SemiAccurate Forums  

 
Go Back   SemiAccurate Forums > Search Forums

Showing results 1 to 25 of 500
Search took 0.02 seconds.
Search: Posts Made By: aaronspink
Forum: Process technology 12-14-2014, 03:45 AM
Replies: 41
Views: 3,597
Posted By aaronspink
You've managed to equate Vin with actual Core...

You've managed to equate Vin with actual Core Vdd. The only think that voltage listing tells you about core voltage is basically absolutely completely nothing. In fact, using that as your point of...
Forum: Process technology 12-13-2014, 05:01 PM
Replies: 41
Views: 3,597
Posted By aaronspink
Then you are going to have to provide the proof...

Then you are going to have to provide the proof on this one.



The RTL may be substantially the same, that doesn't however say anything about the actual implementation in gates, wires, etc. And...
Forum: Process technology 12-12-2014, 03:09 PM
Replies: 41
Views: 3,597
Posted By aaronspink
A) we don't know what voltage is required for...

A) we don't know what voltage is required for 2.6Ghz on a given part and it will not be the same for a given part. Voltages are set on a part by part basis during test. There is no external voltage...
Forum: Process technology 12-12-2014, 01:58 AM
Replies: 41
Views: 3,597
Posted By aaronspink
No it is not at all clear. The F/V curve is of...

No it is not at all clear. The F/V curve is of absolute relevance! So is Idsat vs Voltage.



And I think you need at least basic semiconductor theory which you are completely lacking.
Forum: Process technology 12-11-2014, 04:57 AM
Replies: 41
Views: 3,597
Posted By aaronspink
The thermal limits are entirely relevant and no...

The thermal limits are entirely relevant and no one knows what voltage is needed to operate at 2.6Ghz outside of Intel. And those voltages are likely to vary from chip to chip AND core to core! Its...
Forum: Process technology 12-11-2014, 04:11 AM
Replies: 41
Views: 3,597
Posted By aaronspink
Apples meet oranges, oranges meet apples. Your...

Apples meet oranges, oranges meet apples. Your comparison is basically irrelevant as the Core M is thermally limited. And FYI, the Core M operates in a larger voltage range than .988-1V.




...
Forum: Process technology 12-10-2014, 09:33 PM
Replies: 41
Views: 3,597
Posted By aaronspink
And which pray tell electrical characteristics...

And which pray tell electrical characteristics are leading you to believe its a failure?



Actually, if you go through the options library at say TSMC or GF, you'll learn that they have 10s of...
Forum: Article discussion 12-10-2014, 01:14 AM
Replies: 20
Views: 1,615
Posted By aaronspink
From everything I've seen and read, they are...

From everything I've seen and read, they are fully symmetrical. Now in software(ie firmware), they may treat them as asymmetrical, but that doesn't mean they are aSMP.





If you actually...
Forum: Process technology 12-06-2014, 10:59 AM
Replies: 15
Views: 1,218
Posted By aaronspink
All depends on what you are doing with that MCM...

All depends on what you are doing with that MCM and what the components are. For most mobile SoCs? No. It doesn't make much sense. Most are in the range of 100-150mm2. If the fab can't get...
Forum: Process technology 12-06-2014, 10:54 AM
Replies: 106
Views: 13,361
Posted By aaronspink
I know right? They are only shipping millions of...

I know right? They are only shipping millions of 14nm chips right now!
Forum: Process technology 12-05-2014, 08:58 PM
Replies: 106
Views: 13,361
Posted By aaronspink
that's not what it actually says, but good try. ...

that's not what it actually says, but good try. What it says is that Intel is currently asking 3rd parties 25% more than Samsung. It doesn't say anything about actual cost.
Forum: Process technology 12-05-2014, 08:56 PM
Replies: 106
Views: 13,361
Posted By aaronspink
Numbers are readily available if you would take...

Numbers are readily available if you would take the time to look.
Forum: Process technology 12-04-2014, 03:59 PM
Replies: 106
Views: 13,361
Posted By aaronspink
Still the volume of Intel at say 22nm is...

Still the volume of Intel at say 22nm is significantly greater than the volume of TSMC at 28/20nm.




You handle demand via throughput not by binning unless you are incredibly constrained...
Forum: Process technology 12-04-2014, 03:50 PM
Replies: 106
Views: 13,361
Posted By aaronspink
I could of sworn that Intel was using multiple...

I could of sworn that Intel was using multiple patterning at 22nm as well.



It somewhat depends on what your limiters are. If you are wire limited on the upper layers, it basically has no...
Forum: Process technology 12-04-2014, 01:12 PM
Replies: 106
Views: 13,361
Posted By aaronspink
Intel has always named their nodes based on...

Intel has always named their nodes based on transistor performance. If you haven't figured that out by now...



LOL, Intel is in volume production of *RETAIL* 14nm silicon. The closest thing...
Forum: Process technology 12-04-2014, 05:18 AM
Replies: 106
Views: 13,361
Posted By aaronspink
I haven't worked for Intel in years, and would be...

I haven't worked for Intel in years, and would be making the same comments even if I had never worked there. Intel's per wafer costs have generally been the lowest in the industry for a leading edge...
Forum: Process technology 12-04-2014, 05:13 AM
Replies: 106
Views: 13,361
Posted By aaronspink
Total fab cost per year is going to be in the...

Total fab cost per year is going to be in the 10-14bn range at most. And that's with significantly more capacity at the leading edge than TSMC has. Generally about 2-3x.



60-70% of TSMC...
Forum: Process technology 12-04-2014, 02:51 AM
Replies: 106
Views: 13,361
Posted By aaronspink
LOL, your numbers are so off as to once again be...

LOL, your numbers are so off as to once again be nothing but a joke. Even if you accounted ALL capital expenses and ALL R&D at Intel to the manufacturing, it is significantly less than 25bn. The...
Forum: Process technology 12-01-2014, 08:19 PM
Replies: 106
Views: 13,361
Posted By aaronspink
It increases per wafer costs, does not...

It increases per wafer costs, does not necessarily increase per transistor costs. In fact some IDM are saying they still have lower per transistor costs. I would be weary of those that don't.
Forum: Process technology 12-01-2014, 08:18 PM
Replies: 106
Views: 13,361
Posted By aaronspink
You mean from the company that's been doing SADP...

You mean from the company that's been doing SADP from day 1 of DP? And switched to unidirectional layers multiple processes ago? The question you should be asking is how much TP the foundries will...
Forum: Process technology 11-29-2014, 02:08 AM
Replies: 106
Views: 13,361
Posted By aaronspink
Exactly. Anyone thinking Intel is PRQing a part...

Exactly. Anyone thinking Intel is PRQing a part at less than 80+% yield is fooling themselves. Intel takes yield and production costs very very seriously. It is, has been, and always will be the...
Forum: Process technology 11-29-2014, 01:55 AM
Replies: 106
Views: 13,361
Posted By aaronspink
You are off on the high side and getting more off...

You are off on the high side and getting more off as you continue.



And an equal number of chipsets, and an equal number of aux chips.



Intel's yearly R&D numbers are clearly spelled...
Forum: Process technology 11-28-2014, 03:13 PM
Replies: 106
Views: 13,361
Posted By aaronspink
Porting any design to ANY OTHER process is far...

Porting any design to ANY OTHER process is far from easy and is going to cost quite a bit of money. The restrictive design rules add less than marginal noise to that cost. In fact, many of Intel's...
Forum: Process technology 11-28-2014, 09:18 AM
Replies: 106
Views: 13,361
Posted By aaronspink
8k per wafer? You are rather significantly off. ...

8k per wafer? You are rather significantly off. And your yield numbers might be in the ballpark for some of the foundries, but people would be fired left and right if 100mm2 chips were yielding at...
Forum: Process technology 11-28-2014, 09:13 AM
Replies: 106
Views: 13,361
Posted By aaronspink
Extensive experience in the semi industry and the...

Extensive experience in the semi industry and the costs a multiple different semi companies. Its not exactly like its some big secret. Intel's manufacturing costs have been the lowest in the...
Showing results 1 to 25 of 500

 
Forum Jump

All times are GMT -5. The time now is 12:36 AM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2014, vBulletin Solutions, Inc.
SemiAccurate is a division of Stone Arch Networking Services, Inc. Copyright 2009 Stone Arch Networking Services, Inc, all rights reserved.