very very high IPC (much higher than comparable x86-64 parts), meaning they could run the chip at a much lower max clock speed (3.2 GHz versus boosting to 5 GHz for most competitive x86-64 CPUs) for similar overall performance. This makes a huge difference because power consumption increases exponentially with clock speed. edit: thinking about how it relates to Intel's process–architecture–optimization, it feels a bit tricky to compare. Apple's process seems to be something like: architecture(A-series chip for iPhones)-optimize-with-new-die-reusing-basic-cores-in-diff-arrangement-targetting-perf-in-small-thermal-envelope(M1 for iPad Pro and small Macbooks)-optimize-again-with-another-new-die-reusing-basic-cores-but-in-yet-another-arrangement-targetting-perf-in-a-wider-thermal-envelope(M1 Pro/Max/Ultra), and that all happens before you get to the next M-series increment, which begins with an A-series increment. So the M2 is less the optimization of the M1 than it is the re-use of the cores in the new A-series chip preceding it, which was an architectural change, plus optimizations and other SoC differences.
- megous 4y ago
  > This makes a huge difference because power consumption increases exponentially with clock speed. It scales linearly with clock speed.
  - kec 4y ago
    P = V^2 * f, yes, but V and f are correlated. You can't increase f without also increasing V, hence: exponential scaling.
    - wormslayer666 4y ago
      That's polynomial scaling, not exponential.
    - WithinReason 4y ago
      That's still a polynomial, not an exponential
  - msbarnett 4y ago
    These graphs look linear to you? https://blog.stuffedcow.net/2012/07/ivy-bridge-power-consumption/ https://blog.stuffedcow.net/2012/07/ivy-bridge-power-consump... https://qph.cf2.quoracdn.net/main-qimg-61820a6b86759c354d41dbbd16a0385b https://qph.cf2.quoracdn.net/main-qimg-61820a6b86759c354d41d...
    - megous 4y ago
      These graphs look like they depict f only scaling?
      - msbarnett 4y ago
        Who said anything about f? I said "power consumption increases exponentially with clock speed" not "power consumption increases exponentially with f". That any modern processor has to crank up V to crank up its clock speed is a given. Hence, power consumption is exponential with clock speed (which is achieved by cranking up both V - the polynomial term in P – and f). (Which, fine, if you want to be pedantic, is polynomial growth, but that changes nothing about the point, which is that you have to burn a shitload more power at 5 GHz than at 3.2 GHz, because your consumption isn't scaling anything close to linear).

msbarnett 4y ago

> I believe Process has the most to deal with energy efficiency due to the shrinkage, which is why M1 was so energy efficient when it was launched. Eh, that's part of it, but a lot of it has to do with the M1 having very very high IPC (much higher than comparable x86-64 parts), meaning they could run the chip at a much lower max clock speed (3.2 GHz versus boosting to 5 GHz for most competitive x86-64 CPUs) for similar overall performance. This makes a huge difference because power consumption increases exponentially with clock speed. edit: thinking about how it relates to Intel's process–architecture–optimization, it feels a bit tricky to compare. Apple's process seems to be something like: architecture(A-series chip for iPhones)-optimize-with-new-die-reusing-basic-cores-in-diff-arrangement-targetting-perf-in-small-thermal-envelope(M1 for iPad Pro and small Macbooks)-optimize-again-with-another-new-die-reusing-basic-cores-but-in-yet-another-arrangement-targetting-perf-in-a-wider-thermal-envelope(M1 Pro/Max/Ultra), and that all happens before you get to the next M-series increment, which begins with an A-series increment. So the M2 is less the optimization of the M1 than it is the re-use of the cores in the new A-series chip preceding it, which was an architectural change, plus optimizations and other SoC differences.

megous 4y ago
> This makes a huge difference because power consumption increases exponentially with clock speed. It scales linearly with clock speed.
- kec 4y ago
  P = V^2 * f, yes, but V and f are correlated. You can't increase f without also increasing V, hence: exponential scaling.
  - wormslayer666 4y ago
    That's polynomial scaling, not exponential.
  - WithinReason 4y ago
    That's still a polynomial, not an exponential
- msbarnett 4y ago
  These graphs look linear to you? https://blog.stuffedcow.net/2012/07/ivy-bridge-power-consumption/ https://blog.stuffedcow.net/2012/07/ivy-bridge-power-consump... https://qph.cf2.quoracdn.net/main-qimg-61820a6b86759c354d41dbbd16a0385b https://qph.cf2.quoracdn.net/main-qimg-61820a6b86759c354d41d...
  - megous 4y ago
    These graphs look like they depict f only scaling?
    - msbarnett 4y ago
      Who said anything about f? I said "power consumption increases exponentially with clock speed" not "power consumption increases exponentially with f". That any modern processor has to crank up V to crank up its clock speed is a given. Hence, power consumption is exponential with clock speed (which is achieved by cranking up both V - the polynomial term in P – and f). (Which, fine, if you want to be pedantic, is polynomial growth, but that changes nothing about the point, which is that you have to burn a shitload more power at 5 GHz than at 3.2 GHz, because your consumption isn't scaling anything close to linear).