I’ve got an 8-core i7 running at 3.6 GHz.
I used it to do the math. 8*3.6 GHz = 28.8 GHz, or 0.0288 THz. That’s 576 times faster than 50 MHz.
576 << 200000.
And its throughput is nowhere near that good due to coordination overhead between the CPUs.
576x = difference in clock cycles per second
200,000x = claimed difference in performance
Two totally different animals.
How deep was the pipeline on the old bird? How many registers? FPU? Memory access time? L1-L3 cache? Advances in OS and compilers (probably offset somewhat by sloppier programming). What about swapping?