High Performance Numerical Computing for High Energy Physics: A New Challenge for Big Data Science

Table 2

Isoefficiency for a hypercube architecture: 𝑛=Θ(𝑝log𝑝) and 𝑛=Θ(𝑝(log𝑝)²). We marked with (*) the limitations imposed by Formula (33).

Scenario	Architecture size (𝑝)	𝑛=Θ(𝑝log𝑝)	𝑛=Θ(𝑝(log𝑝)^2)
1	10^1	1.0×10^1	1.00×10^1
2	10^2	2.0×10^2	8.00×10^2
3	10^3	3.0×10^3	2.70×10^4
4	10^4	4.0×10^4	6.40×10^5∗
5	10^5	5.0×10^5∗	1.25×10^7
6	10^6	6.0×10^6	2.16×10^8
7	10^7	7.0×10^7	3.43×10^9
8	10^8	8.0×10^8	5.12×10^10
9	10^9	9.0×10^9	7.29×10^11