why? might helpus to understand what you mean.

simplest, take cost per 10,000 tray (or whatever) and divide by 20,000 for c2q or 10,000 for i7/i5.

or take cost of a fab, estimate the number of dies per day, add in the cost of raw materials and staff and R&D costs and hey presto.

or assume a 20-40% profit margin to allow for more R&D and shareholder value, and subtract that % from the shop price, subtract their margin and there you are...
 

chedderslam1976

Distinguished
Nov 24, 2008
463
0
18,790


This is the most commonly used layman's equation:

(cost of wafer) / (dies per wafer)

However, you can use the Moore-Jobs equation for even better results:

[(dies per wafer) ^2 / ((dies per wafer) % (cost per wafer) - (cost per wafer * waffle size)] * (waffle diameter - (die size ^ 2))

Hope that heps!
 

chedderslam1976

Distinguished
Nov 24, 2008
463
0
18,790
Yes, of course, I should have mentioned that. I always use the Moore-Jobs equation which takes into account the waffle bandwidth so it is easy to forget.

Thanks!
 


Doesn't really take into account the r&D costs, or the costs associated with the plant to make the dies, the cost per wafer should account for this but just pushes the problem of cost one further stage down the line.

Knowning the point of the question would relly help.
 

chedderslam1976

Distinguished
Nov 24, 2008
463
0
18,790


The waffle factor takes all this into account.
 

4745454b

Titan
Moderator
I don't think it takes into account die harvesting however. All companies seem to be doing a better job of disabling parts to put out usable chips. The formula is fine for determining the cost per die, but ignores the value of the wafer. The question seems to be fail however, so without more info its impossible to understand what the OP was talking about.