Skip to main content
Figure 3 | BMC Bioinformatics

Figure 3

From: The top-scoring ‘N’ algorithm: a generalized relative expression classification method from small numbers of biomolecules

Figure 3

GPU vs. CPU running times. Running times for N = 2, N = 3, and N = 4 over a range of input feature sizes. Each point is the mean of three independent runs of the software. The CPU running time for N = 2 over 20,000 features is similar to the running times for N = 3 over 1000 features and N = 4 over 200 features. The CPU version of TSN was run on a single core of a 2.4 GHz Intel Core 2 processor. The GPU version of TSN was run on an NVIDIA Tesla C2050. The speedup due to the GPU improves as the value of N gets higher: for N = 2, the speedup is 2.3X, for N = 3 the speedup is 2.8X, and for N = 4 the speedup is 4.4X. Running times reflect a single iteration of the algorithm and do not include multiple iterations such as cross validation. Note that running times are also a function of the number of samples in the dataset; there were 70 samples in this dataset.

Back to article page