On multi-core machines, a single model can be divided into smaller parts, all running simultaneously. ASReml-R 4.2 has been optimised for multi-threaded processing, enabling you to use multiple threads. This can significantly improve processing speed, allowing you to fit models faster.
You can also increase the number of concurrent sessions to further improve performance. Each license has 2 sessions by default, but you can contact our support team to increase this number.
Memory access (workspace)
Memory access has been increased, enabling you to analyse larger problems. On multi-user systems, memory efficiency is maximised by allowing each user to specify the amount of memory needed for their current session. (This maximum allocation depends on what is available on your PC and your operating system).
Optimisation of internal routines
Several core routines from ASReml-R 4.2 have been reworked, making it much faster than the previous versions. For example, in jobs with relatively dense genomic matrices computation times are often reduced by more than 40%.
This new version of ASReml-R has enhanced stability and optimisation. We have also strengthened our software testing tools, for example, by expanding our model tests with complex and large datasets.
Improvements have been made to the R help files and documentation from ASReml-R and its associated functions. Additional technical details and code examples will facilitate the fitting of models, their interpretation, and obtaining the required output.
|Type of analyses||4.1 Elapsed time (min)||4.2 Elapsed time (min)||Ratio gain 4.2 vs 4.1||Dataset size
|Univariate Animal model GBLUP||475.5||18.8||25.3||7,643|
|Multi-Environment Trial (MET) GBLUP||277.1||42.3||6.5||28,920|
|Multi Environment Trial (MET) ABLUP||136.8||3.3||41.8||12,678|
|Univariate Animal Model ABLUP||64.9||30.9||2.1||263,864|
|Bivariate Animal Model ABLUP||50.4||24.1||2.1||263,864|
|Type of analyses||# genotypes (phenotypic)||# genotypes (relationship)||# environment/ experiment||Model keywords|
|Univariate Animal model GBLUP||7,643||7,643||1||GBLUP|
|Multi-Environment Trial (MET) GBLUP||5,244||7,250||40||GBLUP, MET, multiple designs, heterogeneous variances|
|Multi Environment Trial (MET) ABLUP||884||884||24||ABLUP, factor analytic, multi-experiment|
|Univariate Animal Model ABLUP||263,864||396,734||1||ABLUP, random regression, heterogeneous variances|
|Bivariate Animal Model ABLUP||263,864||396,734||1||ABLUP, multi-trait|
- Computer specifications: Intel(R) Core(TM) i7-3770K CPU @ 3.50GHz, 16 Gb of RAM, Ubuntu 20.04 LTS.
- Analyses were carried out using these settings in ASReml-R 4.2.
asreml.options(threads = 1, use.blas = “standard”)
Further speed improvements can be achieved by increasing the number of threads and enabling Intel’s Math Kernel Library (MKL).