Table 1 Relative performance of the sparse optimization on 2000 genes and 2000 cells, baseline is the standard algorithm with 1 thread and no sparse optimization

From: CoGAPS 3: Bayesian non-negative matrix factorization for single-cell analysis with asynchronous updates and sparse data structures

Data sparsity (%)	Memory (MB)	Runtime (1 thread)	Runtime (4 threads)
70	0.14	1.92	0.62
80	0.09	1.24	0.42
90	0.04	0.42	0.20

Back to article page