Their financial model is already in jeopardy considering it’s been underwritten by charging a premium for their tech because they’ve always been at least a cycle ahead of everyone else. Now that’s no longer the case, are they going to be able to continue charging that premium? I don’t think this is going to spell the end, but I think their absolute dominance is over and they’re going to be just one player in the industry going forward... from a consumer PoV that can only be a good thing
Most likely they are going to 10nm at a less ambitious density bump - they had aimed for 2.7x over 14nm, they’ve probably slashed that to 2x or less just to float themselves off the 14nm rock they’re stuck on...
I think this will be interesting to watch. As Mac consumers, most of us are concerned about the transition from Intel to ARM if it occurs because of compatibility. However, I'm wondering at this point if Microsoft will just throw their support towards it too.
This doesn't really seem to be true anymore when iPhones are beating Macbook Pros in benchmarks.
Based on what? TSMC 5nm will be here in 2020 and 10nm Intel isn’t ready until next year.
Based on specs. nm designation has been marketing for quite some time.
I left Intel some time ago, and even if I knew I wouldn’t have the right to say it, since NDA in place. But the way Intel works, is to have several developing lines, with isolated teams in different countries working. So carbon nanotubes are a good option for the middle time but the big future is Quantum cpu’s doing the parallel work together with classic CPUs.
Here's what a lot of Mac users don't want to see, one OS for both desktop and touch. This has been Microsoft's approach with Windows 10, and has resulted in an experience that works on, but doesn't feel optimized for both platforms.
It’s a complex situation, but a big part of it is transistor density, for ages Intel have been a node ahead of anyone else, so at best TSMCs 10nm would have equal density to Intel’s 14nm. Now that the competition is getting more than a cycle ahead they’re beginning to pull away on this metric, which is a big factor underlying Intel’s superlative single core performance. Brian Krzanich said this was a big part of the issue with their 10nm plans - they aimed for a 2.7x density increase over 14nm and couldn’t make it work. I expect they’ve now binned that and are working on just porting the 14nm chips over to a 10nm process just to get moving again
1. There was a day that Intel almost got their SOC in the first IPAD. It didn't cost them big money losing that deal. However the fun started after all the money went into A-chip and later TSMC became Apple sole chip manufacturer resulted in quick maturing in 7nm.
Benchmarks are a good way of comparing the amount of work two chips can get done in a certain time full stop though - that’s how they’re designed to work. What you say about software optimisation is undoubtedly true, but then Intel will never beat Apple on that, software-hardware integration is Apple’s tour de force.
