Walkthrough on PCA through the command line - Revision history

Andrés Molina: /* GPU computing */

2025-11-24T15:38:20Z

GPU computing

2024-12-02T12:50:32Z

2021-02-08T21:58:49Z

Size of parallel blocks

2020-05-04T15:27:33Z

Correlation matrix

2020-05-04T15:26:50Z

Correlation matrix

2020-05-04T15:24:05Z

Custom approach

2020-05-04T15:21:58Z

TSNE reduction

2020-04-06T11:50:39Z

Syntax

2020-04-05T11:19:49Z

Utilities to set the batch number

2020-04-05T10:30:37Z

@@ Line 54: / Line 54: @@
   <tt>wb.settings.computing.useGpus.value = true;</tt>
 and the specify the device numbers to use:
-  <tt>wb.settings.computing.gpus.value = [0:7];</tt>
+  <tt>wb.settings.computing.gpus.value = [1:8];</tt>
-In this example, you would be using 8 GPUs, marked by the systems with numbers 0 to 7.
+In this example, you would be using 8 GPUs, marked by the systems with numbers 1 to 8.
 Take into account that your system needs an independent CPU for each used GPU.

← Older revision		Revision as of 12:50, 2 December 2024
Line 1:		Line 1:
		+	[[Category:Walkthroughs]]
		+
	[[Principal component analysis\|PCA]] computations through the command line are governed through ''PCA workflow'' objects. We describe here how to create and handle them:		[[Principal component analysis\|PCA]] computations through the command line are governed through ''PCA workflow'' objects. We describe here how to create and handle them:

@@ Line 62: / Line 62: @@
 meaning that each block executed in parallel will preload two sets of 100 (prealigned) particles into memory, and then proceed to cross correlate all possible pairs. Ideally, you should tune your <tt>batch</tt> parameters l to have as many blocks as processing units (CPU cores or GPUs).
 IThis might not be possible, as it might crowd the memory of your device. Thus, in real applications you should aim at using the <tt>batch</tt> parameter to enforce a tradeoff between performance and  memory usage suited to your system.
 ==== Utilities to set the batch number ====

@@ Line 103: / Line 103: @@
 == Correlation matrix==
 <tt>m=wb.getCCMatrix();</tt>
-<tt>figure;dshow(m);h=gca();h.YDir = 'reverse'; </tt>
+<tt>figure;dshow(m);h=gca();</tt>
 == Eigencomponents ==

@@ Line 106: / Line 106: @@
 <tt>m=wb.getCCMatrix();</tt>
-<tt>figure;dshow(cmm);h=gca();h.YDir = 'reverse'; </tt>
+<tt>figure;dshow(m);h=gca();h.YDir = 'reverse'; </tt>
 == Eigencomponents ==

@@ Line 133: / Line 133: @@
 === Custom approach===
 You can use your own methods to visualize the eigencomponents. They can be accessed through:
-  <tt>m = w.getEigencomponents();</tt>
+  <tt>m = wb.getEigencomponents();</tt>
 will produce a matrix <tt>m</tt> where each column represents an eigenvector and each row a particle. Thus, to see how a particular eigencomponent distributes among the particles, you can just write:

@@ Line 96: / Line 96: @@
 == TSNE reduction ==
 TSNE remaps the particles into 2D maps which can be visualised and operated interactively.
-  <tt>wb.steps.items.tsene.compute(); </tt>
+  <tt>wb.steps.items.tsne.compute(); </tt>
 =Visualization=

@@ Line 28: / Line 28: @@
   <tt>name = 'classtest128';</tt>
 Now we are ready to create the workflow:
-   <tt>wb = dpkpca.new(name,'t',tableFile,'d',dataFolder,'m',mask);]]</tt>
+   <tt>wb = dpkpca.new(name,'t',tableFile,'d',dataFolder,'m',mask);</tt>
 This creates an workflow object (arbitrarily called <tt>wb</tt>) in the workspace during the current session). It also creates a folder called classtest128.PCA where results will be stored as they are produced.

@@ Line 66: / Line 66: @@
 You can get some guidance for the selection of the batch number with
   <tt>wb.getBlockSize()</tt>
 which will return the size of each block (in Gibabytes) for the current batch parameter.
 == Unfolding==