This is a small sampling of some of the iconic CPUs released over the years. The oscillating circles are used to show the relative cycle time of the CPUs. Use the exponential-scale slider to change the amount of slowdown. I hope you enjoy!
Missing your favorite CPU? Go make a PR on GitHub to get it added in for the next version.
This is meant to be a fun CPU comparison that provides a broad, historic overview of how CPU speed has progressed over time, and it's quite impressive! Even so, there's a number of "gotchas" people may point out with how the comparison is being made. I get it - comparing things like this can be tricky. I'll point out a few caveats here for the sake of general awareness. There's probably others, but a few things to consider:
Measuring the speed of a processor purely by it's clock speed is nowhere near the whole picture. One of of several reasons for this is that not all instructions execute in the same amount of time. In addition, not all systems support the same types of instructions. Therefore, it's possible for a CPU with a slower clock speed to perform faster than another system with a higher clock speed but a less robust instruction set.
As en extreme example of this, consider that the ENIAC ran at a clock speed of 100,000 cycles per second but this only translated to about 5,000 instructions per second. This mean it took approximately 20 clock cycles per "instruction" to execute. Compare that to modern machines, where an ADD instruction typically takes anywhere from 1 to 4 cycles.
The rate pf increase in raw, straight-line speed of CPUs has slowed significantly in the past 15 years or so. Much of this is due to physical limitations. Much of how CPUs have gotten better over the past 2 decades has not been through clock speed improvements, but through increased parallelism. In the early 200s, basically all consumer CPUs were single-core machines. Nowadays, most computers being release have at least 4 cores, if not 8, 12, or 16. Some companies, such as Intel offer CPUs with over over 100 cores.
Beyond what was already mentioned, there's many other factors that play a part in practical CPU performance. Factors like cache sizes and latencies (L1, L2, L3), RISC vs CISC instruction sets, instruction bit width (8, 16, 32, 64), and many other play a part.
Made by Ben