01-parallelization.Rmd

# Parallelization

Whenever possible, *CopyKit* uses the [BiocParallel](https://www.bioconductor.org/packages/devel/bioc/vignettes/BiocParallel/inst/doc/Introduction_To_BiocParallel.pdf) framework. 

Running parallel processes is highly recommended to speed up *CopyKit* calculations.

Make sure to read the BiocParallel [vignette](https://www.bioconductor.org/packages/devel/bioc/vignettes/BiocParallel/inst/doc/Introduction_To_BiocParallel.pdf) to choose the best parameters for your machine and OS. 

Use `BiocParallel::register()` function to register the number of workers (threads) desired.
To visually monitor the progress, you can use the argument *progressbar*. 

```{r registering_fake, eval=FALSE}
library(BiocParallel)
register(MulticoreParam(progressbar = T, workers = 8), default = T)
```

```{r registering_true, echo=FALSE}
library(BiocParallel)
register(MulticoreParam(progressbar = F, workers = 30), default = T)
```

Confirm parameters:
```{r default, eval=FALSE}
BiocParallel::bpparam()
```

*Note about plotHeatmap() function parallelization*:
**plotHeatmap()** uses a different framework and requires the argument *n_threads* to set the number of threads when in use with the argument *order_cells = hclust*. This is to speed up the calculation of a distance matrix. Using more threads when ordering the heatmap with *hclust* will significantly speed up the plotting of the heatmap. See example in plotHeatmap help.