colRowMads_subset - HenrikBengtsson/matrixStats GitHub Wiki

matrixStats: Benchmark report

colMads() and rowMads() benchmarks on subsetted computation

This report benchmark the performance of colMads() and rowMads() on subsetted computation.

Data type "integer"

Data

> rmatrix <- function(nrow, ncol, mode = c("logical", "double", "integer", "index"), range = c(-100, 
+     +100), na_prob = 0) {
+     mode <- match.arg(mode)
+     n <- nrow * ncol
+     if (mode == "logical") {
+         x <- sample(c(FALSE, TRUE), size = n, replace = TRUE)
+     }     else if (mode == "index") {
+         x <- seq_len(n)
+         mode <- "integer"
+     }     else {
+         x <- runif(n, min = range[1], max = range[2])
+     }
+     storage.mode(x) <- mode
+     if (na_prob > 0) 
+         x[sample(n, size = na_prob * n)] <- NA
+     dim(x) <- c(nrow, ncol)
+     x
+ }
> rmatrices <- function(scale = 10, seed = 1, ...) {
+     set.seed(seed)
+     data <- list()
+     data[[1]] <- rmatrix(nrow = scale * 1, ncol = scale * 1, ...)
+     data[[2]] <- rmatrix(nrow = scale * 10, ncol = scale * 10, ...)
+     data[[3]] <- rmatrix(nrow = scale * 100, ncol = scale * 1, ...)
+     data[[4]] <- t(data[[3]])
+     data[[5]] <- rmatrix(nrow = scale * 10, ncol = scale * 100, ...)
+     data[[6]] <- t(data[[5]])
+     names(data) <- sapply(data, FUN = function(x) paste(dim(x), collapse = "x"))
+     data
+ }
> data <- rmatrices(mode = mode)

Results

10x10 integer matrix

> X <- data[["10x10"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3127712 167.1    5709258 305.0  5709258 305.0
Vcells 5944525  45.4   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3126871 167.0    5709258 305.0  5709258 305.0
Vcells 5942293  45.4   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+10x10 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 0.001618 0.001650 0.0030489 0.0016855 0.0017835 0.133118
2 colMads(X, rows, cols) 0.001788 0.001845 0.0019660 0.0018920 0.0019995 0.004344
3 colMads(X[rows, cols]) 0.002188 0.002389 0.0025937 0.0025155 0.0026230 0.006317
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.0000000 1.000000 1.000000 1.0000000
2 colMads(X, rows, cols) 1.105068 1.118182 0.6448183 1.122516 1.121110 0.0326327
3 colMads(X[rows, cols]) 1.352287 1.447879 0.8507063 1.492436 1.470704 0.0474541

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on integer+10x10 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 0.001594 0.0016560 0.0017765 0.0016845 0.0018295 0.004182
2 rowMads(X, cols, rows) 0.001782 0.0018410 0.0034641 0.0019110 0.0020485 0.152137
3 rowMads(X[cols, rows]) 0.002171 0.0023605 0.0026019 0.0024825 0.0026305 0.007368
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 rowMads(X, cols, rows) 1.117942 1.111715 1.949963 1.134461 1.119705 36.379005
3 rowMads(X[cols, rows]) 1.361982 1.425423 1.464599 1.473731 1.437825 1.761836

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+10x10 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+10x10 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
2 rowMads_X_S 1.594 1.656 1.77650 1.6845 1.8295 4.182
1 colMads_X_S 1.618 1.650 3.04889 1.6855 1.7835 133.118
expr min lq mean median uq max
2 rowMads_X_S 1.000000 1.0000000 1.000000 1.000000 1.0000000 1.00000
1 colMads_X_S 1.015057 0.9963768 1.716234 1.000594 0.9748565 31.83118

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+10x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x100 integer matrix

> X <- data[["100x100"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3125581 167.0    5709258 305.0  5709258 305.0
Vcells 5610824  42.9   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3125575 167.0    5709258 305.0  5709258 305.0
Vcells 5615907  42.9   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+100x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 0.124557 0.1253345 0.1259331 0.125692 0.1261295 0.130745
2 colMads(X, rows, cols) 0.125732 0.1274145 0.1280266 0.127867 0.1283200 0.136184
3 colMads(X[rows, cols]) 0.133008 0.1335405 0.1346972 0.134001 0.1345355 0.185799
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 colMads(X, rows, cols) 1.009433 1.016596 1.016624 1.017304 1.017367 1.041600
3 colMads(X[rows, cols]) 1.067848 1.065473 1.069593 1.066106 1.066646 1.421079

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on integer+100x100 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 0.124224 0.1251330 0.1258091 0.1256310 0.1260185 0.133772
2 rowMads(X, cols, rows) 0.126334 0.1271960 0.1278689 0.1277505 0.1283000 0.132051
3 rowMads(X[cols, rows]) 0.132569 0.1335745 0.1347110 0.1338725 0.1342685 0.181574
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.0000000
2 rowMads(X, cols, rows) 1.016985 1.016487 1.016372 1.016871 1.018104 0.9871348
3 rowMads(X[cols, rows]) 1.067177 1.067460 1.070757 1.065601 1.065467 1.3573394

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+100x100 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+100x100 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
2 rowMads_X_S 124.224 125.1330 125.8091 125.631 126.0185 133.772
1 colMads_X_S 124.557 125.3345 125.9331 125.692 126.1295 130.745
expr min lq mean median uq max
2 rowMads_X_S 1.000000 1.00000 1.000000 1.000000 1.000000 1.0000000
1 colMads_X_S 1.002681 1.00161 1.000986 1.000485 1.000881 0.9773719

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+100x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x10 integer matrix

> X <- data[["1000x10"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3126332 167.0    5709258 305.0  5709258 305.0
Vcells 5614885  42.9   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3126323 167.0    5709258 305.0  5709258 305.0
Vcells 5619963  42.9   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+1000x10 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 0.081427 0.0826875 0.0832570 0.0830020 0.0833390 0.100612
2 colMads(X, rows, cols) 0.083962 0.0847270 0.0854519 0.0850420 0.0853190 0.126803
3 colMads(X[rows, cols]) 0.089906 0.0907615 0.0916832 0.0911845 0.0917125 0.106254
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 colMads(X, rows, cols) 1.031132 1.024665 1.026362 1.024578 1.023758 1.260317
3 colMads(X[rows, cols]) 1.104130 1.097645 1.101206 1.098582 1.100475 1.056077

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on integer+1000x10 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 0.081849 0.0825785 0.0836888 0.0829080 0.0833035 0.118436
2 rowMads(X, cols, rows) 0.084582 0.0850780 0.0855695 0.0852765 0.0857000 0.096658
3 rowMads(X[cols, rows]) 0.091102 0.0920475 0.0925392 0.0924060 0.0925715 0.102606
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.0000000
2 rowMads(X, cols, rows) 1.033391 1.030268 1.022472 1.028568 1.028768 0.8161201
3 rowMads(X[cols, rows]) 1.113050 1.114667 1.105754 1.114561 1.111256 0.8663413

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+1000x10 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+1000x10 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
2 rowMads_X_S 81.849 82.5785 83.68883 82.908 83.3035 118.436
1 colMads_X_S 81.427 82.6875 83.25705 83.002 83.3390 100.612
expr min lq mean median uq max
2 rowMads_X_S 1.0000000 1.00000 1.0000000 1.000000 1.000000 1.0000000
1 colMads_X_S 0.9948442 1.00132 0.9948406 1.001134 1.000426 0.8495052

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+1000x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

10x1000 integer matrix

> X <- data[["10x1000"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3126535 167.0    5709258 305.0  5709258 305.0
Vcells 5615703  42.9   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3126529 167.0    5709258 305.0  5709258 305.0
Vcells 5620786  42.9   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+10x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 0.074979 0.076251 0.0785998 0.0776810 0.0789350 0.144374
2 colMads(X, rows, cols) 0.079670 0.081699 0.0835705 0.0828720 0.0837255 0.132708
3 colMads(X[rows, cols]) 0.083884 0.085756 0.0876781 0.0871725 0.0884730 0.119358
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.00000 1.000000 1.000000 1.0000000
2 colMads(X, rows, cols) 1.062564 1.071448 1.06324 1.066825 1.060689 0.9191960
3 colMads(X[rows, cols]) 1.118767 1.124654 1.11550 1.122186 1.120834 0.8267278

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on integer+10x1000 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 0.075136 0.0763025 0.0781011 0.0775200 0.0785300 0.121054
2 rowMads(X, cols, rows) 0.078360 0.0810635 0.0822967 0.0825625 0.0833075 0.087298
3 rowMads(X[cols, rows]) 0.082860 0.0845020 0.0859525 0.0856260 0.0866880 0.102774
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.0000000
2 rowMads(X, cols, rows) 1.042909 1.062396 1.053720 1.065048 1.060837 0.7211492
3 rowMads(X[cols, rows]) 1.102800 1.107460 1.100528 1.104567 1.103884 0.8489930

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+10x1000 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+10x1000 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
2 rowMads_X_S 75.136 76.3025 78.10114 77.520 78.530 121.054
1 colMads_X_S 74.979 76.2510 78.59984 77.681 78.935 144.374
expr min lq mean median uq max
2 rowMads_X_S 1.0000000 1.0000000 1.000000 1.000000 1.000000 1.000000
1 colMads_X_S 0.9979105 0.9993251 1.006385 1.002077 1.005157 1.192641

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+10x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x1000 integer matrix

> X <- data[["100x1000"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3126734 167.0    5709258 305.0  5709258 305.0
Vcells 5638357  43.1   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3126728 167.0    5709258 305.0  5709258 305.0
Vcells 5688440  43.4   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+100x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 1.266993 1.270186 1.294061 1.271557 1.274584 1.750929
2 colMads(X, rows, cols) 1.284522 1.287359 1.331201 1.289322 1.298381 1.795537
3 colMads(X[rows, cols]) 1.340125 1.344108 1.382902 1.346618 1.359126 2.112093
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 colMads(X, rows, cols) 1.013835 1.013520 1.028701 1.013971 1.018670 1.025477
3 colMads(X[rows, cols]) 1.057721 1.058198 1.068653 1.059030 1.066329 1.206270

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on integer+100x1000 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 1.269649 1.271853 1.276187 1.273303 1.275263 1.361940
2 rowMads(X, cols, rows) 1.300770 1.302888 1.306933 1.304312 1.307311 1.372135
3 rowMads(X[cols, rows]) 1.334043 1.336432 1.345289 1.338763 1.341876 1.669967
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 rowMads(X, cols, rows) 1.024512 1.024401 1.024092 1.024353 1.025130 1.007486
3 rowMads(X[cols, rows]) 1.050718 1.050776 1.054147 1.051410 1.052235 1.226168

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+100x1000 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+100x1000 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 1.266993 1.270186 1.294061 1.271557 1.274584 1.750929
2 rowMads_X_S 1.269649 1.271853 1.276187 1.273303 1.275263 1.361940
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.0000000 1.000000 1.000000 1.0000000
2 rowMads_X_S 1.002096 1.001313 0.9861873 1.001373 1.000533 0.7778385

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+100x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x100 integer matrix

> X <- data[["1000x100"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3126950 167.0    5709258 305.0  5709258 305.0
Vcells 5639132  43.1   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3126944 167.0    5709258 305.0  5709258 305.0
Vcells 5689215  43.5   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+1000x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 0.946344 0.9490525 0.9556377 0.9504085 0.952523 1.170666
2 colMads(X, rows, cols) 0.958893 0.9625085 0.9665988 0.9634775 0.965221 1.054095
3 colMads(X[rows, cols]) 1.009027 1.0123290 1.0228459 1.0142980 1.016363 1.735138
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.0000000
2 colMads(X, rows, cols) 1.013260 1.014178 1.011470 1.013751 1.013331 0.9004233
3 colMads(X[rows, cols]) 1.066237 1.066673 1.070328 1.067223 1.067022 1.4821802

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on integer+1000x100 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 0.954634 0.9583080 0.9617993 0.9604665 0.9628435 0.990129
2 rowMads(X, cols, rows) 0.980087 0.9830515 0.9857112 0.9844675 0.9868645 1.043447
3 rowMads(X[cols, rows]) 1.028572 1.0316400 1.0438749 1.0347785 1.0389705 1.586369
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 rowMads(X, cols, rows) 1.026663 1.025820 1.024862 1.024989 1.024948 1.053850
3 rowMads(X[cols, rows]) 1.077452 1.076522 1.085336 1.077371 1.079065 1.602184

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on integer+1000x100 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+1000x100 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 946.344 949.0525 955.6377 950.4085 952.5230 1170.666
2 rowMads_X_S 954.634 958.3080 961.7993 960.4665 962.8435 990.129
expr min lq mean median uq max
1 colMads_X_S 1.00000 1.000000 1.000000 1.000000 1.000000 1.0000000
2 rowMads_X_S 1.00876 1.009752 1.006448 1.010583 1.010835 0.8457827

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on integer+1000x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

Data type "double"

Data

> rmatrix <- function(nrow, ncol, mode = c("logical", "double", "integer", "index"), range = c(-100, 
+     +100), na_prob = 0) {
+     mode <- match.arg(mode)
+     n <- nrow * ncol
+     if (mode == "logical") {
+         x <- sample(c(FALSE, TRUE), size = n, replace = TRUE)
+     }     else if (mode == "index") {
+         x <- seq_len(n)
+         mode <- "integer"
+     }     else {
+         x <- runif(n, min = range[1], max = range[2])
+     }
+     storage.mode(x) <- mode
+     if (na_prob > 0) 
+         x[sample(n, size = na_prob * n)] <- NA
+     dim(x) <- c(nrow, ncol)
+     x
+ }
> rmatrices <- function(scale = 10, seed = 1, ...) {
+     set.seed(seed)
+     data <- list()
+     data[[1]] <- rmatrix(nrow = scale * 1, ncol = scale * 1, ...)
+     data[[2]] <- rmatrix(nrow = scale * 10, ncol = scale * 10, ...)
+     data[[3]] <- rmatrix(nrow = scale * 100, ncol = scale * 1, ...)
+     data[[4]] <- t(data[[3]])
+     data[[5]] <- rmatrix(nrow = scale * 10, ncol = scale * 100, ...)
+     data[[6]] <- t(data[[5]])
+     names(data) <- sapply(data, FUN = function(x) paste(dim(x), collapse = "x"))
+     data
+ }
> data <- rmatrices(mode = mode)

Results

10x10 double matrix

> X <- data[["10x10"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3127168 167.1    5709258 305.0  5709258 305.0
Vcells 5730221  43.8   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3127153 167.1    5709258 305.0  5709258 305.0
Vcells 5730389  43.8   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+10x10 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 0.002072 0.0021300 0.0024325 0.00220 0.0022990 0.020894
2 colMads(X, rows, cols) 0.002278 0.0023440 0.0024972 0.00242 0.0025555 0.005687
3 colMads(X[rows, cols]) 0.002730 0.0029925 0.0031582 0.00307 0.0031870 0.007702
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.0000000
2 colMads(X, rows, cols) 1.099421 1.100469 1.026619 1.100000 1.111570 0.2721834
3 colMads(X[rows, cols]) 1.317568 1.404930 1.298357 1.395455 1.386255 0.3686226

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on double+10x10 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 0.002087 0.0021295 0.0022628 0.0021765 0.0023210 0.005127
2 rowMads(X, cols, rows) 0.002264 0.0023235 0.0027862 0.0023875 0.0025560 0.034763
3 rowMads(X[cols, rows]) 0.002641 0.0029090 0.0031310 0.0030215 0.0031795 0.009112
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 rowMads(X, cols, rows) 1.084811 1.091101 1.231295 1.096945 1.101249 6.780378
3 rowMads(X[cols, rows]) 1.265453 1.366048 1.383718 1.388238 1.369884 1.777258

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+10x10 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on double+10x10 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
2 rowMads_X_S 2.087 2.1295 2.26278 2.1765 2.321 5.127
1 colMads_X_S 2.072 2.1300 2.43249 2.2000 2.299 20.894
expr min lq mean median uq max
2 rowMads_X_S 1.0000000 1.000000 1.000000 1.000000 1.0000000 1.000000
1 colMads_X_S 0.9928126 1.000235 1.075001 1.010797 0.9905213 4.075288

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on double+10x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x100 double matrix

> X <- data[["100x100"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3127370 167.1    5709258 305.0  5709258 305.0
Vcells 5736165  43.8   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3127364 167.1    5709258 305.0  5709258 305.0
Vcells 5746248  43.9   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+100x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 0.175992 0.1771095 0.1790892 0.1775235 0.1781175 0.208572
2 colMads(X, rows, cols) 0.178243 0.1791985 0.1820410 0.1796175 0.1808025 0.237350
3 colMads(X[rows, cols]) 0.192604 0.1929905 0.1983176 0.1936020 0.1967005 0.279061
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 colMads(X, rows, cols) 1.012790 1.011795 1.016482 1.011796 1.015074 1.137976
3 colMads(X[rows, cols]) 1.094391 1.089668 1.107368 1.090571 1.104330 1.337960

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on double+100x100 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 0.176137 0.1768865 0.1776898 0.177294 0.1777170 0.190034
2 rowMads(X, cols, rows) 0.178403 0.1790985 0.1799678 0.179428 0.1799115 0.194887
3 rowMads(X[cols, rows]) 0.184098 0.1848915 0.1863995 0.185234 0.1856800 0.235587
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 rowMads(X, cols, rows) 1.012865 1.012505 1.012820 1.012037 1.012348 1.025537
3 rowMads(X[cols, rows]) 1.045198 1.045255 1.049017 1.044784 1.044807 1.239710

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+100x100 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on double+100x100 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
2 rowMads_X_S 176.137 176.8865 177.6898 177.2940 177.7170 190.034
1 colMads_X_S 175.992 177.1095 179.0892 177.5235 178.1175 208.572
expr min lq mean median uq max
2 rowMads_X_S 1.0000000 1.000000 1.000000 1.000000 1.000000 1.000000
1 colMads_X_S 0.9991768 1.001261 1.007876 1.001294 1.002254 1.097551

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on double+100x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x10 double matrix

> X <- data[["1000x10"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3127579 167.1    5709258 305.0  5709258 305.0
Vcells 5737570  43.8   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3127573 167.1    5709258 305.0  5709258 305.0
Vcells 5747653  43.9   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+1000x10 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 0.149481 0.151056 0.1538088 0.1515275 0.1556925 0.172223
2 colMads(X, rows, cols) 0.151703 0.153301 0.1569379 0.1566535 0.1574605 0.211940
3 colMads(X[rows, cols]) 0.166276 0.167164 0.1699902 0.1692095 0.1713810 0.216658
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 colMads(X, rows, cols) 1.014865 1.014862 1.020344 1.033829 1.011356 1.230614
3 colMads(X[rows, cols]) 1.112355 1.106636 1.105204 1.116692 1.100766 1.258009

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on double+1000x10 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 0.150658 0.1518865 0.1558518 0.152485 0.1556505 0.268986
2 rowMads(X, cols, rows) 0.152853 0.1537925 0.1555724 0.154502 0.1574575 0.173438
3 rowMads(X[cols, rows]) 0.160257 0.1613410 0.1646386 0.161975 0.1652725 0.265232
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.0000000 1.000000 1.000000 1.0000000
2 rowMads(X, cols, rows) 1.014569 1.012549 0.9982073 1.013227 1.011609 0.6447845
3 rowMads(X[cols, rows]) 1.063714 1.062247 1.0563792 1.062236 1.061818 0.9860439

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+1000x10 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on double+1000x10 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 149.481 151.0560 153.8088 151.5275 155.6925 172.223
2 rowMads_X_S 150.658 151.8865 155.8518 152.4850 155.6505 268.986
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.0000000 1.000000
2 rowMads_X_S 1.007874 1.005498 1.013283 1.006319 0.9997302 1.561847

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on double+1000x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

10x1000 double matrix

> X <- data[["10x1000"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3127785 167.1    5709258 305.0  5709258 305.0
Vcells 5737707  43.8   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3127779 167.1    5709258 305.0  5709258 305.0
Vcells 5747790  43.9   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+10x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 0.134798 0.1359665 0.1382682 0.136553 0.137308 0.294956
2 colMads(X, rows, cols) 0.137468 0.1390495 0.1403244 0.139813 0.140279 0.194123
3 colMads(X[rows, cols]) 0.144962 0.1457125 0.1465781 0.146069 0.146608 0.167936
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.0000000
2 colMads(X, rows, cols) 1.019807 1.022675 1.014871 1.023874 1.021637 0.6581422
3 colMads(X[rows, cols]) 1.075402 1.071679 1.060100 1.069687 1.067731 0.5693595

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on double+10x1000 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 0.134364 0.1358465 0.1368990 0.1363815 0.1369665 0.176040
2 rowMads(X, cols, rows) 0.137478 0.1390235 0.1398765 0.1399440 0.1404265 0.142730
3 rowMads(X[cols, rows]) 0.143316 0.1441095 0.1450114 0.1446385 0.1452310 0.163589
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.0000000
2 rowMads(X, cols, rows) 1.023176 1.023387 1.021749 1.026122 1.025262 0.8107816
3 rowMads(X[cols, rows]) 1.066625 1.060826 1.059259 1.060543 1.060340 0.9292718

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+10x1000 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on double+10x1000 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
2 rowMads_X_S 134.364 135.8465 136.8990 136.3815 136.9665 176.040
1 colMads_X_S 134.798 135.9665 138.2682 136.5530 137.3080 294.956
expr min lq mean median uq max
2 rowMads_X_S 1.00000 1.000000 1.000000 1.000000 1.000000 1.000000
1 colMads_X_S 1.00323 1.000883 1.010001 1.001257 1.002493 1.675506

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on double+10x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x1000 double matrix

> X <- data[["100x1000"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3127984 167.1    5709258 305.0  5709258 305.0
Vcells 5783145  44.2   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3127978 167.1    5709258 305.0  5709258 305.0
Vcells 5883228  44.9   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+100x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 1.787152 1.789640 1.797647 1.792476 1.795505 1.902440
2 colMads(X, rows, cols) 1.801339 1.805511 1.816375 1.808668 1.813373 1.951420
3 colMads(X[rows, cols]) 1.933494 1.939144 1.951791 1.941461 1.945466 2.535541
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 colMads(X, rows, cols) 1.007938 1.008868 1.010418 1.009033 1.009951 1.025746
3 colMads(X[rows, cols]) 1.081886 1.083538 1.085748 1.083117 1.083520 1.332784

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on double+100x1000 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 1.789684 1.792204 1.806411 1.793496 1.795767 2.251923
2 rowMads(X, cols, rows) 1.824163 1.828401 1.832759 1.830094 1.832475 1.921615
3 rowMads(X[cols, rows]) 1.861569 1.873429 1.886224 1.876369 1.882099 2.245709
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.0000000
2 rowMads(X, cols, rows) 1.019265 1.020197 1.014586 1.020406 1.020442 0.8533218
3 rowMads(X[cols, rows]) 1.040166 1.045322 1.044183 1.046207 1.048075 0.9972406

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+100x1000 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on double+100x1000 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 1.787152 1.789640 1.797647 1.792476 1.795505 1.902440
2 rowMads_X_S 1.789684 1.792204 1.806411 1.793496 1.795767 2.251923
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 rowMads_X_S 1.001417 1.001432 1.004875 1.000569 1.000146 1.183703

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on double+100x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x100 double matrix

> X <- data[["1000x100"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3128200 167.1    5709258 305.0  5709258 305.0
Vcells 5783292  44.2   22267496 169.9 56666022 432.4
> colStats <- microbenchmark(colMads_X_S = colMads(X_S, na.rm = FALSE), `colMads(X, rows, cols)` = colMads(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMads(X[rows, cols])` = colMads(X[rows, cols], na.rm = FALSE), 
+     unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3128194 167.1    5709258 305.0  5709258 305.0
Vcells 5883375  44.9   22267496 169.9 56666022 432.4
> rowStats <- microbenchmark(rowMads_X_S = rowMads(X_S, na.rm = FALSE), `rowMads(X, cols, rows)` = rowMads(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMads(X[cols, rows])` = rowMads(X[cols, rows], na.rm = FALSE), 
+     unit = "ms")

Table: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+1000x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 colMads_X_S 1.626832 1.640737 1.741858 1.668980 1.722789 2.576678
2 colMads(X, rows, cols) 1.640423 1.670357 1.824553 1.703631 1.914551 2.833088
3 colMads(X[rows, cols]) 1.693803 1.725707 1.898453 1.774982 1.913741 2.805058
expr min lq mean median uq max
1 colMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
2 colMads(X, rows, cols) 1.008354 1.018053 1.047475 1.020762 1.111309 1.099512
3 colMads(X[rows, cols]) 1.041167 1.051788 1.089901 1.063514 1.110839 1.088634

Table: Benchmarking of rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on double+1000x100 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
1 rowMads_X_S 1.637332 1.642232 1.665525 1.649308 1.677045 1.938316
2 rowMads(X, cols, rows) 1.671836 1.676944 1.701091 1.684074 1.717871 1.993383
3 rowMads(X[cols, rows]) 1.718952 1.728274 1.748624 1.736360 1.758601 2.062640
expr min lq mean median uq max
1 rowMads_X_S 1.000000 1.000000 1.000000 1.000000 1.000000 1.00000
2 rowMads(X, cols, rows) 1.021073 1.021137 1.021355 1.021079 1.024344 1.02841
3 rowMads(X[cols, rows]) 1.049849 1.052393 1.049894 1.052781 1.048631 1.06414

Figure: Benchmarking of colMads_X_S(), colMads(X, rows, cols)() and colMads(X[rows, cols])() on double+1000x100 data as well as rowMads_X_S(), rowMads(X, cols, rows)() and rowMads(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMads_X_S() and rowMads_X_S() on double+1000x100 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr min lq mean median uq max
2 rowMads_X_S 1.637332 1.642232 1.665525 1.649308 1.677045 1.938316
1 colMads_X_S 1.626832 1.640737 1.741858 1.668980 1.722789 2.576678
expr min lq mean median uq max
2 rowMads_X_S 1.0000000 1.0000000 1.000000 1.000000 1.000000 1.000000
1 colMads_X_S 0.9935871 0.9990897 1.045832 1.011927 1.027277 1.329338

Figure: Benchmarking of colMads_X_S() and rowMads_X_S() on double+1000x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

Appendix

Session information

R version 3.6.1 Patched (2019-08-27 r77078)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.3 LTS

Matrix products: default
BLAS:   /home/hb/software/R-devel/R-3-6-branch/lib/R/lib/libRblas.so
LAPACK: /home/hb/software/R-devel/R-3-6-branch/lib/R/lib/libRlapack.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] microbenchmark_1.4-6    matrixStats_0.55.0-9000 ggplot2_3.2.1          
[4] knitr_1.24              R.devices_2.16.0        R.utils_2.9.0          
[7] R.oo_1.22.0             R.methodsS3_1.7.1       history_0.0.0-9002     

loaded via a namespace (and not attached):
 [1] Biobase_2.45.0       bit64_0.9-7          splines_3.6.1       
 [4] network_1.15         assertthat_0.2.1     highr_0.8           
 [7] stats4_3.6.1         blob_1.2.0           robustbase_0.93-5   
[10] pillar_1.4.2         RSQLite_2.1.2        backports_1.1.4     
[13] lattice_0.20-38      glue_1.3.1           digest_0.6.20       
[16] colorspace_1.4-1     sandwich_2.5-1       Matrix_1.2-17       
[19] XML_3.98-1.20        lpSolve_5.6.13.3     pkgconfig_2.0.2     
[22] genefilter_1.66.0    purrr_0.3.2          ergm_3.10.4         
[25] xtable_1.8-4         mvtnorm_1.0-11       scales_1.0.0        
[28] tibble_2.1.3         annotate_1.62.0      IRanges_2.18.2      
[31] TH.data_1.0-10       withr_2.1.2          BiocGenerics_0.30.0 
[34] lazyeval_0.2.2       mime_0.7             survival_2.44-1.1   
[37] magrittr_1.5         crayon_1.3.4         statnet.common_4.3.0
[40] memoise_1.1.0        laeken_0.5.0         R.cache_0.13.0      
[43] MASS_7.3-51.4        R.rsp_0.43.1         tools_3.6.1         
[46] multcomp_1.4-10      S4Vectors_0.22.1     trust_0.1-7         
[49] munsell_0.5.0        AnnotationDbi_1.46.1 compiler_3.6.1      
[52] rlang_0.4.0          grid_3.6.1           RCurl_1.95-4.12     
[55] cwhmisc_6.6          rappdirs_0.3.1       labeling_0.3        
[58] bitops_1.0-6         base64enc_0.1-3      boot_1.3-23         
[61] gtable_0.3.0         codetools_0.2-16     DBI_1.0.0           
[64] markdown_1.1         R6_2.4.0             zoo_1.8-6           
[67] dplyr_0.8.3          bit_1.1-14           zeallot_0.1.0       
[70] parallel_3.6.1       Rcpp_1.0.2           vctrs_0.2.0         
[73] DEoptimR_1.0-8       tidyselect_0.2.5     xfun_0.9            
[76] coda_0.19-3

Total processing time was 26.41 secs.

Reproducibility

To reproduce this report, do:

html <- matrixStats:::benchmark('colRowMads_subset')

Copyright Dongcan Jiang. Last updated on 2019-09-10 20:41:51 (-0700 UTC). Powered by RSP.

<script> var link = document.createElement('link'); link.rel = 'icon'; link.href = "data:image/png;base64,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" document.getElementsByTagName('head')[0].appendChild(link); </script>
Add a custom footer
⚠️ **GitHub.com Fallback** ⚠️