i5/i7 Difference - Hyperthreading?

[witeken]

I'm not arguing that Intel is going to launch 7nm at a specific time, because I clearly have no idea what is going to happen in the next 4 years. Neither do you. And you know what? Neither does Intel!

I am pointing out that your claims that Intel "knows" that it will outperform Moore's Law are ridiculous. They can certainly predict and hope that they will, but there are always unknown unknowns. Never take anything as a certainty, especially when dealing with risky cutting edge technology.

I was referring to this slide:

Transcript:

“So I think you talk about the cost chart where at the very end I put a few little dotted points. That was intentially ambiguous. What we want to make clear to you is that we believe that the cost reduction historical trend can be continued. So we're extending the dotted line on the historical trend. What we didn't say is exactly how far below that line we think we'll be and how that trend will be the previous generation exactly because that stuff we don't know. It will be below the historical trend, that much we know. [...] Yes [we can get below that line without EUV], but I don't want to: I want to be further below the line.” William Holt, Intel

This slide, by the way, is for a mature process. If Intel has yield issues, which like I said is possible, although might be much less of a concern/issue because they'll probably use EUV, then the slide will not be true. So if Intel can get successfully back into Tick-Tock cadence with Skylake and Cannonalake, I wouldn't worry too much about the next few generations.

 

[NTMBK]

I was referring to this slide:

Transcript:



This slide, by the way, is for a mature process. If Intel has yield issues, which like I said is possible, although might be much less of a concern/issue because they'll probably use EUV, then the slide will not be true. So if Intel can get successfully back into Tick-Tock cadence with Skylake and Cannonalake, I wouldn't worry too much about the next few generations.

And yet, again, that slide does not represent Moore's Law. It does not take into account the time component, and hence cannot demonstrate whether or not their progress is in line with Moore's Law. And notice how the graph labels 10nm and 7nm with little (*)s, indicating "Forecast". As in, we don't know yet, but here's our best estimate.

Please bear in mind that I am not predicting that Intel are going to fail to hit their targets. They have a great track record, apart from the recent 14nm slip, and could well get it back on track. But I don't know.

 

[witeken]

And yet, again, that slide does not represent Moore's Law. It does not take into account the time component, and hence cannot demonstrate whether or not their progress is in line with Moore's Law. And notice how the graph labels 10nm and 7nm with little (*)s, indicating "Forecast". As in, we don't know yet, but here's our best estimate.

Please bear in mind that I am not predicting that Intel are going to fail to hit their targets. They have a great track record, apart from the recent 14nm slip, and could well get it back on track. But I don't know.

Moore's Law is all about cost per transistor. Holt unfortunately didn't use years but node names, and BK said they'd be more prudent with that information, so of course we can't predict how well they'll do on outperforming Moore's Law (average cost per transistor reduction per year), but the mere fact that they are improving the cost faster than the trend while other companies are stagnating should be noteworthy. Intel also said they have no intention of slowing down, so I keep looking at 14nm as an exception, not a new rule.

Let's not forget the point we are discussing: you said that it was unknown whether they would actually outperform Moore's Law on an absolute (time scale) until they ramp 7nm, which I agree with.