matmult {float} | R Documentation |
Matrix multiplication for numeric/float matrices.
## S4 method for signature 'float32,float32'
x %*% y
## S4 method for signature 'float32,matrix'
x %*% y
## S4 method for signature 'matrix,float32'
x %*% y
x, y |
Numeric/float matrices. |
If a numeric matrix is multiplied against a float matrix, then if the "numeric" matrix is integers, the integers are promoted to floats. Otherwise, the float matrix is promoted to doubles.
A matrix of the same type as the highest precision input.
library(float)
s1 = flrunif(5, 5)
s2 = flrunif(5, 2)
x = matrix(1:25, 5)
s1 %*% s2 # float
storage.mode(x) # integer
x %*% s2 # float
storage.mode(x) = "double"
x %*% s2 # double