I used the wrong graph for that crap and the scaling in this one is more relative Ph2 -> FX
An AMD Zambezi Module provides 200% performance most programs can use all(100% per core)
An AMD Deneb/Zosma core provides 134% performance but most programs only used (2/3)s of that performance (90% per core)
Zambezi
100% -> 200% -> 300% -> 400% -> 500% -> 600% -> 700% -> 800%
Deneb/Zosma
134% -> 268% -> 402% -> 536% -> 670% -> 804%
Most programs(99% of the programs you use):
90% -> 180% -> 270% -> 360% -> 450% -> 540%
More realistic and easier to under the change from CMP to CMT
AMD could have easily stayed on CMP and just changed the "phenom II" core but
8x100% = 200% larger die
6x100% = 600%
4x150% = 600%
8 cores on one die with shared components instead of upgrading the socket to allow for a much larger die
Maybe we will see 4 core modules
3x200% = 600%
One warning: Do not compare AMD architectures to Intel architectures, vice versa
Brain seizures happen and you lose reasoning and logic skills and make math errors