[R] How to build a large identity matrix faster?

[Uwe Ligges]

I quickly looked at it, and the difference comes from:

n <- 5e3
system.time(x <- array(0, c(n, n))) # from diag()
system.time(x <- matrix(0, n, n))   # from Rdiag()

Replaced in R-devel.

Best,
Uwe Ligges



On 07.06.2012 12:11, Spencer Graves wrote:
> On 6/7/2012 2:27 AM, Rui Barradas wrote:
>> Hello,
>>
>> To my great surprise, on my system, Windows 7, R 15.0, 32 bits, an R
>> version is faster!
>
> I was also surprised, Windows 7, R 2.15.0, 64-bit
>
>
>  > rbind(diag=t1, Rdiag=t2, ratio=t1/t2)
> user.self sys.self elapsed user.child sys.child
> diag 0.72 0.080000 0.81 NA NA
> Rdiag 0.09 0.030000 0.12 NA NA
> ratio 8.00 2.666667 6.75 NA NA
>  >
>  > sessionInfo()
> R version 2.15.0 (2012-03-30)
> Platform: x86_64-pc-mingw32/x64 (64-bit)
>
> locale:
> [1] LC_COLLATE=English_United States.1252
> [2] LC_CTYPE=English_United States.1252
> [3] LC_MONETARY=English_United States.1252
> [4] LC_NUMERIC=C
> [5] LC_TIME=English_United States.1252
>
> attached base packages:
> [1] splines stats graphics grDevices utils datasets methods
> [8] base
>
> other attached packages:
> [1] fda_2.2.9 Matrix_1.0-6 lattice_0.20-6 zoo_1.7-7
>
> loaded via a namespace (and not attached):
> [1] grid_2.15.0 tools_2.15.0
>  >
>
>
> Spencer
>>
>>
>> Rdiag <- function(n){
>> m <- matrix(0, nrow=n, ncol=n)
>> m[matrix(rep(seq_len(n), 2), ncol=2)] <- 1
>> m
>> }
>>
>> Rdiag(4)
>>
>> n <- 5e3
>> t1 <- system.time(d1 <- diag(n))
>> t2 <- system.time(d2 <- Rdiag(n))
>> all.equal(d1, d2)
>> rbind(diag=t1, Rdiag=t2, ratio=t1/t2)
>>
>>
>> Anyway, why don't you create it once, save a copy and use it many times?
>>
>> Hope this helps,
>>
>> Rui Barradas
>>
>> Em 07-06-2012 08:55, Ceci Tam escreveu:
>>> Hello, I am trying to build a large size identity matrix using
>>> diag(). The
>>> size is around 23000 and I've tried diag(23000), that took a long time.
>>> Since I have to use this operation several times in my program, the
>>> running
>>> time is too long to be tolerable. Are there any alternative for diag(N)?
>>> Thanks
>>>
>>> Cheers,
>>> yct
>>>
>>> [[alternative HTML version deleted]]
>>>
>>> ______________________________________________
>>> R-help at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-help
>>> PLEASE do read the posting guide
>>> http://www.R-project.org/posting-guide.html
>>> and provide commented, minimal, self-contained, reproducible code.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.