Skip to main content

Table 3 Training time (in seconds) for a single iteration under different implementations of MedGP

From: Sparse multi-output Gaussian processes for online medical time series prediction

Implementation Sequential Multithreading
Computing Gram matrix 11 2
Inverting Gram matrix 13 3
Computing gradients 2497 97
Total per iteration 2521 102
  1. The total number of observations across time for this patient is 6679. The sequential test used a single CPU, while the multithreading test used 35 CPUs—one thread per CPU