5 ms·

Tin foil hat warning - but how much of the M1 performance improvement is from optimizations made in Big Sur for Apple Silicon that they just didn't bother imple

Tin foil hat warning - but how much of the M1 performance improvement is from optimizations made in Big Sur for Apple Silicon that they just didn't bother implementing for x86 since it's now the outgoing technology for Apple? I realize the reducing in power consumed for any given quantity of work is downright amazing for laptops, but I guess I'm more curious about workstation and (build) server kinds of applications. Also, how many of these benchmarks are x86 versus Apple Silicon where both are running Big Sur. I've been seeing so much "Xcode on Catalina" versus "Xcode on Apple Silicon on Big Sur"

whynotminot 6y ago
Apple has been using x86 for practically 15 years now. Your theory is this whole time they've been too lazy to bother to implement "god mode" ? What about Windows and Linux on x86? Also too lazy? It can be fun to put on the tin foil hat, but at least make sure the theory makes even a little bit of sense first.
- lern_too_spel 6y ago
  > What about Windows and Linux on x86? Also too lazy? No, Windows and Linux are significantly faster. https://www.phoronix.com/scan.php?page=article&item=macos1015-win10-ubuntu&num=10 https://www.phoronix.com/scan.php?page=article&item=macos101... Ubuntu has gotten significantly faster since 19.10, and other distros like ClearLinux are much faster still. https://www.phoronix.com/scan.php?page=article&item=clear-fall-2020&num=8 https://www.phoronix.com/scan.php?page=article&item=clear-fa...
- Teknoman117 6y ago
  I was not saying in any way that they didn't implement optimizations for x86 over the time that they used them. Of course they did. What I meant was that Apple themselves talked about significant optimization for Apple Silicon in Big Sur. The question is if any of these optimizations could have also been applied to x86 but aren't because x86 is the outgoing platform. I'm skeptical of the assertion (without supporting documentation) that there is some hardware design choice in Apple Silicon that makes it _drastically_ more memory (quantity) efficient than x86 when using presumably the same kernel, same toolchain frameworks (LLVM), etc.
blihp 6y ago
That's basically what they did when they made the switch from PPC to x86. IIRC, there was an edict a few years before the switch (Jobs talked about it in an interview at the time) where teams were told something along the lines of not making any design/optimization decisions that were tied to PPC. They likely did the same thing with x86 this time.
- saagarjha 6y ago
  I’ve seen x86_64 improvements go into macOS as recently as last year, so I don’t think this is really true.
kllrnohj 6y ago
The better tinfoil hat question is how much did they intentionally sandbag the outgoing Intel models vs. how much was it a case that M1 was delayed? Such as the Intel MacBook Air whose fan wasn't even connected to anything? Was that bad design because it wasn't supposed to exist, or was that bad design intentional to help drive up the huge generation over generation gains?

whynotminot 6y ago

Apple has been using x86 for practically 15 years now. Your theory is this whole time they've been too lazy to bother to implement "god mode" ? What about Windows and Linux on x86? Also too lazy? It can be fun to put on the tin foil hat, but at least make sure the theory makes even a little bit of sense first.

lern_too_spel 6y ago
> What about Windows and Linux on x86? Also too lazy? No, Windows and Linux are significantly faster. https://www.phoronix.com/scan.php?page=article&item=macos1015-win10-ubuntu&num=10 https://www.phoronix.com/scan.php?page=article&item=macos101... Ubuntu has gotten significantly faster since 19.10, and other distros like ClearLinux are much faster still. https://www.phoronix.com/scan.php?page=article&item=clear-fall-2020&num=8 https://www.phoronix.com/scan.php?page=article&item=clear-fa...
Teknoman117 6y ago
I was not saying in any way that they didn't implement optimizations for x86 over the time that they used them. Of course they did. What I meant was that Apple themselves talked about significant optimization for Apple Silicon in Big Sur. The question is if any of these optimizations could have also been applied to x86 but aren't because x86 is the outgoing platform. I'm skeptical of the assertion (without supporting documentation) that there is some hardware design choice in Apple Silicon that makes it _drastically_ more memory (quantity) efficient than x86 when using presumably the same kernel, same toolchain frameworks (LLVM), etc.

blihp 6y ago

That's basically what they did when they made the switch from PPC to x86. IIRC, there was an edict a few years before the switch (Jobs talked about it in an interview at the time) where teams were told something along the lines of not making any design/optimization decisions that were tied to PPC. They likely did the same thing with x86 this time.

saagarjha 6y ago
I’ve seen x86_64 improvements go into macOS as recently as last year, so I don’t think this is really true.

kllrnohj 6y ago

The better tinfoil hat question is how much did they intentionally sandbag the outgoing Intel models vs. how much was it a case that M1 was delayed? Such as the Intel MacBook Air whose fan wasn't even connected to anything? Was that bad design because it wasn't supposed to exist, or was that bad design intentional to help drive up the huge generation over generation gains?