News Apple's A17 Pro Challenges Core i9-13900K, Ryzen 7950X in Single-Core Performance

[Admin]

Apple's A17 Pro scores 15% higher than A16 Bionic in Geekbench 6 CPU performance test.

Apple's A17 Pro Challenges Core i9-13900K, Ryzen 7950X in Single-Core Performance : Read more

 

[Order 66]

It will be interesting to see how the A17 pro performs in RT if/when mobile games release supporting it. My prediction is that it will do well but I don't really see the point of RT for a phone soc because the chip is going to have to use so much more power to use RT that battery life probably won't be great when using RT.

 

[gg83]

This seems like an advertisement and what's this chip "challenging"? Useless benchmark?

 

[teodoreh]

In other news, RTX4070's single CUDA core performance is identical to RTX4090's!

 

[hotaru251]

to be fair I will say one thing in defense of the intel/amd cpu's....They are always going to be less "efficient" due to how much mroe complex the pipeline for any windows/linux (not sure on macs as havent used one in 25+yrs) OS is. Same reason consoles can do so well vs pc even if specs are vastly different that much less complex pipeline.


when your device has a lot less stuff to run in background and hoops to jump through you get better performance.


I don't like Apple as a company nor would I use their stuff due to that, but as a technology their chips probably best designed than any other.

 

[Makaveli]

I remember all the noise Apple made about the M1 that also had good ST performance but also a node advantage. Once AMD and Intel moved to newer nodes that advantage went away. And even if you compare apple now on 3nm still behind AMD on 5nm and intel on its current.

 

[zecoeco]

Apple should care less about synthetic benchamarks and more about real-world performance that is achieveable through the optimizations and uniformity of their ecosystem. Hardware is way ahead, while software is crappled, unoptimized and unable to utilize this performance effectively. Winning the "watts" war isn't going to benefit anyone but the "hardcore" people out there looking for the smallest yet most expensive drop of perfomance at the cost of power and price.

 

[JamesJones44]

to be fair I will say one thing in defense of the intel/amd cpu's....They are always going to be less "efficient" due to how much mroe complex the pipeline for any windows/linux (not sure on macs as havent used one in 25+yrs) OS is. Same reason consoles can do so well vs pc even if specs are vastly different that much less complex pipeline.

macOS is Unix based (OpenBSD to be exact), the "pipeline" is pretty much the same as Linux (In fact it can run almost all the same things Linux can). It's also still runs on Intel based Macs.

One can argue that Apple has tighter integration and maybe using custom OP codes to optimize their software on their own chips, but I don't think generic OS pipelines are the reason why they are more/less efficient.

 

[JamesJones44]

I remember all the noise Apple made about the M1 that also had good ST performance but also a node advantage. Once AMD and Intel moved to newer nodes that advantage went away. And even if you compare apple now on 3nm still behind AMD on 5nm and intel on its current.

Your comparing an 8.5 watt TDP part to a 100+ watt TDP part. In no world is that an apples to apples comparison and not even worth mentioning.

 

[Makaveli]

Your comparing an 8.5 watt TDP part to a 100+ watt TDP part. In no world is that an apples to apples comparison and not even worth mentioning.

I know its apples vs orange but the article itself has the comparison so why not point it out.

 

[The Historical Fidelity]

macOS is Unix based (OpenBSD to be exact), the "pipeline" is pretty much the same as Linux. This has nothing to do with the chips efficiency.

You can argue that Apple has tighter integration and maybe using custom OP codes to optimize their software on their own chips, but generic OS reasons are not why they are more/less efficient.

I think what he is trying to get at is the inefficiency of x86 architecture with all of its useless legacy micro operations. Every x86 design has to provide hardware that can execute this legacy code and this both takes up valuable space and prevents micro-architects from fully optimizing logic features without having to compromise to ensure legacy features will still work.

Intel has said they plan to remove useless legacy operations from x86 to improve its performance and allow full optimization of logic units for modern operations.

 

[JamesJones44]

I know its apples vs orange but the article itself has the comparison so why not point it out.

Fair enough. The article's comparison is drawing a comparison that doesn't really make much sense.

 

[JamesJones44]

I think what he is trying to get at is the inefficiency of x86 architecture with all of its useless legacy micro operations. Every x86 design has to provide hardware that can execute this legacy code and this both takes up valuable space and prevents micro-architects from fully optimizing logic features without having to compromise to ensure legacy features will still work.

Intel has said they plan to remove useless legacy operations from x86 to improve its performance and allow full optimization of logic units for modern operations.

That part I agree with. I was just disagreeing with the idea that macOS was somehow less complex, when in reality it's basically identical to Linux in kernel and system level complexity.

 

[TJ Hooker]

I think what he is trying to get at is the inefficiency of x86 architecture with all of its useless legacy micro operations. Every x86 design has to provide hardware that can execute this legacy code and this both takes up valuable space and prevents micro-architects from fully optimizing logic features without having to compromise to ensure legacy features will still work.

Intel has said they plan to remove useless legacy operations from x86 to improve its performance and allow full optimization of logic units for modern operations.

They also said this doesn't apply to consoles, but current and last gen consoles are also x86. So it's still not clear what they were referring to.

And the changes Intel is proposing would just remove some steps in the boot sequence and unused modes/options that only exist to maintain an interface in common with ancient hardware (and associated motherboard FW and OSs). There's nothing to indicate it would improve performance, or that maintaining this compatibility until now hindered the hardware. Assuming this is what you are referring to: https://www.intel.com/content/www/u...visioning-future-simplified-architecture.html

 

[hotaru251]

I was just disagreeing with the idea that macOS was somehow less complex

i did say I couldnt say anything about modern Mac :|

They also said this doesn't apply to consoles, but current and last gen consoles are also x86.

misunderstood me.

the console pipeline and the windows piepline are different.
Windows for example has a lot of handshakes for permissions etc.

where as console doesnt need them (so it uses less resources to execute stuff)
and graphics are also done in a much less demanding way (example is pc & say switch dont even do graphics same way so if u tried emulate you'd have worse performace with same spec of hardware)

a eli5 of my original post would be arm and x86 arent apples to apples but apples to oranges.

 

[gg83]

In other news, RTX4070's single CUDA core performance is identical to RTX4090's!

The more I thought about what you said the more I laughed out loud! Lmao

 

[gg83]

I know its apples vs orange but the article itself has the comparison so why not point it out.

Exactly

 

[spongiemaster]

Your comparing an 8.5 watt TDP part to a 100+ watt TDP part. In no world is that an apples to apples comparison and not even worth mentioning.

There are so many asterisks tied to this statement that it pretty much lacks any factual basis. Even the most power hungry single x86 cores don't use remotely close to 100W or even half that.

 

[rtoaht]

Wow, 3% multi traded Gen and gen performance improvement in Apple’s favorite benchmark! Amazing!! Thanks for the ad.

 

[TJ Hooker]

i did say I couldnt say anything about modern Mac :|


misunderstood me.

the console pipeline and the windows piepline are different.
Windows for example has a lot of handshakes for permissions etc.

where as console doesnt need them (so it uses less resources to execute stuff)
and graphics are also done in a much less demanding way (example is pc & say switch dont even do graphics same way so if u tried emulate you'd have worse performace with same spec of hardware)

a eli5 of my original post would be arm and x86 arent apples to apples but apples to oranges.

Xbox OS is Windows-based, and it uses a form of DirectX for graphics API. I think you're overestimating how different it is.

What exactly are you referring to by "handshakes for permissions" in Windows?

 

[bit_user]

I remember all the noise Apple made about the M1 that also had good ST performance but also a node advantage. Once AMD and Intel moved to newer nodes that advantage went away. And even if you compare apple now on 3nm still behind AMD on 5nm and intel on its current.

Apple's cores were never optimized for single-thread performance, which is what made it so impressive that the M1 could actually challenge desktop x86 CPUs, on that metric!

Yeah, when AMD and Intel launched Zen 4 and Alder Lake, they indeed retook a solid lead... but they are still easily outclassed by Apple on perf/W, and that's what Apple's CPUs were truly optimized for.

Winning the "watts" war isn't going to benefit anyone but the "hardcore" people out there looking for the smallest yet most expensive drop of perfomance at the cost of power and price.

It sure matters for phone & laptop users, which are Apple's main markets.

The article's comparison is drawing a comparison that doesn't really make much sense.

I disagree, especially with the apparent rise in popularity of mini-PCs with limited cooling capacity.

There are so many asterisks tied to this statement that it pretty much lacks any factual basis. Even the most power hungry single x86 cores don't use remotely close to 100W or even half that.

That's not true. The current Intel & AMD flagships can easily exceed 50 W on a single-threaded workload.

 

[bit_user]

macOS is Unix based (OpenBSD to be exact), the "pipeline" is pretty much the same as Linux (In fact it can run almost all the same things Linux can). It's also still runs on Intel based Macs.

The relationship between BSD and MacOS X is more tenuous and superficial than that, from what I've read. I'm not an expert on the subject, but I think not much of MacOS X hews from its BSD roots.

One can argue that Apple has tighter integration and maybe using custom OP codes to optimize their software on their own chips,

The only custom opcodes were really just their AMX matrix-multiply extensions. Otherwise, it's implementing straight ARMv8.n-A ISA.

Intel has said they plan to remove useless legacy operations from x86 to improve its performance and allow full optimization of logic units for modern operations.

APX does a lot to close the gap, but it can only do so much without breaking backwards compatibility.

the changes Intel is proposing would just remove some steps in the boot sequence and unused modes/options that only exist to maintain an interface in common with ancient hardware (and associated motherboard FW and OSs). There's nothing to indicate it would improve performance, or that maintaining this compatibility until now hindered the hardware. Assuming this is what you are referring to: https://www.intel.com/content/www/u...visioning-future-simplified-architecture.html

I think removing some of those legacy addressing modes could conceivably benefit parts of the microarchitecture which have to integrate support for them (e.g. the instruction decoders), even though they're basically unused.

APX is the bigger improvement. By my count, it addresses about 3 of the 5 distinct advantages ARM's ISA holds over x86:

Introducing Intel® Advanced Performance Extensions (Intel® APX)

These extensions expand the x86 instruction set with access to registers and features that improve performance.

www.intel.com

No CPU has been announced which will feature it. So, we probably won't see it before 2025, at the earliest.

 

[thestryker]

The biggest thing to me is that this is the fourth year of releases on effectively the same core. Was hoping this would be something new, but it seems like next year at the earliest. They've made all sorts of GPU/connectivity changes, but not much on the CPU side.

 

[The Historical Fidelity]

They also said this doesn't apply to consoles, but current and last gen consoles are also x86. So it's still not clear what they were referring to.

And the changes Intel is proposing would just remove some steps in the boot sequence and unused modes/options that only exist to maintain an interface in common with ancient hardware (and associated motherboard FW and OSs). There's nothing to indicate it would improve performance, or that maintaining this compatibility until now hindered the hardware. Assuming this is what you are referring to: https://www.intel.com/content/www/u...visioning-future-simplified-architecture.html

The white paper synopsis you linked is for engineers so generic phrases like “improved and/ or hindered performance” will not be found.
You need to understand micro-architectures in order to parse the effects on efficiency these x86s changes will impart on future micro-architectures.
If you read the actual white paper (not the synopsis you linked) the levels of removal and change in functionality allows several functional units to be redesigned for efficient execution of modern code.

https://cdrdv2-public.intel.com/776648/x86s-EAS-v1-4-17-23-1.pdf

 

[George³]

How does modern code differ from code written 5 years ago, or even more? During this time, I don't think there are any general changes, at the very bottom of the processor architectures and in how the execution of instructions is ordered and implemented.