The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
https://doi.org/10.32614/CRAN.package.bigalgebra
bigalgebra
provides fast linear algebra primitives that
operate seamlessly on base matrix
objects and
[bigmemory::big.matrix
] containers. The package wraps BLAS
and LAPACK routines with R-friendly helpers so that vector updates,
matrix products, and classic decompositions work the same way in memory
or on disk.
dset()
, dsub()
and ddot()
extend
familiar vector algebra to big.matrix
inputs. See the Level
1 BLAS-Style Helpers vignette.dgemm()
and dsymm()
expose Level 3 BLAS
routines for dense matrix multiplication with optional file-backed
outputs. Explore the Matrix
Wrapper Helpers vignette.dgeqrf()
, dpotrf()
,
dgeev()
, dgesdd()
) bring advanced
factorisations to large datasets. Walk through the LAPACK
Decompositions vignette.The package defines a number of global options that begin with
bigalgebra
:
Option Default value * bigalgebra.temp_pattern
with
default matrix_
* bigalgebra.tempdir
with
default tempdir
*
bigalgebra.mixed_arithmetic_returns_R_matrix
with default
TRUE
* bigalgebra.DEBUG
with default
FALSE
The bigalgebra.tempdir
option must be a function that
returns a temporary directory path used to store big matrix results of
BLAS and LAPACK operations. The default value is simply the base R
tempdir()
function.
The bigalgebra.temp_pattern
option is a name prefix for
file names of generated big matrix objects output as a result of BLAS
and LAPACK operations.
The bigalgebra.mixed_arithmetic_returns_R_matrix
option
determines whether arithmetic operations involving an R matrix or vector
and a big.matrix
matrix or vector return a big matrix (when
the option is FALSE
), or return a normal R matrix
(TRUE
).
The package is built, by default, with R’s native BLAS libraries, which use 32-bit signed integer indexing. The default build is limited to vectors of at most 2^31 − 1 entries and matrices with at most 2^31 − 1 rows and 2^31 − 1 columns (note that standard R matrices are limited to 2^31 − 1 total entries).
The package includes a reference BLAS implementation that supports 64-bit integer indexing, relaxing the limitation on vector lengths and matrix row and column limits. Installation of this package with the 64-bit reference BLAS implementation may be performed from the command-line install:
REFBLAS=1 R CMD INSTALL bigalgebra
where bigalgebra
is the source package (for example,
bigalgebra_0.9.0.tar.gz
).
The package may also be built with user-supplied external BLAS and
LAPACK libraries, in either 32- or 64-bit varieties. This is an advanced
topic that requires additional Makevars
modification, and
may include adjustment of the low-level calling syntax depending on the
library used.
Feel free to contact us for help installing and running the package.
This website, the unit tests, some C code fixes and improvements as well as these examples were created by F. Bertrand.
Maintainer: Frédéric Bertrand frederic.bertrand@lecnam.net.
You can install the released version of bigalgebra from CRAN with:
install.packages("bigalgebra")
You can install the development version of bigalgebra from GitHub with:
::install_github("fbertran/bigalgebra") devtools
The snippets below mirror the worked examples in the vignettes and show how the helpers behave with in-memory and file-backed matrices.
These helpers cover vector updates, reductions, and element-wise
transforms such as the in-place square root provided by
dsqrt()
.
library(bigmemory)
library(bigalgebra)
<- bigmemory::big.matrix(5, 1, init = 0)
x dset(ALPHA = 9, X = x)
dsqrt(X = x)
x[]#> [1] 3 3 3 3 3
<- bigmemory::big.matrix(5, 1, init = 1)
y dvcal(ALPHA = 0.5, X = x, BETA = 2, Y = y)
y[]#> [1] 3.5 3.5 3.5 3.5 3.5
dgemm()
<- bigmemory::big.matrix(5, 4, init = 1)
A <- bigmemory::big.matrix(4, 4, init = 2)
B <- bigmemory::big.matrix(5, 4, init = 0)
C
dgemm(A = A, B = B, C = C, ALPHA = 1, BETA = 0)
C[]#> [,1] [,2] [,3] [,4]
#> [1,] 8 8 8 8
#> [2,] 8 8 8 8
#> [3,] 8 8 8 8
#> [4,] 8 8 8 8
#> [5,] 8 8 8 8
set.seed(1)
<- matrix(rnorm(9), 3)
M <- crossprod(M)
SPD <- as.big.matrix(SPD)
SPD_big dpotrf(A = SPD_big)
#> [1] 0
<- SPD_big[,]
chol_factor lower.tri(chol_factor)] <- 0
chol_factor[
chol_factor#> [,1] [,2] [,3]
#> [1,] 1.060398 -0.2388263 -0.6138286
#> [2,] 0.000000 1.8082109 0.2222424
#> [3,] 0.000000 0.0000000 0.8294922
big.matrix
workflows<- tempdir()
tmpdir <- filebacked.big.matrix(3, 3, init = diag(3),
file_big backingpath = tmpdir,
backingfile = "example.bin")
#> Warning in filebacked.big.matrix(3, 3, init = diag(3), backingpath = tmpdir, : No
#> descriptor file given, it will be named example.bin.desc
1, 3] <- 5
file_big[
file_big[]#> [,1] [,2] [,3]
#> [1,] 1 1 5
#> [2,] 1 1 1
#> [3,] 1 1 1
rm(file_big)
gc()
#> used (Mb) gc trigger (Mb) limit (Mb) max used (Mb)
#> Ncells 900616 48.1 1699095 90.8 NA 1377768 73.6
#> Vcells 2253326 17.2 8388608 64.0 65536 3247956 24.8
The full vignette set expands on the topics above and demonstrates how the routines interact:
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.