Plus matrices in a list

# RAM超過6GB，才可以run!!
# 否則請更改矩陣大小(mat_size)或是列表的長度(length_list)
length_list <- 30
mat_size <- c(3000, 3000)

mat <- matrix(rnorm(prod(mat_size)), mat_size[1], mat_size[2])
mat.list <- lapply(1:length_list, function(i) mat)

res1 <- function(){
  mat_sum <- matrix(rep(0, prod(mat_size)), mat_size[1], mat_size[2])
  for (i in 1:length_list){
    mat_sum <- mat_sum + mat.list[[i]]
  }
  mat_sum
}
res2 <- function() Reduce('+', mat.list)
res3 <- function() matrix(rowSums(matrix(unlist(mat.list), nrow=prod(mat_size))), nrow=mat_size[1])
res4 <- function() matrix(colSums(do.call(rbind, lapply(mat.list, as.vector))), nrow=mat_size[1])
res5 <- function() matrix(rowSums(do.call(cbind, lapply(mat.list, as.vector))), nrow=mat_size[1])
library(Rcpp)
library(RcppArmadillo)
sourceCpp(code = '
// [[Rcpp::depends(RcppArmadillo)]]
#include <RcppArmadillo.h>
using namespace Rcpp;
using namespace arma;

// [[Rcpp::export]]
mat list_mat_sum_f(List data_list){
 int n = data_list.size();
 SEXP mat1 = data_list[0];
 NumericMatrix mat1_rmat(mat1);
 mat result_mat(mat1_rmat.begin(), mat1_rmat.nrow(), mat1_rmat.ncol());
 for(int i = 1; i < n; i++) {
   SEXP tmp_m = data_list[i];
   NumericMatrix data_m(tmp_m);
   mat working_m(data_m.begin(), data_m.nrow(), data_m.ncol(), false);
   result_mat += working_m;
 }
 return result_mat;
}')
res6 <- function() list_mat_sum_f(mat.list)

# cmpfun
library(compiler)
res1_cmp <- cmpfun(res1)
res2_cmp <- cmpfun(res2)
res3_cmp <- cmpfun(res3)
res4_cmp <- cmpfun(res4)
res5_cmp <- cmpfun(res5)

all.equal(tmp <- res1(), res2())
# TRUE
all.equal(tmp, res3())
# TRUE
all.equal(tmp, res4())
# TRUE
all.equal(tmp, res5())
# TRUE
all.equal(tmp, res6())
# TRUE

library(microbenchmark)
microbenchmark(res1(),res2(),res3(),res4(),res5(),res6(),
          res1_cmp(),res2_cmp(),res3_cmp(),res4_cmp(),res5_cmp(),
          times = 20)

# Unit: milliseconds
#       expr       min        lq      mean    median        uq       max neval
#     res1()  446.3391  597.8725  783.2508  862.2031  964.0379 1033.5585    20
#     res2()  402.5949  466.2898  570.2002  508.7840  611.2877  876.8133    20
#     res3() 2134.7643 2225.5092 2305.0105 2310.8695 2373.2789 2454.5463    20
#     res4() 2434.3700 2558.5332 2760.0291 2825.5692 2926.9820 3104.8740    20
#     res5() 1641.4066 1822.6534 2022.1378 2087.0276 2168.8398 2392.9408    20
#     res6()  265.8707  281.7982  294.2491  286.9239  296.3758  384.0111    20
# res1_cmp()  442.7341  651.8927  754.0913  770.4271  861.9162  973.8377    20
# res2_cmp()  395.7517  457.7643  528.6478  472.3783  527.4379  873.7846    20
# res3_cmp() 1932.8973 2156.7042 2244.6550 2274.8265 2341.0351 2469.4029    20
# res4_cmp() 2442.3234 2748.4403 2836.7386 2874.3035 2963.9890 3117.3893    20
# res5_cmp() 1655.1716 1976.7308 2061.0744 2086.4023 2194.5759 2248.9206    20