The AMD Execution Thread [2007 - 2017]

Status
Not open for further replies.
A single tape-out delivered into multiple leading edge foundries, even if they're implementing the same technology, is pretty impressive.
 
The last time something like this was done, it was called this:
http://www.commonplatform.com/manufacturing/

My memories of some interview quotes after that fizzled was that there was little demand at the time.

What might make this time different is that this is not a common platform initiative as much as Globalfoundries is paying for a Samsung process and ditched its own 14nm tech, and that there is finally a customer in Apple that might actually care to have that kind of option.
 
Even with a common platform, customers can still create custom lib's for their architectures right?
Yep, but then you (high likely) can't do what AMD are trying to do (single GDS2 delivery to multi-foundry manufacture).
 
Yes, but in practice it seems that GloFo will always lag behind Samsung a bit, or perhaps a lot, and as far as I know it doesn't change the WSA.

Edit: yes, WSA = wafer supply agreement.


I dont see how Glofo could have bad yields (( when they have just start the production anyway of 14nm ) when it is not the case for Samsung.. or if you prefer, i dont see looking at the arrangement between Samsung and Glofo, if Glofo have a problem, why Samsung will not goes help them to solve it quickly Asap as they have the knowledge for solve it really quickly and with a good reason, Samsung use the Glofo line for his own productions too. This permit them to have a lot flexibility...

Im sorry, untill we got more information who push to the same conclusion.. for today, the article of digitimes is a bit short.

They will allways lag behind Samsung, because Samsung lead the research and developpement on the semi conductor industry.. ( ok, some could say Intel, but thats a bit different, anyway this make years than Samsung and Intel are trusting the first and the second place in this domain, some years it is Samsung on the first place, some years it is Intel ) ..

Today, theres not really any funders who have the foce that can have Samsung in this domain ( ootside Intel, but again this is a different case, so far Intel dont " rent " or sell his tech, this will change soon ), For be honest, i dont even see how TSMC could fight Samsung this next years... Not an a technology point of view, i mean i think the tech of TSMC can match Samsung one, but not on the other aspect..
I dont say they are the best at it or have the best technology... i just say, this will be hard to rivalize with them, financially or economically, and on the workforce. Do you imagine the volume of production that have Samsung today..
 
Last edited:
I dont see how Glofo could have bad yields (( when they have just start the production anyway of 14nm ) when it is not the case for Samsung..
It's not uncommon for different fabs of the same company having different yields.

That's why silicon sometimes gets certified for a specific fab, not just for a specific process.
 
It's not uncommon for different fabs of the same company having different yields.

That's why silicon sometimes gets certified for a specific fab, not just for a specific process.


Yes, but there's a difference between having "not so good yields" than the other, and have so bad yields, that they need to push back the production.
 
Yes, but there's a difference between having "not so good yields" than the other, and have so bad yields, that they need to push back the production.
When you're dealing with 14nm structures, a few innocent looking differences in the production pipeline can go a long way in explaining dramatically different yields.

(Only tangentially related, but it reminds me of a case study in college where a glass production plant looked for weeks to explain a difference in production quality result between daytime and nighttime. It was ultimately root caused by a small window that was opened in some room during the day.)
 
Sweclockers.com published a news blurb today with speculation that Microsoft is in talks to buy parts of, or the whole company.

Interesting.
 
Sweclockers.com published a news blurb today with speculation that Microsoft is in talks to buy parts of, or the whole company.

Interesting.

They are just reporting what Fudzilla reported a few days ago. And we all know how reliable Fudzilla is.

Microsoft buying AMD makes no sense whatsoever to me. Competing with Nvidia and Intel while at the same time controlling the PC gaming platform and API? How exactly would that work?

But then again Fudzilla is also saying Intel is also buying AMD so yeah...
 
If the early tape-outs of Zen are showing promise, Microsoft buying AMD right now would actually make a lot more sense than, say, Microsoft buying Minecraft for $2.5B.

It would put Sony and Nintendo into a lot of pressure for their next consoles, plus they'd get their own ARM and x86 and GPU teams to fight Apple with custom designs.
 
If the early tape-outs of Zen are showing promise, Microsoft buying AMD right now would actually make a lot more sense than, say, Microsoft buying Minecraft for $2.5B.
Am I missing a /s ?

Minecraft was a money making machine before they were bought by MS. And judging by the way kids are still addicted to it, there's no reason it won't stay that way for a long time to come. If anything, that $2.5B was cheap. They bought the attention of a whole generation.
 
If the early tape-outs of Zen are showing promise, Microsoft buying AMD right now would actually make a lot more sense than, say, Microsoft buying Minecraft for $2.5B.

It would put Sony and Nintendo into a lot of pressure for their next consoles, plus they'd get their own ARM and x86 and GPU teams to fight Apple with custom designs.

Minecraft actually generates profits.
Does AMD generate profits?
 
Minecraft actually generates profits.
Does AMD generate profits?

Isn't this a rather narrow view? Do we live in a world where only the healthy companies are bought?

As stated above, Microsoft seems to be fine with loss-leading divisions as long as it brings mind-share and market presence.
AMD's acquisition would be made for IP and technological dominance, while taking advantage of a period of poor performance to buy it cheap.
Just like Nokia before them.

Obviously, it wouldn't be a straight-forward "this is making money let's own them" purchase. It would be a strategic one.


Am I missing a /s ?
Minecraft was a money making machine before they were bought by MS. And judging by the way kids are still addicted to it, there's no reason it won't stay that way for a long time to come. If anything, that $2.5B was cheap. They bought the attention of a whole generation.

No, what you're missing is a little research over the subject.
Mojang was making around $130 million a year before the acquisition in late 2014.
For $2.5B this means that it'd take over 20 years to get the money back, assuming Minecraft would sell just as much during that time, which is ridiculous from a business POV.

What buying Mojang meant is that Microsoft will prevent further Minecraft iterations from ever appearing in anything other than Windows/XBox platforms. They bought mind-share and market presence.
 
This would have made sense in the Ballmer era, at this time when you watch an Ms representative showing off Office at an Apple even doesn´t make much sense

What about Nvidia? :mrgreen:

About Minecraft, IMO the very same time that MS tries to tie it to one device or one OS, I willl lose relevance I think
 
No, what you're missing is a little research over the subject.
Mojang was making around $130 million a year before the acquisition in late 2014.
For $2.5B this means that it'd take over 20 years to get the money back, assuming Minecraft would sell just as much during that time, which is ridiculous from a business POV.
Actually they stated that the purchase would be positive for them in the next fiscal year after purchase. Why? Because they bought the company with overseas funds that is very unlikely to have ever made it back to the US because of tax charges so all it was doing is sitting there earning meagre amounts of interest - the profits they are projected to generate with the purchase of Minecraft are greater than what that money would have otherwise generated.

Of note, the Nokia deal was the same, with 5.44 Billion Euro's being drawn from overseas funds to finance the transaction.

What buying Mojang meant is that Microsoft will prevent further Minecraft iterations from ever appearing in anything other than Windows/XBox platforms. They bought mind-share and market presence.
Except they have so far stated the opposite and have continued to develop and support other platforms. MS of late have been very much expanding their ecosystem to other mobile devices
 
Status
Not open for further replies.
Back
Top