Slashdot Mirror

Actually, Intel's 10nm process results in features of almost exactly the same size as TSMC's 7nm process.

So yes, they're behind, but not by nearly as far as you think.

Another victim of Intel FP rounding errors!

Actually, Intel's 10nm process results in features of almost exactly the same size as TSMC's 7nm process.

So yes, they're behind, but not by nearly as far as you think.

Intel has effectively missed it's 10nm die shrink when Samsung and TSMC are on 7nm. Intel better have 5nm in it's back pocket because it's pointless building any 10nm CPU's now (maybe other chips instead.)

These values (14 nm, 10 nm, 7 nm) are not directly comparable. These days, the values seems more like marketing numbers.

Intel does need to focus more on innovation and implementation and less on customer segmentation, though. Intel used to be so far ahead that even an inferior processor design was on par with the competition, due to the advantage in fabrication.

Would be interesting, except 10nm, and 7nm are marketing terms with no basis in reality, and are actually the same. https://www.eejournal.com/arti...

I recently read this article which gives an excellent historical perspective on ISAs, RISC, CISC, VLIW, etc. To me, it also shows why very long upgrade cycles (like 2013->2019 in this case) might make quite a bit of sense now a days. We may be heading to a period of expensive, long-lived machines. Interesting times.

Calling them "Itanics" has been a long running joke.

Well, it may be multi-billion-dollar industry, but it spectacularly failed to meet its sales projections. My absolute favorite Itanium sales chart can be found here. Granted some of those initial projections were crazy stupid. But it fell short of even the much more modest, revised projections from 2002 and 2003.

Getting a robot to write it's own program has been done for a long time (10 years at least). This isn't even a new way of doing it, they are just using this guys code http://www.eejournal.com/archives/fresh-bytes/baxter-is-the-humanoid-robot-you-can-teach-without-programming/ .

Can Baxter (feminine of Baker) make twinkies?

This isn't even a new way of doing it, they are just using this guys code http://www.eejournal.com/archives/fresh-bytes/baxter-is-the-humanoid-robot-you-can-teach-without-programming/ .

How do you figure?

The project in this article (and all the software that the PR2 is run on) is open source ros.org. Baxter is closed source but reportedly uses ROS under the hood, it is in fact baxter leveraging the code from this article.

Getting a robot to write it's own program has been done for a long time (10 years at least). This isn't even a new way of doing it, they are just using this guys code http://www.eejournal.com/archives/fresh-bytes/baxter-is-the-humanoid-robot-you-can-teach-without-programming/ .

Link to a better writeup, one that doesn't attempt strained architectural analogies (ignore the first paragraph or three, but do look at the comments).