4 ms·

If it is, maybe related to LTT's live stream yesterday that used it? https://youtu.be/jtHhvcdjNWw https://youtu.be/jtHhvcdjNWw

JumpCrisscross 8y ago
Counter argument is the suite of tests he used were optimised for the i7 and not the i9 [1]. [1] https://www.computerworld.com/article/3291760/apple-mac/about-that-macbook-pro-core-i9-throttling-story.html https://www.computerworld.com/article/3291760/apple-mac/abou...
- simcop2387 8y ago
  The problem with that is that, while yes it means that performance can be improved (faster rendering times, more computations per second etc.) it doesn't change the thermal issues at all with the particular implementation from apple.
- seabrookmx 8y ago
  While it's true optimizing the software could increase render times, the fact that the CPU does run regularly below the base clock with heavy workloads (ie: throttles) remains true. I know everyone likes to bash Apple at every opportunity but stating the fact that the i9 model throttles is _not_ inflammatory. The better question is, how much does it matter? The machine still has 6-cores. MKBHD was still noticing a 15% performance increase (vs. last gen) without new, updated builds of Final cut etc.
  - simcop2387 8y ago
    Exactly, and that's why the freezer test mentioned in the article the grandparent linked to was done. To try to get an idea of how much the throttling was affecting.
    - Jnr 8y ago
      It doesn't matter if it is faster until it starts throttling. i9 is intended for pros who need that CPU power for longer periods of time. And if the issue appears after a minute of usage and performance becomes worse than i7 (as shown in initial Dave's video), then it's a bad product (Macbook, because of not being able to cool it, not the i9 itself). And benchmarks should be changed to run for longer time periods or else it's not showing actual performance indication.
      - matthewmacleod 8y ago
        It doesn't matter if it is faster until it starts throttling. i9 is intended for pros who need that CPU power for longer periods of time What's the evidence for this? I'd love a faster multi-core processor, but only to accelerate tasks which complete in a relatively small amount of time already.
        simcop2387 8y ago
        Intel's own marketing talks about it being used by these kinds of people https://newsroom.intel.com/news/intel-core-i9-processor-comes-mobile-best-gaming-creation-laptop-processor-intel-ever-built/ https://newsroom.intel.com/news/intel-core-i9-processor-come...
      - seabrookmx 8y ago
        The optimization argument is that even after throttling, if the software was optimized for the 6core machines it would no longer be slower than the i7 after one minute. But that's yet to be seen. As of this moment in time you're correct.. buying the i9 model makes zero sense.
  - threeseed 8y ago
    This doesn't get posted enough. But the issue seems to be that Apple favours a quieter and cooler machine over performance. As such they run the fans a lot slower and later than they could. So if you manually adjust the fan speeds (plenty of apps that do this) you can get significantly (10-15%) better performance.
- stefan_ 8y ago
  What counter argument is this? The i9 is a 6-core i7. That's basically all.
  - yoz-y 8y ago
    What is the actual 6-core i7 then? https://www.apple.com/shop/buy-mac/macbook-pro/15-inch-space-gray-2.6ghz-6-core-512gb# https://www.apple.com/shop/buy-mac/macbook-pro/15-inch-space...
- userbinator 8y ago
  How does changing the benchmarking software in any way not invalidate the comparison, when trying to compare two compatible hardware implementations? To see if the hardware has gotten faster, you must run the same software on it. IMHO this "grasping at straws" is a sign of the end of actual performance improvements. From the start of the PC through perhaps the mid 2000s you could run the same software, and each new generation of processor would show a measurable (and often dramatic) improvement.
- Jnr 8y ago
  That is not a counter argument! If software is optimized for multiple cores then it is optimized for multiple cores. It is not like they tested it on a single core CPU with higher frequency first. It could be that i9 might even have new instruction set with faster instructions but it should still have the same ones that i7 have as well.

JumpCrisscross 8y ago

Counter argument is the suite of tests he used were optimised for the i7 and not the i9 [1]. [1] https://www.computerworld.com/article/3291760/apple-mac/about-that-macbook-pro-core-i9-throttling-story.html https://www.computerworld.com/article/3291760/apple-mac/abou...

simcop2387 8y ago
The problem with that is that, while yes it means that performance can be improved (faster rendering times, more computations per second etc.) it doesn't change the thermal issues at all with the particular implementation from apple.
seabrookmx 8y ago
While it's true optimizing the software could increase render times, the fact that the CPU does run regularly below the base clock with heavy workloads (ie: throttles) remains true. I know everyone likes to bash Apple at every opportunity but stating the fact that the i9 model throttles is _not_ inflammatory. The better question is, how much does it matter? The machine still has 6-cores. MKBHD was still noticing a 15% performance increase (vs. last gen) without new, updated builds of Final cut etc.
- simcop2387 8y ago
  Exactly, and that's why the freezer test mentioned in the article the grandparent linked to was done. To try to get an idea of how much the throttling was affecting.
  - Jnr 8y ago
    It doesn't matter if it is faster until it starts throttling. i9 is intended for pros who need that CPU power for longer periods of time. And if the issue appears after a minute of usage and performance becomes worse than i7 (as shown in initial Dave's video), then it's a bad product (Macbook, because of not being able to cool it, not the i9 itself). And benchmarks should be changed to run for longer time periods or else it's not showing actual performance indication.
    - matthewmacleod 8y ago
      It doesn't matter if it is faster until it starts throttling. i9 is intended for pros who need that CPU power for longer periods of time What's the evidence for this? I'd love a faster multi-core processor, but only to accelerate tasks which complete in a relatively small amount of time already.
      - simcop2387 8y ago
        Intel's own marketing talks about it being used by these kinds of people https://newsroom.intel.com/news/intel-core-i9-processor-comes-mobile-best-gaming-creation-laptop-processor-intel-ever-built/ https://newsroom.intel.com/news/intel-core-i9-processor-come...
    - seabrookmx 8y ago
      The optimization argument is that even after throttling, if the software was optimized for the 6core machines it would no longer be slower than the i7 after one minute. But that's yet to be seen. As of this moment in time you're correct.. buying the i9 model makes zero sense.
- threeseed 8y ago
  This doesn't get posted enough. But the issue seems to be that Apple favours a quieter and cooler machine over performance. As such they run the fans a lot slower and later than they could. So if you manually adjust the fan speeds (plenty of apps that do this) you can get significantly (10-15%) better performance.
stefan_ 8y ago
What counter argument is this? The i9 is a 6-core i7. That's basically all.
- yoz-y 8y ago
  What is the actual 6-core i7 then? https://www.apple.com/shop/buy-mac/macbook-pro/15-inch-space-gray-2.6ghz-6-core-512gb# https://www.apple.com/shop/buy-mac/macbook-pro/15-inch-space...
userbinator 8y ago
How does changing the benchmarking software in any way not invalidate the comparison, when trying to compare two compatible hardware implementations? To see if the hardware has gotten faster, you must run the same software on it. IMHO this "grasping at straws" is a sign of the end of actual performance improvements. From the start of the PC through perhaps the mid 2000s you could run the same software, and each new generation of processor would show a measurable (and often dramatic) improvement.
Jnr 8y ago
That is not a counter argument! If software is optimized for multiple cores then it is optimized for multiple cores. It is not like they tested it on a single core CPU with higher frequency first. It could be that i9 might even have new instruction set with faster instructions but it should still have the same ones that i7 have as well.