melgross
About
- Username
- melgross
- Joined
- Visits
- 122
- Last Active
- Roles
- member
- Points
- 10,933
- Badges
- 2
- Posts
- 33,687
Reactions
-
Apple's C1 modem is a quiet game-changer that's mostly flying under the radar
Intel got itself in trouble years ago at the end of the 10nm period. Before then other fabs had used nm loosely.10nm wasn’t really 10nm in that much of what was on the chips were 12 and 14nm. Intel had the idealistic idea that 10nm should actually be 10nm. They couldn’t do it. So a good two years after “10nm” came out elsewhere, they finally decided to follow the industry so they could catch up. By that time “7nm” had come out and they were forever behind. They proposed a new nomenclature which was number of transistors per square mm, or something similar, because they still had the better numbers, but unsurprisingly, no other company wanted to go along.SuntanIronMan said:
I know of a company that has shiny new leading-edge fab in Arizona, is struggling financially, has been selling/spinning-off various parts of its company and now has a new CEO.melgross said:
A modern fab costs up to $25 billion, and rising. It also takes years of experience to get the fab up to optimal performance. I doubt Apple will be willing to spend that money and take years to get it running properly. Then there are the risks of having a problem as fab companies gave. So now, 2nm is a year late from tank and others. But these companies have a number of favs. The build a new one and one by one, upgrade their older ones. How could Apple do that? They would have to have a fully functioning 3nm fab and build a 2nm fab.blastdoor said:Apple's silicon design team appears to be among the strongest teams in the company, and among the best silicon design teams in the world (if not the best).
I wonder if Apple will buy or build their own fab so that they are both designing and manufacturing their chips. I know what the knee jerk reaction to that suggestion will be, but TSMC margins are steadily going up, which means that's profit Apple is missing out on. Time and time again, we have seen Apple identify suppliers with fat profit margins and then take over that business themselves.
Maybe a way to start could be a joint venture with TSMC or Intel to build an Apple fab in the US. Apple could finance and own the fab and pay their partner some patent licensing and management/operation fees. Eventually Apple could then take over the management and operation.
no. Apple decided to not do that, and they’re right.
I’m not saying Apple would buy Intel fabrication business. Not only am I not saying that, but I’m darn-near positive that Apple would never, lol.
BUT (just to speculate wildly): If Apple did want to get into fabrication (which it doesn’t), there is a quick way to skip all the time consuming parts that you mentioned. It would just require a large sum of money (which Apple has), Intel’s willingness to sell (which it might) and some government approval (since Intel took CHIPS Act money that requires Intel to not sell their fabrication business for some time).
Again, I’m not at all saying this is something that would happen. And it wouldn’t never be as simple as that. Just having fun speculating about something that definitely won’t happen, lol.
at any rate, refusing to cooperate with Apple in making the A series of SoCs because they didn’t believe Apple’s numbers, which they later said were correct, was a reason TSMC has risen, and they have fallen.Nah, it’s not that simple. Remember Intel had a pretty decent modem that Apple bought. But they stopped making them as Apple stopped buying them because of bad publicity. Apple had its own team and then they bought Intel’s. You would think that they could have popped out a new modem in a couple of years with all of that, but it took six. I was a manufacturer a while ago and I can tell you that, no pun intended, there are an awful lot of moving parts to manufacturing. You can’t just take over a chip plant. Then it needs to be upgraded to 3nm. And you have to keep your eye on 2nm. And how do you go to 2nm? What is the path? Intel, TSMC and Samsung all have multiple plants. They do a shuffle. They run most of their older plants on current process, but upgrade one or two to the new process. Then they build a new one or maybe two. Then they close some old ones.If Apple buys one plant, how does that work? Intel has no current 3nm fabs they would sell. They would have to immediately upgrade it to what, 3nm which will not be used for bleeding edge chips in 18 months, or so? Or work on 2nm which would, since they would be going from scratch, take three years, if everything works out as expected? Remember that TSMC is a year behind on 2nm now. And Apple management has no hands on experience with this. You buy the plant and hire the people and within a few months, top management leaves as do engineering personnel. No, terrible idea. Apple is well aware of this too.
i don’t understand why people make these proposals. Years ago it would have been easier. It’s far more difficult now. If they were going to do this they should have started before the first A series was ready, back in 2007. They could have bought a number of state of the art fabs from IBM, AMD and others at the time really cheaply, with a much easier learning curve. It’s too late now.
-
Trump's team promises to keep TikTok from being banned
Trump was the one who wanted Tick Tock banned in the first place. Now, he wants it because he thinks the viewers will want to see HIM on it. What a hypocrite.
-
Apple-Nvidia collaboration triples speed of AI model production
Good one! Apple has had a feud with Nvidia for a long time. It started with CUDA. Apple came up with their own software for that which was better. ATI embraced it, but Nvidia didn’t allow it in their GPUs. That was the first problem. The second was as linked to above, when Nvidia changed their production to different solder balls. It didn’t tell the manufacturer of the sub boards the chips were then soldered to. When the incompatible solder connections began to fail for all manufacturers, and we had two iMacs that failed because of this, Nvidia refused to take responsibility. Eventually they had to put $500,000,000 into an escrow account for manufacturers to dip into, but it wasn’t adequate.netrox said:
What makes you think Apple wasn't friendly with Nvidia?elijahg said:So Apple is finally friendly with Nvidia again?
Apple hasn’t used Nvidia since. So this is interesting and somewhat surprising.
-
NeoRuler and M-Cube review: Smarter, iPhone-connected precision measuring
These seem nice. But neither is a replacement for any of the precision measuring equipment I use. While most calipers are accurate to about plus or minus 1.5 thousandths, with more expensive models managing plus or minus one thousandth, they aren’t accurate enough for precise work. I don’t rely on them when machining other than as an approximation. These would be useless for that as well. I’m trying to decide where they would be useful, but other than for the built-in scales, all of which I have in apps in my phone now, I can’t really think of anything I would use them for.
-
Apple takes over more NYC office space from Macy's
I wouldn’t walk cross 5th avenue most of the day during a red light. At least, not if you expect to get to the other side.anonymouse said:
Now that jaywalking is officially legal.melgross said:That’s not really far. I’ve done more than that through Manhattan many times over the years. It’s at most a 25 minute walk, depending on traffic lights. Often, it takes less time. I can make it in 15 minutes the way I walk.
