They’ve never successfully built up a non-x86 userbase on Windows.
An architecture switch where a lot of software won’t run well (or at all) is a big risk for them that people could start to go elsewhere in non-trivial numbers.
Apple, Linux, Chromebook.
Yes they have a compatibility layer for running x86 on ARM. They could make it for other archs. But that’s still a big effort and consumers may not trust.
I think at this point, Azure and their OpenAI partnership are enough to weather the transition to ARM. It is just software, and Apple has (TWICE) shown how to do it right.
That said, has anyone checked in with Lisa Su on this?
I felt 2012 - 2015 Intel Macs were pretty decent: the first Retina MacBook Pro, the iMac 5K, for example. It definitely went down hill with the Touch Bar and the MacBook Pros with the Core i9 chips. Those seemed like they were constantly throttling, fans sounded like a jet engine...
IMHO that is because Intel wasn't delivering chips meeting the specs they promised. Once you dropped it into a system and put a thermal/cooling profile in, the new chip didn't really perform better than the old one.
On the lower end, Apple just stopped releasing updates because there was no useful advantage to new chips.
On the high end, Apple was fighting between their desire to have a machine pleasant to use, and one that would fire the fans full speed at boot to keep up maximum performance without thermal throttling.
Kinda? Apple was known to tune their ACPI tables pretty hard, it wasn't impossible for them to put a hard-limit of 70c like the other Wintel machines at the time. Instead they seemed to push the Turbo mode until you approached junction temp, which didn't seem like a smart idea for a mobile device. Especially those embarrassingly thin i9 workstations Apple tried shipping.
The behavior persists on Apple Silicon, it just gets there slower. Someone internally at Apple must have a vendetta against CPU throttling, I guess.
Heat transfer rates are higher when the difference in temperature is higher. So if the CPU temperature is below the maximum you're leaving some theoretical performance on the table. Someone at Apple has a hard on for performance data that's making them ignore real world consequences.
I don’t think that’s fair. The move to x86 was a huge step up from the PowerPC G5s. That said, the move to the M1 was even bigger, so I can relate to the sentiment.
Where's Microsoft? They have the most to lose if Intel goes under.
What is Microsoft losing? AMD is still selling x86 chips. There's ARM. They can port to RISC-V. OS isn't even their bread and butter anymore.
They’ve never successfully built up a non-x86 userbase on Windows.
An architecture switch where a lot of software won’t run well (or at all) is a big risk for them that people could start to go elsewhere in non-trivial numbers.
Apple, Linux, Chromebook.
Yes they have a compatibility layer for running x86 on ARM. They could make it for other archs. But that’s still a big effort and consumers may not trust.
I think at this point, Azure and their OpenAI partnership are enough to weather the transition to ARM. It is just software, and Apple has (TWICE) shown how to do it right.
That said, has anyone checked in with Lisa Su on this?
As someone who has used an Intel mac, then moved up to a M3, please god no. Intel macs were some of the worst things Apple has ever made.
I felt 2012 - 2015 Intel Macs were pretty decent: the first Retina MacBook Pro, the iMac 5K, for example. It definitely went down hill with the Touch Bar and the MacBook Pros with the Core i9 chips. Those seemed like they were constantly throttling, fans sounded like a jet engine...
IMHO that is because Intel wasn't delivering chips meeting the specs they promised. Once you dropped it into a system and put a thermal/cooling profile in, the new chip didn't really perform better than the old one.
On the lower end, Apple just stopped releasing updates because there was no useful advantage to new chips.
On the high end, Apple was fighting between their desire to have a machine pleasant to use, and one that would fire the fans full speed at boot to keep up maximum performance without thermal throttling.
Kinda? Apple was known to tune their ACPI tables pretty hard, it wasn't impossible for them to put a hard-limit of 70c like the other Wintel machines at the time. Instead they seemed to push the Turbo mode until you approached junction temp, which didn't seem like a smart idea for a mobile device. Especially those embarrassingly thin i9 workstations Apple tried shipping.
The behavior persists on Apple Silicon, it just gets there slower. Someone internally at Apple must have a vendetta against CPU throttling, I guess.
Heat transfer rates are higher when the difference in temperature is higher. So if the CPU temperature is below the maximum you're leaving some theoretical performance on the table. Someone at Apple has a hard on for performance data that's making them ignore real world consequences.
I don’t think that’s fair. The move to x86 was a huge step up from the PowerPC G5s. That said, the move to the M1 was even bigger, so I can relate to the sentiment.
Most likely for the foundry business
Is this Tim Cook’s way of getting preferential treatment from the Trump administration?
China is taking Taiwan unless Xi randomly dies of a heart attack.
It’s safer to have an alternative to tsmc knowing this.
China has been going to take Taiwan for a long time now, hasn't.
They have finished construction on the largest field hospitals and civilian bunkers across the strait.
It’s part of official Xi CCP policy to acquire Taiwan.
Most experts put the move around 2027 when the next tranche of ships are finished.
You could at least do readers the justice of citing which policy you're referencing offhand (Chinese Dream 2049: https://en.wikipedia.org/wiki/Chinese_Dream)
Intel has never sold their EULV nodes before. They're a "substitute" in the same way eating from the trash substitutes a McDonalds meal.