Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

M3 Macs: there's more to performance than counting cores

eclecticlight.co

Post reply on HN
Print view View original

171–180 of 419 posts

Re: M3 Macs: there's more to performance than counting cores

#171

Post by hajile »

_ph_ wrote: post #101
hajile wrote: post #60

Earlier quoted context omitted.

Actual IPC increase is 1-2%. The rest is from ramping the clockspeeds. This is a problem because power consumption goes up exponentially with frequency. Go up too high and they'll be doing what AMD or Intel does where a single core is using 50+ watts to hit those peak numbers.

How is this a problem? It just looks that they didn't try to improve the general architecture but make one step to ramp up the clock speeds without increasing the power consumption thanks to the process step. Which is a great achievement, because any tape out on a new advanced process - here for the first time a "3nm" - is a big achievement. One has to consider that Apple now has yearly updates in its processor lineu…

With each recent node step, you basically get +10-15% clockspeed or -30% power consumption. They just blew their entire node on a small clockspeed ramp.

Now, if they want a wider core for M4, that means more transistors and more heat. They are then forced to: not go wider, decrease max clockspeed, hold max clockspeed for a pitiful amount of time, or increase power consumption.

On the whole, I'd rather have a wider core and lower clockspeeds then turn the other power savings into either battery life or a few more E-cores.

Re: M3 Macs: there's more to performance than counting cores

#172

Post by speedgoose »

grecy wrote: post #79
speedgoose wrote: post #70

Earlier quoted context omitted.

Wow, that’s overpriced for such low specs.

Please show me a laptop from another manufacturer that has similar specs, similar screen, similar battery life that costs less.

It’s more that those specs are not worth the price bump. I have a M1 that is slightly less capable but cost much much less. I use it to access machines that are faster than this M3 laptop when needed.

Re: M3 Macs: there's more to performance than counting cores

#173

Post by twoodfin »

the8472 wrote: post #132
bloopernova wrote: post #98

I want a better system load metric now that we've got heterogenous cores in CPUs. The Pixel 8 Pro has 3 types, efficiency, performance, and one single ultra performance core. If your efficiency cores are always close to max usage, but your performance cores are idle, is your system being heavily or barely used? I understand that system load isn't really a useful metric between systems, but it's useful to compare on a…

linux has pressure stall information[0] which tracks the time where some processes couldn't be run because they were waiting for contended resources (cpu/io/memory). An N-parallel compute-bound job on an M-thread CPU will stay at nearly 0 CPU pressure if N=M (assuming no background tasks) because they're not stepping on each other's toes. At N>M pressure will start to rise. [0] https://docs.kernel.org/accounting/psi.…

Note that for memory, as I understand the documentation, it's less about "couldn't be run" than "couldn't have the real memory allocation necessary to avoid paging".

This is in contrast to another contended resource, memory bandwidth, "waiting" for which manifests at the process level as CPU cycles like any other code. It's possible to use profiling tools to distinguish memory waits from other CPU activity, but as far as I know nobody has built the infrastructure to bubble that data up in some form so it could be tracked systemically by the OS.

I can guess why: The memory hierarchy is complicated, and what you can derive about it from CPU performance counters is indirect and limited. Still, even some basic estimation for how many cache lines a process was responsible for driving over the memory bus would be helpful for those building high-performance systems and applications.

Re: M3 Macs: there's more to performance than counting cores

#174

Post by DesiLurker »

deagle50 wrote: post #52
cubefox wrote: post #3

This seems to be a lot of effort to rationalize the surprisingly small performance increase from M2 to M3. Initially the assumption was that M2 to M3 would be a bigger step than M1 to M2, not a smaller one. Perhaps TSMC 3nm is showing the limits of scaling?

The surface area of the M3 Pro vs M2 Pro tells you everything you need to know.

have a handy link and few more sentences for my friend who does not gets it?

Re: M3 Macs: there's more to performance than counting cores

#176

Post by wubrr »

toast0 wrote: post #168
wubrr wrote: post #136

Earlier quoted context omitted.

Not at all. As a single example - macos comes with an extremely outdated version of GNU bash (3.2) from 2007. And yes, this has consequences in that many modern bash scripts depend on bash 4+ (current latest GNU bash is 5.2+).

Using tools from the base OS is really holding it wrong. Just like ftp.exe/Internet Explorer/Edge and Safari should only be used to download a usable browser, the Apple provided CLI tools should only be used to download the tools and versions of tools that you actually want. Otherwise, you have no way to control versioning anyway. When the OS version is tied to the version of so many other tools, it's a nightmare. Ap…

> Apple provided CLI tools should only be used to download the tools and versions of tools that you actually want.

Why? Why do I need to use 3rd party package managers, or manual installations from 3rd party sources to install common tools that aren't a decade+ outdated?

> When the OS version is tied to the version of so many other tools, it's a nightmare.

It doesn't need to be tied to anything. If the OS needs specific libraries/tools then they can be installed in a separate location. If there are new major versions of user-level tools (like bash) those should come by default, not some 15 year old version of the same tool, with the same license. Multiple linux distros have solved these problems in different ways 10+ years ago.

> Apple is trying to push you in the right direction by basically not updating the CLI tools ever

No, they are just neglecting the CLI ecosystem, the package management, and the many many outstanding bugs that have existed and been ignored for years.

Re: M3 Macs: there's more to performance than counting cores

#177

Post by Tagbert »

Keyframe wrote: post #86
cubefox wrote: post #3

This seems to be a lot of effort to rationalize the surprisingly small performance increase from M2 to M3. Initially the assumption was that M2 to M3 would be a bigger step than M1 to M2, not a smaller one. Perhaps TSMC 3nm is showing the limits of scaling?

Everytime Apple oversells there are apologists putting a spin on it. Nothing new to see.

And there are detractors downplaying improvements. Nothing new to see.

Re: M3 Macs: there's more to performance than counting cores

#178

Post by Exoristos »

wubrr wrote: post #110
asylteltine wrote: post #67

And yet the air can’t support more than one monitor. I know Apple is a small company with only a few engineers so it must have been too hard. Sarcasm obviously. As an Apple enthusiast it’s so annoying how they sniff farts.

I mean, the support for even one monitor is by far worse than windows or linux IMO. - Monitor randomly resets/readjusts for no apparent reason. Like multiple times every hour. - Windows disappear, become inaccessible after moving between monitors, even though it still open and active according to doc/task list. - Why can't I move a window to a monitor/workspace that has a maximized window? Like, you can do it by un-m…

I've both used Macs for decades and at one time worked in enterprise Mac technical support, and I've never seen this stuff happen.

Re: M3 Macs: there's more to performance than counting cores

#179

Post by GeekyBear »

exabrial wrote: post #158

I think I'll hold out to the M4 and make sure memory bandwidth and latency is restored to it's previous value.

Given that Intel and AMD tend to choke the memory bandwidth on anything below their server chips, this seems short sighted.

The M3 Pro's "reduced memory bandwidth" is still double the memory bandwidth you see on Intel and AMD's HEDT chips.

Step up to the M3 Max and you're looking at five times the memory bandwidth of Intel and AMD's chips.

Re: M3 Macs: there's more to performance than counting cores

#180

Post by tuetuopay »

willsmith72 wrote: post #122
user_7832 wrote: post #81

Earlier quoted context omitted.

High refresh rates aren't just for games, they make the entire interface smoother. Things like frame rate or latency are quality of life things that you don't need , but once you upgrade, the difference becomes clear. Sort of like going from a slowly accelerating station wagon to a Porsche/sports car even when (capped by the speed limit) if you've experienced that. Hopefully someone else has a better analogy.

sounds like something I'd rather never want to get used to then. More expensive, uses more power so kills battery life faster and worse for the environment. Where's the win?

> uses more power so kills battery life faster

not necessarily. on promotion iPhones (fancy name for variable refresh rates), the panel is not stuck at 120Hz all the time. it varies from 1Hz when the screen is static up to 120Hz when it's getting animated, down to whatever framerate your content is at (e.g. movies and videos). it actually is a battery saver whenever the phone's screen is static (common on most apps with text) or with <60Hz content (youtube, movies, etc).

Post reply on HN