In Dec. 2005, I bought a new CPU for $389. It was an AMD Athlon 64 X2, with dual 2.0GHz processors. I also bought 1GB of RAM for $120.
In September 2008, I bought a new CPU. It was an AMD Phenom II X4, with quad 3.0 GHz processors, for $190. I also bought 8GB of RAM for $270.
Now I’m looking for a newer CPU and more RAM. I can buy an AMD Phenom II X6, with six 3.2 GHz processors, for $170. I can also buy 16GB of RAM for $100.
Dec. 2005 was 5 years and 10 months ago. So we should have had 3 to 4 Moore doublings. From .0333 GHz/$, that would take us to .266 to .533 GHz/$. From 8.3 MB/$, we should have gone to 66.4 or 132.8 MB/$.
Instead, I am at 6x3.2 GHz/$170 = .113 GHz/$ and 160 MB/$. CPU speed is lagging far behind Moore’s Law, while RAM is outstripping it.
Both depend on transistor density. So why are they so different?
Why is memory but not CPU following Moore’s law?
In Dec. 2005, I bought a new CPU for $389. It was an AMD Athlon 64 X2, with dual 2.0GHz processors. I also bought 1GB of RAM for $120.
In September 2008, I bought a new CPU. It was an AMD Phenom II X4, with quad 3.0 GHz processors, for $190. I also bought 8GB of RAM for $270.
Now I’m looking for a newer CPU and more RAM. I can buy an AMD Phenom II X6, with six 3.2 GHz processors, for $170. I can also buy 16GB of RAM for $100.
Dec. 2005 was 5 years and 10 months ago. So we should have had 3 to 4 Moore doublings. From .0333 GHz/$, that would take us to .266 to .533 GHz/$. From 8.3 MB/$, we should have gone to 66.4 or 132.8 MB/$.
Instead, I am at 6x3.2 GHz/$170 = .113 GHz/$ and 160 MB/$. CPU speed is lagging far behind Moore’s Law, while RAM is outstripping it.
Both depend on transistor density. So why are they so different?