3. WS denerg33184 (a.k.a Titan)

3.1. Overview

Prometheus WS

3.2. Benchmark

We compute the scalability efficiency with the Eq. (3.1)

(3.1)\[\eta = \dfrac{p_N}{p_1 \cdot N} \cdot 100\]

where \(N\) is the number of cores, \(p_N\) is the performance measure for the test with \(N\) cores (e.g ns/day for MD simulations) and \(p_1\) is the performance measured with only one core.

3.2.1. Tensorflow 2

Tensorflow 2 Benchmark

Fig. 3.1 Tensorflow 2 Benchmark results for WS Titan.