AI & Computing

Amdahl's Law Calculator (Parallel Speedup)

Enter the parallel share of the work and the number of processors to see the speedup limit.

Your numbers

%

Your results

Speedup (Amdahl)
Speedup (Amdahl)
—
Parallel efficiency
—
Limit with unlimited processors
—
Scaled speedup (Gustafson)
—

Amdahl's Law Calculator (Parallel Speedup) by TheCalcTools

How the Amdahl's Law Calculator (Parallel Speedup) works

Amdahl’s law says the serial part of a program limits the overall speedup no matter how many processors you add: speedup = 1 ÷ ((1 − p) + p ÷ n).

If 90% of the work runs in parallel, the speedup can never exceed 10×. Gustafson’s law takes the more optimistic view that larger machines are used for larger problems.

The formula

S = 1 ÷ ((1 − p) + p ÷ n); maximum = 1 ÷ (1 − p)

Worked example

Using the default values shown in the calculator:

Parallelisable share of the work90 %
Processors or cores16
  1. 90% parallel on 16 cores: 1 ÷ (0.1 + 0.9 ÷ 16) = 6.40×.
  2. Even with unlimited cores the maximum is 10.00×.

Result:

Speedup (Amdahl)6.40
Parallel efficiency40.0%
Limit with unlimited processors10.00×
Scaled speedup (Gustafson)14.50

Frequently asked questions

Why does adding cores give diminishing returns?

Each extra core only speeds up the parallel part, so the fixed serial part takes a growing share of the time.

Sources and further reading

More in AI & Computing · All calculators

This calculator gives estimates for information only. Check important results independently before you rely on them. See our disclaimer.