Geekbench score for the Intel i9-13900K is out and it’s not looking pretty for Apple …

[Colstan]

"Science project" is apt. It's a mashup of far-out ideas. Each is interesting in the academic sense, as in it might have made sense for a researcher to write a paper about the idea and figure out whether it had any merit. Instead of doing something like that, Intel just jammed a bunch of these ideas into their ISA of the future and pushed it into production without doing the homework to rigorously check whether any of it was likely to work out.

It seems that, often times, when Intel gets into trouble it's because they are trying to experiment too much with a shipping product. I believe the rumor for why they were stuck on 14nm++++++ for so long was because they tried to cram too much into 10nm, while TSMC does far more incremental tweaks. Arc is running into problems because they decided to compete with AMD and Nvidia's top-end with their first mass-market release. They've stuffed all sorts of gimmicks into the drivers without making sure the basics even work properly.

Moon shot products are fine for companies like Google that have plenty of funds to waste, and perhaps one-in-ten will work out and be successful. As long as the search engine works, they can do as they please. When Intel does it with some of its most important products it blows up in their face.

 

[diamond.g]

It seems that, often times, when Intel gets into trouble it's because they are trying to experiment too much with a shipping product. I believe the rumor for why they were stuck on 14nm++++++ for so long was because they tried to cram too much into 10nm, while TSMC does far more incremental tweaks. Arc is running into problems because they decided to compete with AMD and Nvidia's top-end with their first mass-market release. They've stuffed all sorts of gimmicks into the drivers without making sure the basics even work properly.

Moon shot products are fine for companies like Google that have plenty of funds to waste, and perhaps one-in-ten will work out and be successful. As long as the search engine works, they can do as they please. When Intel does it with some of its most important products it blows up in their face.

To be clear/fair Intel isn't competing with Nvidia and AMD top end GPU's they have admitted that they cannot do that until Celestial is released.

 

[pdoherty]

If it has to be in a Windows machine, then to me it doesn't matter what specs it has. I'm not in the market to ever purchase or care about buying anything that runs that crappy software.

I don't know... this guy makes some excellent points that show where Windows is better than Mac OS. A lot of inconsistent behaviors in Mac, and annoying behaviors.

 

[pshufd]

I don't know... this guy makes some excellent points that show where Windows is better than Mac OS. A lot of inconsistent behaviors in Mac, and annoying behaviors.

His take on multi-monitor systems is odd.

I far prefer macOS precisely for how macOS handles multiple monitors. One of the most annoying things about Windows multiple monitor handling is that you can't use virtual independent desktops. When you change desktops, it changes the desktop on all of your monitors at the same time.

A lot of his complaints are about specific software or how to do things. There are lots of annoying things on Windows too. One of the biggest was their Windows 11 rollout. I went to do the installation and it bricked my Windows desktop. It took me four hours to get it back up and running.

 

[Jumpthesnark]

I don't know... this guy makes some excellent points that show where Windows is better than Mac OS. A lot of inconsistent behaviors in Mac, and annoying behaviors.

He's free to like and use the machines he wants to, of course. Like everyone. I've used Windows machines and I find they hurt my productivity in a way the Mac doesn't, so I'm not really interested in his points. Also clickbaity YT videos aren't really my thing.

 

[pshufd]

He's free to like and use the machines he wants to, of course. Like everyone. I've used Windows machines and I find they hurt my productivity in a way the Mac doesn't, so I'm not really interested in his points. Also clickbaity YT videos aren't really my thing.

There are so many things that Linux does well that you can't do on Windows and macOS yet I don't use Linux as my daily driver. I find that the macOS desktop just works a lot better for me despite me having a large choice of operating systems and hardware. I have flip-flopped back and forth between Windows, Windows + macOS and macOS. At the moment, my desktop is all macOS.

I have written tools to make Windows more comfortable and there are lots of tools out there but you can't change the base architecture. One of the things that I like most about macOS is that it is based on Unix and the familiar Unix commands and tools are readily available in Terminal - just as I'd expect from Linux, Solaris and even OEL.

 

[Colstan]

To be clear/fair Intel isn't competing with Nvidia and AMD top end GPU's they have admitted that they cannot do that until Celestial is released.

I should have clarified that they were targeting 3070 performance, very much intentional hyperbole on my part, but Arc is thus far nowhere near that level of performance. It's fine being clear, but I'm starting to think like our friend Cliff Maier; Intel doesn't deserve fairness because they are in a state of constant delays and mismanagement.

Meteor Lake, Sapphire Rapids and Arc have all been delayed, showing a pattern with the company. I don't want Intel to fail, they're doing that on their own. Intel is too large to go under, plus they've got government handouts to pay for stock dividends, so there's that, too. I just see very little technical leadership or innovation from them, which is frustrating, because they had been an innovator for decades.

 

[pshufd]

I should have clarified that they were targeting 3070 performance, very much intentional hyperbole on my part, but Arc is thus far nowhere near that level of performance. It's fine being clear, but I'm starting to think like our friend Cliff Maier; Intel doesn't deserve fairness because they are in a state of constant delays and mismanagement.

Meteor Lake, Sapphire Rapids and Arc have all been delayed, showing a pattern with the company. I don't want Intel to fail, they're doing that on their own. Intel is too large to go under, plus they've got government handouts to pay for stock dividends, so there's that, too. I just see very little technical leadership or innovation from them, which is frustrating, because they had been an innovator for decades.

But their new chip can run at 6.2 Ghz!

The air temp is 97 degrees where I am. Probably a lot hotter in other parts of the country as we're in a cold climate.

 

[Colstan]

But their new chip can run at 6.2 Ghz!

I assume that's Intel's "Plan B" chip, Molten Lake. Its secondary purpose serves as a space heater specifically designed for tropical areas with high electricity bills.

 

[pshufd]

Intel screwing up and affecting Apple's production plans:

Intel reportedly planned to outsource production of its Meteor Lake tGPU chipset to TSMC, with mass production scheduled for the second half of 2022, before being delayed to the first half of 2023 due to design and verification issues. Now, Intel is said to have delayed mass production to the end of 2023, virtually cancelling the 3nm chip production capacity that it had booked with TSMC for most of next year.

As a result, TSMC has apparently been "greatly affected" by the move, forcing it to slow its expansion of 3nm chip production to ensure that production capacity "is not excessively idle, leading to massive cost amortization pressure."

Apple is believed to be the main customer of TSMC's initial 3nm chip mass production. According to the report, Apple is now the only major company among the first wave of 3nm chip production clients with orders scheduled between the second half of 2022 and the start of 2023. Apple's upcoming 3nm chips reportedly include new M-series chips and the "A17 Bionic."

Apple Chip Supplier Hit by Intel Delay Ahead of 'A17 Bionic' Production in 2023

Apple chip supplier TSMC is scaling back its plans to produce 3nm chips next year after Intel postponed a major order, Taiwanese research firm...

www.macrumors.com

So Meteor Lake is delayed a year? Way to go Intel!

 

[pshufd]

I assume that's Intel's "Plan B" chip, Molten Lake. Its secondary purpose serves as a space heater specifically designed for tropical areas with high electricity bills.

Good thing that Ice Lake was already taken.

 

[pdoherty]

I don't get why people care so much about benchmarks. I mean end of the day you can have the beefiest processor in the world but if the "lower scoring" one still feels and runs better at everyday task then what is the point? Its not like everyone who buys their compute equipment actually needs it to be 1 zillion single score performing but eating 1 billion watts per second you know.

Why would an objectively slower computer “feel and run better at everyday tasks” though?

 

[pshufd]

I got a question on Reddit asking between a 2014 MacBook Pro 13 and a Dell system. I used Geekbench 5 multicore scores to compare these systems with the M1. I personally find Geekbench 5 multicore an accurate reflection of the amount of work that a CPU can do, at least for what I do. It's nice to quantify what an old, slow CPU is like compared to something current and it's something that users can identify with. It can be difficult to get across how bad some CPUs are.

 

[Sydde]

Why would an objectively slower computer “feel and run better at everyday tasks” though?

Because speed is primarily subjective to most users.

 

[leman]

Why would an objectively slower computer “feel and run better at everyday tasks” though?

Because objectively measuring performance is like objectively trying to determine whether someone is a “good person” . Things get complicated in real world compared to a benchmarking suite that tries to “guess” popular workloads.

One part of the problem is that conventional wisdom defines performance as throughput (how much work can be done in a given unit of time). And while this definition works well most of the time, it can also lead to weird results like a bus being faster than a sports car. Many user-facing workloads are all about latency (response time), not throughput. Take the speedometer wen browser test for example. It simulates user interaction with a web app and measures how fast it can respond. Fastest desktop Intel processor (i9-12900K) scores 280 points on that test, which is very very fast. M2 MacBook Air scores… 400.

Another thing is that these benchmarks show the throughput under ideal conditions. They pretty much represent the peak performance you can get when the starts align correctly. Take that Dell XPS laptop off the charger and see the performance tank. Apples performance on the other hand is consistent - you get what you measure. That’s energy efficiency in practice.

And one can go on For instance, I’m quite sure that M2 Air will build software faster than any Alder Lake P-series laptop. Et cetera.

 

[pshufd]

Why would an objectively slower computer “feel and run better at everyday tasks” though?

We have Geekbench 5 single-core and multicore. Single-core generally reflects responsiveness while multicore better represents work that you can get done. So you could have a score of 1,500 for single-core and 3,000 for multicore and another CPU have a single-core of 1,000 and a multicore of 7,000. The first would feel more responsive in day-to-day tasks but the latter would be able to get large and long workloads done more quickly.

In actual reality, responsiveness doesn't mean you get more work done. It just feels more immediate and some translate that to better. Getting a response a few milliseconds or even a hundred milliseconds doesn't translate to more work done.

Sometimes even doing tricks like smoothing out response makes things feel faster or smoother even if they actually take longer.

 

[throAU]

Source

The i9-13900K chip will be out later this year and we now have Geekbench results. Single core: 2133 and Multi core: 23701

In comparison, the M2 in the new MacBook Pro scored: 1919 in single core
8929 in multi core.

Sure, Apple is much better at performance per watt than Intel but it’s not a good look to fall behind in single core performance. Most day to day tasks are single core.

Apple upended the chip industry with the M1 but AMD and Intel came back swinging and it seems like Apple now needs to pull another rabbit out of the hat with the M3.

The i9-13900k is intel's best, to hell with power consumption, DESKTOP model that is not yet shipping.

The m2 is Apple's new 2022 base model mobile laptop processor, and shipping RIGHT NOW.

 

[pshufd]

The i9-13900k is intel's best, to hell with power consumption, DESKTOP model that is not yet shipping.

The m2 is Apple's new 2022 base model mobile laptop processor, and shipping RIGHT NOW.

I'd argue that the M1 is still the chip to beat, even closing in on being two years old. Try finding M1 minis on the refurb store - it's pretty challenging and there are often no models or only oddball configurations that aren't popular. It seems like people are even willing to buy the 10 GB models to get something close to the configurations that they want. It's interesting to see such demand for a two-year-old model.

I have the mini and love it for performance and quiet.

 

[pdoherty]

Because objectively measuring performance is like objectively trying to determine whether someone is a “good person” . Things get complicated in real world compared to a benchmarking suite that tries to “guess” popular workloads.

One part of the problem is that conventional wisdom defines performance as throughput (how much work can be done in a given unit of time). And while this definition works well most of the time, it can also lead to weird results like a bus being faster than a sports car. Many user-facing workloads are all about latency (response time), not throughput. Take the speedometer wen browser test for example. It simulates user interaction with a web app and measures how fast it can respond. Fastest desktop Intel processor (i9-12900K) scores 280 points on that test, which is very very fast. M2 MacBook Air scores… 400.

Another thing is that these benchmarks show the throughput under ideal conditions. They pretty much represent the peak performance you can get when the starts align correctly. Take that Dell XPS laptop off the charger and see the performance tank. Apples performance on the other hand is consistent - you get what you measure. That’s energy efficiency in practice.

And one can go on For instance, I’m quite sure that M2 Air will build software faster than any Alder Lake P-series laptop. Et cetera.

When you start muddling things by switching OSes and which browsers are involved, sure different outcomes can happen. But that isn't how such a comparison should be done. It's done by holding all other aspects constant (OS, apps, etc) while changing only the underlying hardware. And what I said then is true - the faster system will always "feel" faster too.

 

[pdoherty]

We have Geekbench 5 single-core and multicore. Single-core generally reflects responsiveness while multicore better represents work that you can get done. So you could have a score of 1,500 for single-core and 3,000 for multicore and another CPU have a single-core of 1,000 and a multicore of 7,000. The first would feel more responsive in day-to-day tasks but the latter would be able to get large and long workloads done more quickly.

In actual reality, responsiveness doesn't mean you get more work done. It just feels more immediate and some translate that to better. Getting a response a few milliseconds or even a hundred milliseconds doesn't translate to more work done.

Sometimes even doing tricks like smoothing out response makes things feel faster or smoother even if they actually take longer.

And I'm thinking that making things smoother is far more a function of the OS and apps than it is the hardware. So faster hardware can more-easily be made to produce that "smoother" impression.

 

[leman]

When you start muddling things by switching OSes and which browsers are involved, sure different outcomes can happen. But that isn't how such a comparison should be done. It's done by holding all other aspects constant (OS, apps, etc) while changing only the underlying hardware. And what I said then is true - the faster system will always "feel" faster too.

Sure, but it’s not like changing the OS or the browser is going to affect the result this much. Different tests measure different things, all of which can be described as “performance“. You say “faster” system will feel “faster”. Well, what’s faster? A 64-core threadripper is as fast as computers go. And yet it will be outmatched by an iPhone in any kind of interactive app, simply because Threadripper is optimized for throughput and not latency.

 

[jav6454]

[...]

A lot of his complaints are about specific software or how to do things. There are lots of annoying things on Windows too. One of the biggest was their Windows 11 rollout. I went to do the installation and it bricked my Windows desktop. It took me four hours to get it back up and running.

My criticism of Win11 is how now, it is a massive invasion of privacy and spyware.

 

[pshufd]

And I'm thinking that making things smoother is far more a function of the OS and apps than it is the hardware. So faster hardware can more-easily be made to produce that "smoother" impression.

Browsers did this a long time ago putting in controls and wait times to improve the user experience which doesn’t necessarily mean that pages render faster. I was a bit surprised when I saw parameters related to waiting for things.

 

[pshufd]

My criticism of Win11 is how now, it is a massive invasion of privacy and spyware.

That is their revenue model and they really don’t have a choice. Same with Facebook and Google.

 

[jav6454]

That is their revenue model and they really don’t have a choice. Same with Facebook and Google.

Yeah, still doesn't make it ok. I'd rather spend $300 on an OS than have it riddled with spyware.