It's not the increase in transistor count that matters. It's the transistor count.
Imagine what AMD could have achieved with that density in 2007.
I'll quote a famous movie here. "I don't think that word means what you think it means."
If you're arguing that transistor density, as a way to approximate transistor count and assuming planar transistors, is the reason Intel has bettered their graphics you are demonstrably incorrect. I proved that a 9% increase in raw transistor count goes from meh graphical fidelity to very decent performance. This change occured in less than 5 years, which means no substantial shift in software could account for the increased performance.
Assuming that your argument is that simply throwing more transistors at the problem is the solution, that is also a demonstrably incorrect. A 10000% increase in transistor count had us go from 2.5 dimensional characters to rough 3 dimensional characters. This took a decade, patents limiting competition, and fundamentally new software.
Where is your argument? I cannot see any salient point, unless your argument is that Intel has discovered a fundamentally different way to build transistors in the last 5 years, which allow them to work much faster.
As to AMD having that transistor count in 2007, who gives a crap? That transistor count would have cost AMD an insane fortune. As they had only recently acquired ATI, where was that money going to come from? AMD made the only fiscally responsible choice, and they put out what they could afford. It wasn't the fastest chip, it was a budget performer. Phenom competed well enough with Core2, if only in the budgetary arena. None of this even begins to address our current situation.
My God, this is so pathetic. Seriously, how many unlocked CPU users need the iGPU? I am so sick of Intel's bullshit decisions. the iGPU takes up more than half the die, and most people have no use for it.
"Most people" is a useless term. You don't provide any view of what this elusive majority is. You don't even cite a basic knowledge of the market. Let's rectify that, for your sake, and those who espouse the same ideology.
http://www.macrumors.com/2014/11/07/apple-mac-us-pc-record/
-This article cites an IDC study, in which the top two PC creators are Lenovo and HP.
http://www.statista.com/statistics/233818/share-of-product-sales-in-total-sales-of-lenovo/
-This article shows that most of Lenovo's products are business oriented. In decline, year after year, is the Desktop PC. The desktop is where a GPU added would not significantly impact lifespan of the product, as they would be plugged in constantly.
It therefore is reasonable to conclude the following:
1) Most PC sales are driven by business users.
2) Most PC sales are of goods where increased battery life would be a big benefit.
3) iGPUs can save a substantial amount of energy, by removing a discrete GPU.
4) Intel's largest market is the one they cater to, and it is the market of business users.
It is understandable that a power user, on a desktop, would not want an iGPU. It is not reasonable for Intel to design a radically different piece of silicon for markets that would not have enough sales to justify the expenditure. It is not reasonable for business users to lug 10 pounds of battery around, so they can have a 3 hour battery life to look at a bunch of spreadsheets (business usage, in a nut shell). Intel caters to their largest consumer base, they include an iGPU, and you have the chutzpah to say their decisions are "bullshit." Perhaps a moment more introspection, and a bit less focus on what you believe you are entitled to, is something sorely lacking. You may not like the suggestion, but people rarely like to be called on their entitlement.
