SNOW.R

# Mon Aug 15 09:35:32 KST 2016
# 
# Is there a common abstraction for using apply with the parallel backend?
# Yes. So lets use it.

library(parallel)

# Don't think that I want to use this.
Sys.setenv(MC_CORES = 3L)

options("mc.cores")

x1 = list(1:10, letters, rnorm(5))
x2 = list(101:110, LETTERS, rnorm(5))

expensive_function <- function(...){
system.time({
    bigmatrix <- matrix(rnorm(4e6L), nrow=2e3L)
    solve(bigmatrix)
})
}

cl_PSOCK <- makeCluster(3L, "PSOCK")

clusterMap(cl_PSOCK, expensive_function, x1, x2)

cl_FORK <- makeCluster(3L, "FORK")

clusterMap(cl_FORK, expensive_function, x1, x2)

print("hello \
there")


{
a = 1 + "one"
a = 1 + 1
}

{
a = try(1 + "one")
#a = 1 + 1
}