Yes I stopped reading because you are making an assumption that goes against what the manufacturer is saying and is in fact less than what was announced if you trust what DF and MS are saying. I didn't say it's a negative article and I didn't refer to it or identify it as such. If I am going to read an analysis piece I think it's important (for myself) that the commentary is done based on what we know and not what we assume. In this case we know it's a highly customized Jaguar CPU.
And how does it relate to fill rate? Seriously we don't know how it is customized so we cannot make any assumption. We do know that it is based on Jaguar. That way we can make a conservative estimate, which means we stay on the safe side.
Especially on a requirement that that has nothing to do with that particular part of the project.
But let's take Computers out of this and go to Engineering.
We are building a suspension bridge and we are discussing the support characteristics of the concrete pillars. One of the Engineers says :"Look we know there is this new Alloy, our supplier said it is all the rage but hasn't given any details yet. It is better than the previous alloy, but let's for now just assume it has the same tensile strength and get on with our requirements, for the concrete pillars".
Then some Engineer comes in and goes "NOOO! LET US DOUBLE THE TENSILE STRENGTH! The Manufacturer says it's better than the previous alloy!".
Just to be done with it, the other engineers decide to double the tensile strength specification in the books. It has no effect on the concrete anyway even though it has no effect on the concrete pillars itself. Everybody continues but the specification stays in the records. We use this new alloy and it is magnificent, 70% more tensile strength and a reduction of weight. However on certain parts of the bridge they have use the requirement possibility of "double the tensile strength" as some other engineers looked it up in the books and came about thosenumbers. They didn't check with anybody they just used it in their calculations.
This goes well for ten years, but one day the bridge collapses. Hundreds dead, thousands injured.
So yes, it is better to steer on the conservative side and calculate with what we know, especially in engineering. With a Game Console nobody can die, but in Engineering and even Computer Science there can be lives on the line. For instance the logical steps in a the driver program for a CT machine. There are requirements to be met, like tensile strength, which then dictates the thickness of the steel wire. And in CT machine amount of X-Ray Radiation, so you don't flood the patient with too much radiation and cause for irreparable damage.
You have requirements and those must be met, to meet them you sometimes need to clearly assume the old specification.
And before you go "this doesn't matter.". It does, many lives have been lost due to these kind of decisions and people making a mistake, meaning well, being really motivated, not asking just assuming the numbers are correct, the list goes on.
It isn't just for ease of comparison, but it also is just a good practice. Until Microsoft tells us what the specifications are of their CPU, which will be filled with buzzwords so I'd need time to study it, we can't assume anything.