Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/intel-ceo-pat-gelsinger-on-achieving-chip-independence-goal-is-50-50-by-the-end-of-the-decade.19457/page-2
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Intel CEO Pat Gelsinger on achieving chip independence: Goal is 50/50 by the end of the decade

I believe any chance of Taiwan agreeing to unify with China sailed out after the 2019-2020 Hong Kong protests. The methods and ways which the CCP took control and ended much of the freedoms and rights of the people in Hong Kong horrified Taiwan (and much of the Western World).

Xi Jingping did more for the Taiwan independence movement than any Taiwanese politician could ever hope to do. There isn't a chance that Taiwan will ever voluntary agree (without coercion) to unification under Xi's China.
Taiwan seeking or pursuing independence is almost irrelevant all that matters is US current embargo as policy and Xi’s response.

1) West relaxes or clamps down

2). Xi’s pursue a multiple decade independent approach, or a hardline and embargo or invasion to bring the West to their knees
 
I do think the additional investment in Israel is worthwhile since Intel Israel has been one of the major R&D hubs, and they had contributed significantly to Intel's past. Like Raptor Lake which delivered a significant performance uplift when the architecture and process node stays the same. Also US has plenty of reason to step in middle east. But it is extremely difficult to step in the conflict between Taiwan and China. Let's not forgot that North Korea just declared that their peace treaties have officially breached, and they are ready to attach SK at any time. The east is now a potential battleground.

https://www.pbs.org/newshour/world/...ue-reconciliation-with-south-kim-jong-un-says

If we agree Pat Gelsinger's opinion that Taiwan is an unsafe location for semiconductor supply chain, can we also
agree that his $25 billion new fabs in Israel, an active war zone, does not contradict to his safety concerns?

Or he failed his CEO's fiduciary duty to safeguard shareholders, employees, and company's interests?

What's wrong with Arizona, Ohio, Oregon, Colorado, New Mexico, New York, Ireland, Germany, Japan, or Singapore ? Are those places inferior and unsafe?
 
And let's not forget that Taiwan is part of China (which is the official US policy).
The official US policy is " the United States recognizes the PRC as the sole legal government of China but only acknowledges the Chinese position that Taiwan is part of China ". The USA has explicitly refused to change "acknowledge" to "recognize" - in other words, we hear what they say, but have not said we agree it is effective. Isn't diplomacy fun?
 
When intel planned to invest in Israel Fabs and acquire Tower Semi, one idea was popped out: Why not intel to acquire UMC or even build fab in Taiwan with cheaper utility, high skilled workforces, lower people cost, close to OEM customers and more. If it is true, what will happen then? It will be very interesting.
 
Taiwan seeking or pursuing independence is almost irrelevant all that matters is US current embargo as policy and Xi’s response.

1) West relaxes or clamps down

2). Xi’s pursue a multiple decade independent approach, or a hardline and embargo or invasion to bring the West to their knees
Comrade Xi surely have to have this all locked in before his time ends no?

Its seems like a personal mission , so any timelime is dependent on his health is it not?
 
When intel planned to invest in Israel Fabs and acquire Tower Semi, one idea was popped out: Why not intel to acquire UMC or even build fab in Taiwan with cheaper utility, high skilled workforces, lower people cost, close to OEM customers and more. If it is true, what will happen then? It will be very interesting.

Somebody needs to buy UMC. Maybe Tower?
 
TSMC mentioned AZ during the call. There seemed to be a veiled threat that if they don't get CHIPs Act money there may not be expansion in AZ. Intel did the same thing for Ohio so we shall see. Will customers pay extra for US fabbed wafers? Probably not.

Mark Liu
Yes. The second fab shell is under construction. But what technology in that shell is still in under discussion. I think that also has to do with how much incentives that fab, the U. S. Government can provide. And yes, there will be a gap. At least, current planning is '27 or '28, that will be time frame. To be honest, most of the fab in overseas, what actually be loaded, what technology is being set up, really, it's a decision of customers' demand in that area at that timing. So, nothing is definitive, but we are trying to optimize value for the overseas fab for TSMC.
 
TSMC mentioned AZ during the call. There seemed to be a veiled threat that if they don't get CHIPs Act money there may not be expansion in AZ. Intel did the same thing for Ohio so we shall see. Will customers pay extra for US fabbed wafers? Probably not.

Mark Liu
Yes. The second fab shell is under construction. But what technology in that shell is still in under discussion. I think that also has to do with how much incentives that fab, the U. S. Government can provide. And yes, there will be a gap. At least, current planning is '27 or '28, that will be time frame. To be honest, most of the fab in overseas, what actually be loaded, what technology is being set up, really, it's a decision of customers' demand in that area at that timing. So, nothing is definitive, but we are trying to optimize value for the overseas fab for TSMC.
The threat doesn't sound veiled to me. Nonetheless, I've always thought Apple has more sway with TSMC than the CHIPS Act. Also, the Commerce Department seems rather arrogant, and probably won't respond well to threats.
 
USB was invented by Intel, the lead architect was Ajay Bhatt, who went on to lead the original PCIe definition in Intel, and USB was a project in Gelsinger's organization. I have no idea what if any technical contributions Gelsinger might have made, but he was a senior manager at the time. (Intel even made some silly commercials about Bhatt ages ago, with an actor portraying him.)


I vaguely remember this commercial - that is cool!

Not sure if you know, but USB seems awfully similar to Atari Serial Input/Output (SIO) developed in 1978:

- Serial interface for up to 256 connected devices
- Daisy chainable and/or connectable via a dumb hub
- +5V and +12V power supplied over the chain
- Each device is required to have it's own firmware, and can upload code to the host - i.e. all devices must be smart and self-identify what they are at least

https://en.wikipedia.org/wiki/Atari_SIO

Certainly the USB cable is a lot smaller, speeds are a lot faster (Atari SIO defaulted to 19.2kbps, but could run up to 80-126 kbps), and less pins are required to achieve the same effect, but there's a lot in common. The guy who designed this also worked on some USB stuff later (Joe Decuir).

(Atari SIO also carried an analog audio signal which was great for allowing audio to play from a cassette.. sometimes even when simultaneously loading data from a cassette).
 
Interesting that Atari SIO gets a few pages of history and explanation, while widespread hub and daisy-chain sytems like IBM 2701 and the related BiSynch with their 10-year head start barely get a list of devices from Wikipedia. I guess mainframes don't have homebrew fan clubs. Even in the late 1980s I remember we could get comms chips from Intel and Zilog which included BiSynch support, but Ethernet controllers were still relatively rare.
 
Interesting that Atari SIO gets a few pages of history and explanation, while widespread hub and daisy-chain sytems like IBM 2701 and the related BiSynch with their 10-year head start barely get a list of devices from Wikipedia. I guess mainframes don't have homebrew fan clubs. Even in the late 1980s I remember we could get comms chips from Intel and Zilog which included BiSynch support, but Ethernet controllers were still relatively rare.
You're right the retro enthusiasm for mainframes isn't as high as retro micros -- due to volume of markets, but there are some Vintage groups out there that keep them alive and even use them occasionally I was able to donate an old IBM System/32 that I got ahold of to the "Vintage Computer Festival East" group and they've actually restored a few that they demo for historical / learning purposes. It's cool still being able to walk up and look at (or touch occasionally) a UNIVAC or similar.

Edit: Found a bit on BiSync: https://en.wikipedia.org/wiki/Binary_Synchronous_Communications
It looks like it's point to point protocol only rather than a full bus and wiring / connector standard like SIO or USB? No power provided over the connection, or firmware-on-device features either.

I can definitely see some elements of SIO/USB here, but I think SIO had every major element of USB before hand, even including third party devices that were allowed to use it.
 
I was able to donate an old IBM System/32 that I got ahold of to the "Vintage Computer Festival East" group and they've actually restored a few that they demo for historical / learning purposes.
Nice! I briefly used an S/32. It had some kind of COBOL-lite (RPG!) language for data manipulation which I studied, and the 96-column cards seemed cool at the time. Almost 3x denser memory than the 80-column, it was easier to carry my code to the submission window!
It looks like it's point to point protocol only rather than a full bus and wiring / connector standard like SIO or USB? No power provided over the connection, or firmware-on-device features either.
Yeah, when your peripheral consumes a kW you don't power it over the wire. For sure, SIO brought some new stuff. The 2701 equipment had daisy chaining as well as hubs, but I suspect everything plugged into a socket for electricity since even a light ticket printer was likely 20W or more. Though we did have some wire-powered equipment it was mostly associated with telephony which had been self-powered over the signal wires since the telegraph. Even today I believe the landlines in most parts of the world still deliver 24V and something like 50mA. It was one of AT&T's arguments controlling 3rd party phones and modems up until the early 80s "we need to certify they won't cause other lines to fail by taking too much power".
 
Yeah, when your peripheral consumes a kW you don't power it over the wire.
I got a really good laugh out of this :) - Thanks for reminding me about how much power even “simple” early data center peripherals used to consume.

Thanks for sharing those other anecdotes too. That’s an interesting spec on the analog lines. I remember holding a phone wire in my mouth one time as a kid, and I think someone called (which raised the voltage) and getting a brief white out in vision. Never did that again..
 
Nobody thought Ukraine would flare up like it did.

Ukraine was preparing for this war for all 30 years of its independence. It's the biggest military in Europe after all.

Russia was visibly preparing for the invasion of Europe as early as 2003, and you can find articles as early as 2004 with detailed Russian plans. Generals were fuming over White House's inaction.

It's a bigger shocker in 2024 that there were so many people who didn't know that.

Similarly, I am absolutely sure communists will attack, and that's the only a question of time. Otherwise, why they spent trillions on preparations for expeditionary warfare?
 
Last edited:
Ukraine was preparing for this war for all 30 years of its independence. It's the biggest military in Europe after all.

Russia was very visibly preparing for the invasion of Europe at least since 2003, and generals were calling out White House's inaction as early as then.
This isn't wrong. But I had assumed the build up on the border was just going to be the prelude to a limited strike elsewhere (like what happened when Russia nabbed Crimea without resistance). Doubly so because of how economically disastrous total war is even before the sanctions that everyone knew would follow a full scale war.
 
This isn't wrong. But I had assumed the build up on the border was just going to be the prelude to a limited strike elsewhere (like what happened when Russia nabbed Crimea without resistance). Doubly so because of how economically disastrous total war is even before the sanctions that everyone knew would follow a full scale war.
If you can grab it all, why would you grab just a part of it? Especially, if you are a dictator with "conquer Europe" obsession.

If Ukraine, the largest military in Europe, had collapsed in 2022, who else would've stood in the path of freaking 40 divisions?

No one. There is simply nobody that big, with large enough ammunition stockpiles in Europe.
 
Back
Top