• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

xBox One's SoC is TSMC

Interesting...

Does anyone know whether TSMC's or GloFo's 28nm process(es) is/are technically superior to the other(s), from an electrical charictaristic standpoint? Is that data available?

I'd imagine yields are substantially in the favor of TSMC for multiple reasons, however I am curious to see if there is any reason to go with GloFo.
 
Interesting. The subpar foundry didn't get one of the console contracts.

That was obvious for quite a while. It also means the "subpar" foundry isn't to blame for the xBox's yield problems, which is something you and a few others would no doubt have revelled in.

Note that it also means AMD can pay the WSA at current levels of share, without the consoles.
 
Last edited:
Interesting...

Does anyone know whether TSMC's or GloFo's 28nm process(es) is/are technically superior to the other(s), from an electrical charictaristic standpoint? Is that data available?

I'd imagine yields are substantially in the favor of TSMC for multiple reasons, however I am curious to see if there is any reason to go with GloFo.

From what I can gather, GF claims performance superiority and TSMC doesn't deny it. TSMC for sure is yielding better currently, and is much less risky overall for AMD so was basically the only choice.
 
TSMC is really the premier foundry - I wouldn't expect something this important to go anywhere else.

Depends on competition and capacity, and how desperate GloFo are for customers. (And how desperate AMD are to clear their WSA!) I wouldn't entirely rule out the PS4 going to GloFo.
 
AMD designed both Jaguar and GCN for TSMC already. Jaguar is supposed to be pretty portable - no idea about GCN - but there'd at least be some non-zero work effort in porting it to GF.

I'd also bet big that TSMC's 28nm is denser than GF's.
 
AMD designed both Jaguar and GCN for TSMC already. Jaguar is supposed to be pretty portable - no idea about GCN - but there'd at least be some non-zero work effort in porting it to GF.

Kaveri will be GCN at GF so porting it won't be an issue.

I'd also bet big that TSMC's 28nm is denser than GF's.
Hard to say, gate first supposedly gives a ~20% area bonus over gate last.
 
ewww eMMC

I wonder what they're using the 8GB for? OS storage would make sense, and the poor speed of eMMC (compared to their "grown-up", SSD brethren) wouldn't be an issue. It could also be a buffer for the video recording, but I'm not too sure about that.
 
It's interesting to see how customised it is to their needs. Maybe AMD does actually have a sound business with this semi-custom stuff - I have to admit I was dubious but if they can make it work for both MS and Sony then they can make it work for anyone, and are clear industry leaders.
 
Interesting...

Does anyone know whether TSMC's or GloFo's 28nm process(es) is/are technically superior to the other(s), from an electrical charictaristic standpoint? Is that data available?

I'd imagine yields are substantially in the favor of TSMC for multiple reasons, however I am curious to see if there is any reason to go with GloFo.
That's hard to know. I think TSMC are ahead of the rest. In the earnings call Chang said they have no serious competition for high-K metal gate. They do for Oxynitride, but that won't be any use for the X-box.
 
It's interesting to see how customised it is to their needs. Maybe AMD does actually have a sound business with this semi-custom stuff - I have to admit I was dubious but if they can make it work for both MS and Sony then they can make it work for anyone, and are clear industry leaders.

You'll need more than just console wins every 7 years to make a business model off of semi custom stuff. You'll need a steady stream of customers to create a critical mass of product lines that you can refresh every generation with a couple tweaks here and there. I'd be interested in seeing in what other industries semi custom designs work well for.
 
You'll need more than just console wins every 7 years to make a business model off of semi custom stuff. You'll need a steady stream of customers to create a critical mass of product lines that you can refresh every generation with a couple tweaks here and there. I'd be interested in seeing in what other industries semi custom designs work well for.

Well that's what I mean. The prestige from developing to delivering semi-custom silicon for two of the world's largest companies is all the marketing it'll ever need.

If the business exists at all, AMD has shown themselves to be the only game in town. Who else is doing it on this scale?

I suppose one way to look at it is that everything from now onwards is just a bonus.
 
Last edited:
Back
Top