Help for package MPCR

Type:

Package

Title:

Multi Precision Computing

Version:

2.1.1

Date:

2026-03-15

Author:

David Helmy [aut, cph], Sameh Abdulah [cre, cph], KAUST King Abdullah University of Science and Technology [fnd, cph], Brightskies [cph]

Maintainer:

Sameh Abdulah <sameh.abdulah@kaust.edu.sa>

Description:

Provides new data-structure support for multi-precision computing for R users. The package supports 16-bit, 32-bit, and 64-bit operations. To the best of our knowledge, 'MPCR' differs from the currently available packages in the following: 'MPCR' introduces a new data structure that supports three different precisions (16-bit, 32-bit, and 64-bit), allowing for optimized memory allocation based on the desired precision. This feature offers significant advantages in memory optimization. 'MPCR' extends support to all basic linear algebra methods across different precisions. Optional GPU acceleration via CUDA is available for 32-bit and 64-bit operations when CUDA Toolkit is detected during installation, while 16-bit operations are GPU-only and limited to matrix-matrix multiplication. 'MPCR' maintains a consistent interface with normal R functions, allowing for seamless code integration and a user-friendly experience.

License:

GPL (≥ 3)

Imports:

methods, Rcpp (≥ 1.0.9)

Depends:

R (≥ 3.6.0)

RoxygenNote:

7.2.3

VignetteBuilder:

R.rsp

Suggests:

R.rsp

SystemRequirements:

C++17, CMake (>=3.20), CUDA (>=11.2)

NeedsCompilation:

yes

OS_type:

unix

URL:

https://github.com/stsds/MPCR

Packaged:

2026-03-15 02:27:01 UTC; ali

Repository:

CRAN

Date/Publication:

2026-03-15 12:10:20 UTC

MPCR S4 Class

Description

MPCR is a multi-precision vector/matrix, that enables the creation of vector/matrix with three different precisions (16-bit (half), 32-bit(single), and 64-bit(double)) on CPU and GPU. MPCR uses smart cache memory allocation system that minimize the number of memory transfers required during the transition from CPU to GPU operations and vice verse. So the object at some point can be allocated on both CPU and GPU, in case, the user needed to free specific allocation, helper functions can be used.

Value

MPCR object (constructor - accessors - methods)

Constructor

new Creates a new instance of zero values of the MPCR class. new(MPCR,size, "precision","placement")

size: The total number of values for which memory needs to be allocated.
precision: String to indicate the precision of MPCR object ("half","single", or "double").
placement: String to indicate whether the allocation should be made on CPU (default) or GPU ("CPU","GPU") .

Accessors

The following accessors can be used to get the values of the slots:

IsMatrix: Boolean to indicate whether the MPCR object is a vector or matrix.
Size: Total number of elements inside the object, (row*col) in the case of matrix, and number of elements in the case of vector.
Row: Number of rows.
Col: Number of cols.

Methods

The following methods are available for objects of class MPCR:

PrintValues

PrintValues(): Prints all the values stored in the matrix or vector, along with metadata about the object.

ToMatrix

ToMatrix(row,col): Changes the object representation to match the new dimensions, no memory overhead.

ToVector

ToVector(): Changes the MPCR matrix to vector, no memory overhead.

IsGPUAllocated

IsGPUAllocated(): Returns TRUE if the MPCR object is allocated on GPU.

IsCPUAllocated

IsCPUAllocated(): Returns TRUE if the MPCR object is allocated on CPU.

FreeGPU

FreeGPU(): Free the data allocated on GPU.

FreeCPU

FreeCPU(): Free the data allocated on CPU.

Examples

  
    # Example usage of the class and its methods
    library(MPCR)
    MPCR_object <- new(MPCR,50,"single","CPU")

    MPCR_object$ToMatrix(5,10)
    MPCR_object$Row       #5
    MPCR_object$Col       #10
    MPCR_object$Size      #50
    MPCR_object$IsMatrix  #TRUE

    MPCR_object$PrintValues()
    MPCR_object$ToVector()
    MPCR_object$IsCPUAllocated() #TRUE
    MPCR_object$IsGPUAllocated() #FALSE

    MPCR_object

Converters

Description

Converters from R to MPCR objects and vice-versa.

Value

An MPCR or R numeric vector/matrix.

MPCR Converter

Convert R object to MPCR object.

MPCR converters

as.MPCR(data,nrow = 0,ncol = 0,precision,placement): Converts R object to MPCR object.

data: R matrix/vector.
nrow: Number of rows of the new MPCR matrix, default = zero which means a vector will be created.
ncol: Number of cols of the new MPCR matrix, default = zero which means a vector will be created.
precision: String indicates the precision of the new MPCR object (half, single, or double).
placement: String indicates whether the data should be allocated on CPU (default) or GPU ("CPU", "GPU")

R Converter

Convert an MPCR object to R object.

R vector converter

MPCR.ToNumericVector(x): Converts an MPCR object to a numeric R vector.

x: MPCR object.

R matrix converter

MPCR.ToNumericMatrix(x): Converts an MPCR object to a numeric R matrix.

x: MPCR object.

Examples

   # Example usage of the class and its methods
   library(MPCR)
   a <- matrix(1:36, 6, 6)
   MPCR_matrix <- as.MPCR(a,nrow=6,ncol=6,precision="single", placement="CPU")
   r_vector <- MPCR.ToNumericVector(MPCR_matrix)
   r_vector
   r_matrix <- MPCR.ToNumericMatrix(MPCR_matrix)
   r_matrix

Binary arithmetic numeric/MPCR objects.

Description

Binary arithmetic for numeric/MPCR objects.

Usage

## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
e1 + e2

## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
e1 - e2

## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
e1 * e2

## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
e1 / e2

## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
e1 ^ e2

## S4 method for signature 'Rcpp_MPCR,BaseLinAlg'
e1 + e2

## S4 method for signature 'Rcpp_MPCR,BaseLinAlg'
e1 * e2

## S4 method for signature 'Rcpp_MPCR,BaseLinAlg'
e1 - e2

## S4 method for signature 'Rcpp_MPCR,BaseLinAlg'
e1 / e2

## S4 method for signature 'Rcpp_MPCR,BaseLinAlg'
e1 ^ e2

Arguments

e1, e2

Numeric/MPCR objects.

Value

An MPCR object, matching the data type of the highest precision input.

Examples

library(MPCR)
s1 <- as.MPCR(1:20,nrow=2,ncol=10,"single")
s2 <- as.MPCR(21:40,nrow=2,ncol=10,"double")

x <- s1 + s2
typeof(x) # A 64-bit precision (double) MPCR matrix.

s3 <- as.MPCR(1:20,nrow=2,ncol=10,"single")
x <- s1 + s3
typeof(x) # A 32-bit precision (single) MPCR matrix.

Binary comparison operators for numeric/MPCR objects.

Description

Binary comparison operators for numeric/MPCR objects.

Usage


## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
e1 < e2

## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
e1 <= e2

## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
e1 == e2

## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
e1 != e2

## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
e1 > e2

## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
e1 >= e2

## S4 method for signature 'Rcpp_MPCR,BaseLinAlg'
e1 < e2

## S4 method for signature 'Rcpp_MPCR,BaseLinAlg'
e1 <= e2

## S4 method for signature 'Rcpp_MPCR,BaseLinAlg'
e1 == e2

## S4 method for signature 'Rcpp_MPCR,BaseLinAlg'
e1 != e2

## S4 method for signature 'Rcpp_MPCR,BaseLinAlg'
e1 > e2

## S4 method for signature 'Rcpp_MPCR,BaseLinAlg'
e1 >= e2

Arguments

e1, e2

Numeric/MPCR objects.

Value

A vector/matrix of logicals.

Examples


  library(MPCR)
  s1 <- as.MPCR(1:20,nrow=2,ncol=10,"single")
  s2 <- as.MPCR(21:40,nrow=2,ncol=10,"double")

  x <- s1 > s2

copy

Description

Functions for copying MPCR objects.

Value

An MPCR copy from the input object.

MPCR deep copy

Create a copy of an MPCR object. Typically, using 'equal' creates a new pointer for the object, resulting in any modifications made to object one affecting object two as well.

copy

MPCR.copy(x): Create a new copy of an MPCR object.

x: MPCR object.

Examples


   library(MPCR)
   # Example usage of the class and its methods
   a <- matrix(1:36, 6, 6)
   MPCR_matrix <- as.MPCR(a,nrow=6,ncol=6,precision="single")

   # Normal equal '=' will create a new pointer of the object, so any change in object A
   # will affect object B
   temp_MPCR_matrix = MPCR_matrix
   temp_MPCR_matrix[2,2] <- 500
   MPCR_matrix[2,2]           #500


   MPCR_matrix_copy <- MPCR.copy(MPCR_matrix)
   MPCR_matrix[2,2] <-100
   MPCR_matrix_copy[2,2] <- 200

   MPCR_matrix[2,2]           #100
   MPCR_matrix_copy[2,2]      #200

dimensions

Description

Returns the number of rows or cols in an MPCR object.

Usage

## S4 method for signature 'Rcpp_MPCR'
nrow(x)

## S4 method for signature 'Rcpp_MPCR'
ncol(x)

Arguments

x

An MPCR object.

Value

The number of rows/cols in an MPCR object.

Examples

    library(MPCR)
    x <- as.MPCR(1:16,4,4,"single")
    y <- as.MPCR(1:20,4,5,"double")
    rows_x <- nrow(x)
    cols_y <- ncol(y)

Extract or replace elements from an MPCR object.

Description

Extract or replace elements from an MPCR object using the '[', '[[', '[<-', and '[[<-' operators. When extracting values, they will be converted to double precision. However, if you update a single object, the double value will be cast down to match the precision. If the MPCR object is a matrix and you access it using the 'i' index, the operation is assumed to be performed in column-major order, or using 'i' and 'j' index.

Usage

  ## S4 method for signature 'Rcpp_MPCR'
x[i, j, drop = TRUE]
  ## S4 replacement method for signature 'Rcpp_MPCR'
x[i, j, ...] <- value
  ## S4 method for signature 'Rcpp_MPCR'
x[[i, drop = TRUE]]
  ## S4 replacement method for signature 'Rcpp_MPCR'
x[[i, ...]] <- value

Arguments

x

An MPCR object.

i

Row index or indices.

j

Column index or indices.

...

ignored.

drop

ignored.

value

A value to replace the selected elements with.

Examples

  library(MPCR)
    x <-as.MPCR(1:50,precision="single")
    ext <- x[5]
    x[5] <- 0
    x$ToMatrix(5,10)
    x[2,5]
    x[3,5] <- 100

concatenate

Description

c() function for MPCR objects.

Usage

## S4 method for signature 'Rcpp_MPCR'
MPCR.Concatenate(x)

Arguments

x

List of MPCR objects.

Value

MPCR object containing values from all objects in the list.

Examples

library(MPCR)
x <- as.MPCR(1:20,precision="single")
y <- as.MPCR(1:20,precision="single")
list <- c(x,y)
new_obj <- MPCR.Concatenate(list)

bind

Description

rbind() and cbind() for MPCR objects.

Usage

## S4 method for signature 'Rcpp_MPCR'
MPCR.rbind(x,y)

## S4 method for signature 'Rcpp_MPCR'
MPCR.cbind(x,y)

Arguments

x

An MPCR object.

y

An MPCR object.

Value

An MPCR object, matching the data type of the highest precision input.

Examples

library(MPCR)
# create 2 MPCR matrix a,b
a <- as.MPCR(1:20,nrow=2,ncol=10,"single")
b <- as.MPCR(21:40,nrow=2,ncol=10,"double")

x <- MPCR.rbind(a,b)
y <- MPCR.cbind(a,b)

diag

Description

Returns the diagonal of an MPCR matrix.

Usage

## S4 method for signature 'Rcpp_MPCR'
diag(x)

Arguments

x

An MPCR matrix.

Value

An MPCR vector contains the main diagonal of the matrix.

Examples

    library(MPCR)
    x <- as.MPCR(1:16,4,4,"single")
    diag_vals <- diag(x)

Min-Max Functions

Description

Min-Max functions for MPCR objects values and indices, all NA values are disregarded.

Usage

## S4 method for signature 'Rcpp_MPCR'
min(x)

## S4 method for signature 'Rcpp_MPCR'
max(x)

## S4 method for signature 'Rcpp_MPCR'
which.min(x)

## S4 method for signature 'Rcpp_MPCR'
which.max(x)

Arguments

x

An MPCR object.

Value

Min/max value/index.

Examples


  library(MPCR)

  x <- as.MPCR(1:20,precision="double")
  min <-min(x)
  min_idx <-which.min(x)

Logarithms and Exponentials

Description

exp/log functions.

Usage

## S4 method for signature 'Rcpp_MPCR'
exp(x)

## S4 method for signature 'Rcpp_MPCR'
expm1(x)

## S4 method for signature 'Rcpp_MPCR'
log(x, base = 1)

## S4 method for signature 'Rcpp_MPCR'
log10(x)

## S4 method for signature 'Rcpp_MPCR'
log2(x)

Arguments

x

An MPCR object.

base

The logarithm base. If base = 1, exp(1) is assumed, only base 1,2, and 10 available.

Value

An MPCR object of the same dimensions as the input.

Examples


library(MPCR)

x <- as.MPCR(1:20,precision="double")
log(x)

Finite, infinite, and NaNs

Description

Finite, infinite, and NaNs.

Usage

## S4 method for signature 'Rcpp_MPCR'
is.finite(x)

## S4 method for signature 'Rcpp_MPCR'
is.infinite(x)

## S4 method for signature 'Rcpp_MPCR'
is.nan(x)

Arguments

x

An MPCR object.

Value

A bool vector/matrix of the same dimensions as the input.

Examples

  
  library(MPCR)

  x <- as.MPCR(1:20,precision="double")
  is.nan(sqrt(x))

Miscellaneous mathematical functions

Description

Miscellaneous mathematical functions.

Usage

## S4 method for signature 'Rcpp_MPCR'
abs(x)

## S4 method for signature 'Rcpp_MPCR'
sqrt(x)

Arguments

x

An MPCR object.

Value

An MPCR object of the same dimensions as the input.

Examples


library(MPCR)

x <- as.MPCR(1:20,precision="double")
sqrt(x)

NA's

Description

is.na() ,na.omit(), and na.exclude() for MPCR objects.

Usage

## S4 method for signature 'Rcpp_MPCR'
MPCR.is.na(object,index=-1)
## S4 method for signature 'Rcpp_MPCR'
MPCR.na.exclude(object,value)
## S4 method for signature 'Rcpp_MPCR'
MPCR.na.omit(object)

Arguments

object

MPCR object.

index

If a particular index in the MPCR matrix/vector is specified, it will be checked. If no index is provided, all elements will be checked.

value

Value to replace all NAN with.

Value

MPCR.is.na will return matrix/vector/bool according to input of the function.
MPCR.na.exclude & MPCR.na.omit will not return anything.

Examples

  
library(MPCR)
x <- as.MPCR(1:20,precision="single")
x[1] <- NaN
MPCR.is.na(x,index=1) #TRUE
MPCR.na.exclude(x,50)
x[1]  #50

replicate

Description

Replicates the given input number of times according to count/len , only one should be set at a time, and in case both values are given, only the len value will have effect.

Usage

## S4 method for signature 'Rcpp_MPCR'
rep(x,count=0,len=0)

Arguments

x

An MPCR object.

count

Value to determine how many times the input value will be replicated.

len

Value to determine the required output size, the input will be replicated until it matches the output len size.

Value

MPCR vector containing the replicated values.

Examples


    library(MPCR)
    x <- as.MPCR(1:16,4,4,"single")
    rep_vals_1 <- rep(x,count=2)  #output size will be 16*2
    rep_vals_2 <- rep(x,len=2)  #output size will be 2

Rounding functions

Description

Rounding functions.

Usage

## S4 method for signature 'Rcpp_MPCR'
ceiling(x)

## S4 method for signature 'Rcpp_MPCR'
floor(x)

## S4 method for signature 'Rcpp_MPCR'
trunc(x)

## S4 method for signature 'Rcpp_MPCR'
round(x, digits = 0)

Arguments

x

An MPCR object.

digits

The number of digits to use in rounding.

Value

An MPCR object of the same dimensions as the input.

Examples

  
  library(MPCR)

  input <- runif(20,-1,1)
  x <- as.MPCR(input,precision="double")
  floor(x)

scale

Description

Center or scale an MPCR object.

Usage

## S4 method for signature 'Rcpp_MPCR'
scale(x, center, scale)

Arguments

x

An MPCR object.

center, scale

Logical or MPCR objects.

Value

An MPCR matrix.

Examples

  
  library(MPCR)
    input <-as.MPCR(1:50,precision="single")
    input$ToMatrix(5, 10)
    temp_center_scale <- as.MPCR(1:10,precision="double")
    z <- scale(x=input, center=FALSE, scale=temp_center_scale)

sweep

Description

Sweep an MPCR vector through an MPCR matrix.

Usage

## S4 method for signature 'Rcpp_MPCR'
sweep(x,stat,margin,FUN)

Arguments

x

An MPCR object.

stat

MPCR vector containing the value(s) that should be used in the operation.

margin

1 means row; otherwise means column.

FUN

Sweeping function; must be one of "+", "-", "*", "/", or "^".

Value

An MPCR matrix of the same type as the highest precision input.

Examples


library(MPCR)
x <- as.MPCR(1:20,10,2,"single")
y <- as.MPCR(1:5,precision="double")
sweep_out <- sweep(x, stat=y, margin=1, FUN="+")
MPCR.is.double(sweep_out)  #TRUE

Special mathematical functions.

Description

Special mathematical functions.

Usage

## S4 method for signature 'Rcpp_MPCR'
gamma(x)

## S4 method for signature 'Rcpp_MPCR'
lgamma(x)

Arguments

x

An MPCR object.

Value

An MPCR object of the same dimensions as the input.

Examples


library(MPCR)

x <- as.MPCR(1:20,precision="double")
lgamma(x)

Trigonometric functions

Description

Basic trig functions.

Usage

## S4 method for signature 'Rcpp_MPCR'
sin(x)

## S4 method for signature 'Rcpp_MPCR'
cos(x)

## S4 method for signature 'Rcpp_MPCR'
tan(x)

## S4 method for signature 'Rcpp_MPCR'
asin(x)

## S4 method for signature 'Rcpp_MPCR'
acos(x)

## S4 method for signature 'Rcpp_MPCR'
atan(x)

Arguments

x

An MPCR object.

Value

An MPCR object of the same dimensions as the input.

Examples


library(MPCR)

mpcr_matrix <- as.MPCR(1:20,nrow=2,ncol=10,"single")
x <- sin(mpcr_matrix)

Hyperbolic functions

Description

These functions give the obvious hyperbolic functions. They respectively compute the hyperbolic cosine, sine, tangent, and their inverses, arc-cosine, arc-sine, arc-tangent (or 'area cosine', etc).

Usage

## S4 method for signature 'Rcpp_MPCR'
sinh(x)
## S4 method for signature 'Rcpp_MPCR'
cosh(x)
## S4 method for signature 'Rcpp_MPCR'
tanh(x)
## S4 method for signature 'Rcpp_MPCR'
asinh(x)
## S4 method for signature 'Rcpp_MPCR'
acosh(x)
## S4 method for signature 'Rcpp_MPCR'
atanh(x)

Arguments

x

An MPCR object.

Value

An MPCR object of the same dimensions as the input.

Examples


    library(MPCR)

    mpcr_matrix <- as.MPCR(1:20,nrow=2,ncol=10,precision="single")
    x <- sinh(mpcr_matrix)

transpose

Description

Transpose an MPCR object.

Usage

## S4 method for signature 'Rcpp_MPCR'
t(x)

Arguments

x

An MPCR object.

Value

An MPCR object.

Examples

  
  library(MPCR)
  a <- matrix(1:20, nrow = 2)
  a_MPCR <- as.MPCR(a,2,10,"double")
  a_MPCR_transpose <- t(a_MPCR)

Metadata functions

Description

Checks the precision of a given MPCR object.

Usage


## S4 method for signature 'Rcpp_MPCR'
MPCR.is.single(x)
## S4 method for signature 'Rcpp_MPCR'
MPCR.is.half(x)
## S4 method for signature 'Rcpp_MPCR'
MPCR.is.double(x)
## S4 method for signature 'Rcpp_MPCR'
MPCR.is.float(x)

Arguments

x

An MPCR object.

Value

Boolean indicates the precision of the object according to the used function.

Examples

  library(MPCR)
  x <- as.MPCR(1:20,precision="double")
  MPCR.is.double(x) #TRUE
  MPCR.is.single(x) #FALSE

Metadata functions

Description

Metadata functions.

Usage

## S4 method for signature 'Rcpp_MPCR'
storage.mode(x)
## S4 method for signature 'Rcpp_MPCR'
typeof(x)
## S4 method for signature 'Rcpp_MPCR'
MPCR.object.size(x)
## S4 method for signature 'Rcpp_MPCR'
MPCR.ChangePrecision(x,precision)

Arguments

x

An MPCR object.

precision

String with the required precision.

Value

Prints/change metadata about an MPCR object.

Examples


  library(MPCR)

  x <- as.MPCR(1:20,precision="double")
  typeof(x)
  MPCR.ChangePrecision(x,"single")
  MPCR.is.single(x) #True

print

Description

Prints the precision and type of the object, and print will print the meta data of the object without printing the values. Function x$PrintValues() should be used to print the values."

Usage

## S4 method for signature 'Rcpp_MPCR'
print(x)

## S4 method for signature 'Rcpp_MPCR'
show(object)

Arguments

x, object

An MPCR objects.

Details

Prints metadata about the object and some values.

Value

A string containing the metadata of the MPCR object.

Examples

  
    library(MPCR)
    x <- as.MPCR(1:16,4,4,"single")
    y <- as.MPCR(1:20,4,5,"double")
    x
    print(y)

cholesky decomposition

Description

Performs the Cholesky factorization of a positive definite MPCR matrix x.

Usage

## S4 method for signature 'Rcpp_MPCR'
chol(x,upper_triangle=TRUE)

Arguments

x

An MPCR matrix.

upper_triangle

Boolean to check on which triangle the cholesky decomposition should be applied.

Value

An MPCR matrix.

Examples

  ## Not run: 
  library(MPCR)
  x <- as.MPCR(c(1.21, 0.18, 0.13, 0.41, 0.06, 0.23,
                 0.18, 0.64, 0.10, -0.16, 0.23, 0.07,
                 0.13, 0.10, 0.36, -0.10, 0.03, 0.18,
                 0.41, -0.16, -0.10, 1.05, -0.29, -0.08,
                 0.06, 0.23, 0.03, -0.29, 1.71, -0.10,
                 0.23, 0.07, 0.18, -0.08, -0.10, 0.36),6,6,precision="double")
  chol_out <- chol(x)
  
## End(Not run)

cholesky inverse

Description

Performs the inverse of the original matrix using the Cholesky factorization of an MPCR matrix x.

Usage

## S4 method for signature 'Rcpp_MPCR'
chol2inv(x, size = NCOL(x))

Arguments

x

An MPCR object.

size

The number of columns to use.

Value

An MPCR object.

Examples

  ## Not run: 
  library(MPCR)
  x <- as.MPCR(c(1.21, 0.18, 0.13, 0.41, 0.06, 0.23,
                 0.18, 0.64, 0.10, -0.16, 0.23, 0.07,
                 0.13, 0.10, 0.36, -0.10, 0.03, 0.18,
                 0.41, -0.16, -0.10, 1.05, -0.29, -0.08,
                 0.06, 0.23, 0.03, -0.29, 1.71, -0.10,
                 0.23, 0.07, 0.18, -0.08, -0.10, 0.36),6,6,precision="single")
  chol_out <- chol(x)
  chol <- chol2inv(chol_out)

## End(Not run)

crossprod

Description

Calculates the cross product of two MPCR matrices. It uses BLAS routine gemm() for A X B operations and syrk() for A X A^T operations.

Usage

## S4 method for signature 'Rcpp_MPCR'
crossprod(x, y = NULL)

## S4 method for signature 'Rcpp_MPCR'
tcrossprod(x, y = NULL)

Arguments

x

An MPCR object.

y

Either NULL, or an MPCR matrix.

Details

Calculates cross product of two MPCR matrices performs:
x %*% y , t(x) %*% x
This function uses blas routine gemm() for A X B operations & syrk() for A X A^T operations.

Value

An MPCR matrix.

Examples

    ## Not run: 
    library(MPCR)
    x <- as.MPCR(1:16,4,4,"single")
    y <- as.MPCR(1:20,4,5,"double")

    z <- crossprod(x)     # t(x) x
    z <- tcrossprod(x)    # x t(x)
    z <- crossprod(x,y)   # x y
    z <- x %*% y          # x y
    
## End(Not run)

eigen decomposition

Description

Solves a system of equations or invert an MPCR matrix, using lapack routine syevr()

Usage

## S4 method for signature 'Rcpp_MPCR'
eigen(x, only.values = FALSE)

Arguments

x

An MPCR object.

only.values

(TRUE/FALSE)?

Value

A list contains MPCR objects describing the values and optionally vectors.

Examples

## Not run: 
library(MPCR)
s <- runif(10, 3)
cross_prod <- crossprod(s)
x <- as.MPCR(cross_prod,nrow(cross_prod),nrow(cross_prod),"double")
y <- eigen(x)

## End(Not run)

isSymmetric

Description

Check if a given MPCR matrix is symmetric.

Usage

## S4 method for signature 'Rcpp_MPCR'
isSymmetric(object, ...)

Arguments

object

An MPCR matrix.

...

Ignored.

Value

A logical value.

Examples

## Not run: 
library(MPCR)

x <- as.MPCR(1:50,25,2,"Single")
isSymmetric(x)                      #false

crossprod_output<-crossprod(x)
isSymmetric(crossprod_output)       #true

## End(Not run)

norm

Description

Compute norm.

Usage

## S4 method for signature 'Rcpp_MPCR'
norm(x, type = "O")

Arguments

x

An MPCR object.

type

"O"-ne, "I"-nfinity, "F"-robenius, "M"-ax modulus, and "1" norms.

Value

An MPCR object.

Examples

library(MPCR)

x <- as.MPCR(1:20,precision="double")
norm(x, type="O")

QR decomposition

Description

QR factorization and related functions.

Usage

## S4 method for signature 'Rcpp_MPCR'
qr(x, tol = 1e-07)

## S4 method for signature 'ANY'
qr.Q(qr, complete = FALSE, Dvec)

## S4 method for signature 'ANY'
qr.R(qr, complete = FALSE)

Arguments

x

An MPCR matrix.

qr

QR decomposition MPCR object.

tol

The tolerance for determining numerical column rank.

complete

Should the complete or truncated factor be returned?

Dvec

Vector of diagonals to use when re-constructing Q (default is 1's).

Details

The factorization is performed by the LAPACK routine geqp3(). This should be similar to calling qr() on an ordinary R matrix with the argument LAPACK=TRUE.

Value

qr

Output of qr().

Examples

## Not run: 

library(MPCR)


qr_input <-as.MPCR( c(1, 2, 3, 2, 4, 6, 3, 3, 3),3,3,"single")
qr_out <- qr(qr_input)
qr_out
qr_out[["qr"]]$PrintValues()
qr_out[["qraux"]]$PrintValues()
qr_out[["pivot"]]$PrintValues()
qr_out[["rank"]]$PrintValues()

qr_q <- qr.Q(qr_out)
qr_q

## End(Not run)

reciprocal condition

Description

Compute matrix norm.

Usage

## S4 method for signature 'Rcpp_MPCR'
rcond(x, norm = "O", useInv = FALSE)

Arguments

x

An MPCR object.

norm

"O"-ne or "I"-nfinity norm.

useInv

TRUE to use the lower triangle only.

Value

An MPCR Object.

Examples


library(MPCR)

x <- as.MPCR(1:20,precision="double")
rcond(x)

solve

Description

Solve a system of equations or invert an MPCR matrix.

Usage

## S4 method for signature 'Rcpp_MPCR'
solve(a, b = NULL, ...)

Arguments

a, b

An MPCR objects.

...

Ignored.

Value

Solves the equation AX=B .and if B=NULL t(A) will be used.

Examples

## Not run: 
library(MPCR)

x <- as.MPCR(1:20,4,5,"double")
y <- crossprod(x)
solve(y)

## End(Not run)

SVD

Description

SVD factorization.

Usage

## S4 method for signature 'Rcpp_MPCR'
La.svd(x, nu = min(n, p), nv = min(n, p))

## S4 method for signature 'Rcpp_MPCR'
svd(x, nu = min(n, p), nv = min(n, p))

Arguments

x

An MPCR matrix.

nu, nv

The number of left/right singular vectors to return.

Details

The factorization is performed by the LAPACK routine gesdd().

Value

The SVD decomposition of the MPCR matrix.

Examples

## Not run: 
library(MPCR)
svd_vals <- c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0,
              0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0,
              0, 0, 0, 1, 1, 1)

x <- as.MPCR(svd_vals,9,4,"single")
y <- svd(x)

## End(Not run)

Back/Forward solve

Description

Solves a system of linear equations where the coefficient matrix is upper or lower triangular. The function solves the equation A X = B, where A is the coefficient matrix, X is the solution vector, and B is the right-hand side vector.

Usage

 ## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
backsolve(r, x, k = ncol(r), upper.tri = TRUE, transpose = FALSE)

 ## S4 method for signature 'Rcpp_MPCR,Rcpp_MPCR'
forwardsolve(l, x, k = ncol(l), upper.tri = FALSE, transpose = FALSE)

Arguments

l

An MPCR object.

r

An MPCR object.

x

An MPCR object whose columns give the right-hand sides for the equations.

k

The number of columns of r and rows of x to use.

upper.tri

logical; if TRUE, the upper triangular part of r is used. Otherwise, the lower one.

transpose

logical; if TRUE, solve for t( l , r ) %*% output == x.

Value

An MPCR object represents the solution to the system of linear equations.

Examples

## Not run: 
  library(MPCR)
  a <- matrix(c(2, 0, 0, 3), nrow = 2)
  b <- matrix(c(1, 2), nrow = 2)
  a_MPCR <- as.MPCR(a,2,2,"single")
  b_MPCR <- as.MPCR(b,2,1,"double")
  x <- forwardsolve(a_MPCR, b_MPCR)
  x

## End(Not run)

MPCR GEMM (Matrix-Matrix Multiplication)

Description

Performs matrix-matrix multiplication of two given MPCR matrices to performs:
C = alpha A * B + beta C
C = alpha A A^T + beta C

Usage


## S4 method for signature 'Rcpp_MPCR'
MPCR.gemm(a,b = NULL,c,transpose_a= FALSE,transpose_b=FALSE,alpha=1,beta=0)

Arguments

a

An MPCR matrix A.

b

An MPCR matrix B, if NULL, the function will perform syrk operation from blas.

c

Input/Output MPCR matrix C.

transpose_a

A flag to indicate whether transpose matrix A should be used, if B is NULL and transpose_a =TRUE
The function will perform the following operation:
C=alphaA^TXA+betaC.

transpose_b

A flag to indicate whether transpose matrix B should be used.

alpha

Specifies the scalar alpha.

beta

Specifies the scalar beta.

Value

An MPCR matrix.

Examples

   ## Not run: 
   library(MPCR)
   # create 3 MPCR matrices a,b,c
   a <- as.MPCR(1:12, 3, 4, "double")
   b <- as.MPCR(1:20, 4, 5, "double")
   c <- as.MPCR(matrix(0, 3, 5), 3, 5, "double")
   print(c)
   MPCR.gemm(a,b,c,transpose_a=FALSE,transpose_b=TRUE,alpha=1,beta=1)
   print(c)
   
## End(Not run)

MPCR TRSM (Triangular Solve)

Description

Solves a triangular matrix equation.
performs:
op(A)*X=alpha*B
X*op(A)=alpha*B

Usage


## S4 method for signature 'Rcpp_MPCR'
MPCR.trsm(a,b,upper_triangle,transpose,side = 'L',alpha =1)

Arguments

a

MPCR Matrix A.

b

MPCR Matrix B.

upper_triangle

If the value is TRUE, the referenced part of matrix A corresponds to the upper triangle, with the opposite triangle assumed to contain zeros.

transpose

If TRUE, the transpose of A is used.

side

'R for Right side, 'L' for Left side.

alpha

Factor used for A, If alpha is zero, A is not accessed.

Value

An MPCR Matrix.

Examples

   ## Not run: 
   library(MPCR)
   a <- matrix(c(3.12393, -1.16854, -0.304408, -2.15901,
                 -1.16854, 1.86968, 1.04094, 1.35925,
                 -0.304408, 1.04094, 4.43374, 1.21072,
                 -2.15901, 1.35925, 1.21072, 5.57265), 4,4)


   mat_a <- as.MPCR(a,4,4,"single")
   mat_b <- as.MPCR(a,4,4,"double")

   MPCR.trsm(a=mat_a,b=mat_b,side='R',upper_triangle=TRUE,transpose=FALSE,alpha=1)
   print(mat_b)
   
## End(Not run)

Context Handling

Description

The Context Placement handling mechanism used by MPCR manages the dispatch of operations to either the CPU or GPU. When the operation placement is set to "GPU," all operations with GPU implementations (e.g., Linear Algebra functions) will execute on the GPU. If an operation is not supported on the GPU, it will execute on the CPU without altering the Context Placement for subsequent code execution.

By default, the placement is set to the CPU at the start of execution. To switch to GPU placement, users must use specific functions to set and get the current placement.

The Context Placement mechanism does not control initial memory allocation. For optimal performance, users should be aware of where data is allocated. If the user is uncertain about data allocation, the package will automatically manage allocation and data movement between the CPU and GPU. It employs an internal caching mechanism to minimize memory transfers, although data might be allocated on both the CPU and GPU during the object's lifetime. Helper functions are available to check if memory is allocated on the CPU/GPU and to free memory on either one.

Value

Operation Context (Setting and Getting).

Set Operation Context

MPCR.SetOperationPlacement Set the placement for the up-coming flow of code ( "CPU", "GPU"). MPCR.SetOperationPlacement(placement)

placement: String to indicate on which hardware should the up-coming flow of code be executed.

Get Operation Context

MPCR.GetOperationPlacement Get the placement used currently for dispatching operations ( "CPU", "GPU"). MPCR.GetOperationPlacement()

returns the placement currently used for operations.

Examples

## Not run: 
  library(MPCR)
  values <- c(3.12393, -1.16854, -0.304408, -2.15901,
              -1.16854, 1.86968, 1.04094, 1.35925, -0.304408,
              1.04094, 4.43374, 1.21072, -2.15901, 1.35925, 1.21072, 5.57265)


  x <- new(MPCR, 16, "float","GPU") # Data will be allocated on GPU
  y <- new(MPCR, 16, "float","GPU")

  # Since this two for loops changes the data inside the MPCR objects,
  # the GPU memory will be freed.

  for (val in 1:16) {
    x[[val]] <- values[[val]]
    y[[val]] <- values[[val]]

  }


  # At this point only CPU memory is allocated.

  x$ToMatrix(4, 4)
  y$ToMatrix(4, 4)
  paste("X and Y values")
  x$PrintValues()  # CPU Function
  y$PrintValues()  # CPU Function

  MPCR.SetOperationPlacement("default", "GPU") # Set Function Dispatching to GPU
  cat("----------------------- CrossProduct C=XY --------------------\n")

  # GPU Cuda Kernel, The data will be automatically copied to GPU.
  crossproduct <- crossprod(x, y)

   # Data won't be moved to CPU, since crossprod didn't change the content.
  x$PrintValues()  # CPU Function
  y$PrintValues()  # CPU Function

  # CPU Function, so the data will be copied to CPU after finalizing the GPU Call.
  crossproduct$PrintValues()

## End(Not run)