matrixStats.benchmarks

x_OP_y() benchmarks on subsetted computation

This report benchmark the performance of x_OP_y() 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 vector

> x <- data[["10x10"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x10+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.004301	0.0045075	0.0061420	0.004634	0.0048170	0.149514
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.004970	0.0052440	0.0054553	0.005406	0.0055405	0.007984
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.005513	0.0059765	0.0061901	0.006101	0.0062800	0.011605

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.155545	1.163394	0.8881962	1.166595	1.150197	0.0533997
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.281795	1.325901	1.0078264	1.316573	1.303716	0.0776181

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x10+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x10+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.004234	0.004568	0.0047562	0.0046955	0.0048160	0.007858
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.005095	0.005311	0.0058629	0.0054710	0.0056495	0.030825
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.005517	0.006083	0.0062789	0.0062375	0.0063515	0.009541

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.203354	1.162653	1.232691	1.165158	1.173069	3.922754
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.303023	1.331655	1.320153	1.328399	1.318833	1.214177

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x10+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x10+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.004280	0.0044885	0.0046744	0.0046455	0.0048045	0.006006
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.004970	0.0051955	0.0054345	0.0053545	0.0055560	0.009982
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.005588	0.0060000	0.0064970	0.0061450	0.0063345	0.033749

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.161215	1.157514	1.162612	1.152621	1.156416	1.662005
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.305608	1.336750	1.389898	1.322785	1.318452	5.619214

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x10+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x10+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.004385	0.004566	0.0051200	0.0046845	0.0048125	0.042345
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.004909	0.005276	0.0054938	0.0053785	0.0055510	0.012264
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.005723	0.006022	0.0063739	0.0061630	0.0064140	0.015456

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.119498	1.155497	1.073004	1.148148	1.153455	0.2896210
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.305131	1.318879	1.244901	1.315615	1.332779	0.3650018

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x10+div data. Outliers are displayed as crosses. Times are in milliseconds.

100x100 vector

> x <- data[["100x100"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x100+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.030722	0.0318580	0.0336242	0.032477	0.0350165	0.066095
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.034107	0.0350375	0.0375856	0.036580	0.0388505	0.052725
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.041637	0.0434030	0.0455389	0.045449	0.0473550	0.060698

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.110182	1.099802	1.117813	1.126336	1.109491	0.7977154
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.355283	1.362389	1.354350	1.399421	1.352362	0.9183448

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x100+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x100+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.027959	0.0302150	0.0317816	0.0309780	0.0331180	0.048815
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.030694	0.0325985	0.0344686	0.0342325	0.0359345	0.046504
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.038741	0.0415725	0.0437470	0.0424100	0.0453460	0.083317

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.097822	1.078885	1.084546	1.105058	1.085044	0.952658
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.385636	1.375889	1.376489	1.369036	1.369225	1.706791

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x100+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x100+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.029126	0.0304600	0.0319780	0.031304	0.0331830	0.061293
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.031056	0.0322445	0.0338619	0.033227	0.0348345	0.049440
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.040009	0.0418450	0.0442752	0.043635	0.0455880	0.064907

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.066264	1.058585	1.058911	1.061430	1.049770	0.8066174
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.373652	1.373769	1.384551	1.393911	1.373836	1.0589627

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x100+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x100+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.026766	0.0274325	0.0289760	0.0287645	0.029319	0.059735
1	x_OP_y_x_y_S	0.028858	0.0298285	0.0317642	0.0315380	0.032823	0.047012
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.039978	0.0417295	0.0439793	0.0435365	0.045425	0.058511

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	x_OP_y_x_y_S	1.078159	1.087342	1.096222	1.096421	1.119513	0.7870093
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.493611	1.521170	1.517780	1.513550	1.549337	0.9795095

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x100+div data. Outliers are displayed as crosses. Times are in milliseconds.

1000x10 vector

> x <- data[["1000x10"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x10+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.029856	0.031577	0.0329250	0.0329365	0.0341370	0.039397
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.037128	0.038577	0.0404953	0.0404185	0.0421510	0.052627
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.043355	0.045181	0.0481844	0.0477040	0.0493415	0.087143

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.243569	1.22168	1.229924	1.227164	1.234760	1.335812
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.452137	1.43082	1.463456	1.448363	1.445396	2.211920

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x10+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x10+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.027611	0.0289790	0.0311371	0.0310470	0.0324020	0.046393
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.033189	0.0352065	0.0371865	0.0369050	0.0384455	0.049039
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.040402	0.0431115	0.0459137	0.0455685	0.0475725	0.084907

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.202021	1.214897	1.194282	1.188682	1.186516	1.057035
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.463257	1.487681	1.474565	1.467726	1.468196	1.830168

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x10+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x10+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.027365	0.0290075	0.0307944	0.0298640	0.0314190	0.047742
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.032633	0.0349620	0.0369028	0.0358315	0.0383185	0.053385
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.040491	0.0437325	0.0461154	0.0457665	0.0475315	0.091787

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.192509	1.205275	1.198360	1.199823	1.219596	1.118198
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.479664	1.507627	1.497526	1.532497	1.512827	1.922563

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x10+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x10+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.027338	0.0290460	0.0305179	0.0298460	0.0317330	0.043400
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.029133	0.0308940	0.0326957	0.0315075	0.0331765	0.066807
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.040056	0.0430065	0.0456537	0.0445890	0.0475905	0.059132

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.065659	1.063623	1.071361	1.055669	1.045489	1.539332
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.465213	1.480634	1.495964	1.493969	1.499716	1.362488

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x10+div data. Outliers are displayed as crosses. Times are in milliseconds.

10x1000 vector

> x <- data[["10x1000"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x1000+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.030967	0.0327145	0.0343475	0.0344375	0.0355545	0.046409
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.035892	0.0372035	0.0394236	0.0393195	0.0412400	0.055153
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.043558	0.0456250	0.0483569	0.0479815	0.0500700	0.086439

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.159040	1.137217	1.147786	1.141764	1.159909	1.188412
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.406594	1.394642	1.407875	1.393292	1.408261	1.862548

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x1000+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x1000+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.028568	0.0302475	0.0326353	0.0322940	0.033914	0.063179
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.032340	0.0337250	0.0364054	0.0365165	0.037871	0.055650
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.041306	0.0429045	0.0462613	0.0463095	0.048222	0.070748

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.132036	1.114968	1.115525	1.130752	1.116677	0.8808307
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.445884	1.418448	1.417524	1.433997	1.421891	1.1198025

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x1000+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x1000+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.029175	0.0307105	0.0330678	0.0322040	0.0346325	0.062226
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.032665	0.0341040	0.0362512	0.0358730	0.0381560	0.052736
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.042831	0.0437320	0.0470083	0.0468335	0.0491600	0.061798

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.119623	1.110500	1.096269	1.113930	1.101740	0.8474914
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.468072	1.424008	1.421574	1.454276	1.419476	0.9931218

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x1000+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x1000+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.028529	0.0297495	0.0314314	0.0308665	0.0320280	0.047564
1	x_OP_y_x_y_S	0.028690	0.0300955	0.0317838	0.0309475	0.0324165	0.045763
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.042791	0.0434645	0.0461065	0.0449095	0.0474145	0.086824

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	x_OP_y_x_y_S	1.005643	1.011630	1.011211	1.002624	1.012130	0.9621352
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.499912	1.461016	1.466892	1.454959	1.480408	1.8254142

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+10x1000+div data. Outliers are displayed as crosses. Times are in milliseconds.

100x1000 vector

> x <- data[["100x1000"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x1000+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.186796	0.197073	0.2214871	0.2053535	0.2428615	0.323011
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.202905	0.213589	0.2394878	0.2256165	0.2586855	0.341566
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.252993	0.267259	0.3017059	0.2754915	0.3257935	0.505944

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.086239	1.083806	1.081272	1.098674	1.065157	1.057444
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.354381	1.356142	1.362183	1.341548	1.341479	1.566337

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x1000+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x1000+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.173985	0.1792525	0.2082686	0.1933385	0.230801	0.308099
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.178718	0.1879090	0.2141691	0.2017905	0.232735	0.296289
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.241083	0.2534490	0.2939490	0.2835990	0.315175	0.499924

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.027203	1.048292	1.028331	1.043716	1.008379	0.9616682
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.385654	1.413922	1.411394	1.466852	1.365570	1.6226083

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x1000+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x1000+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.175891	0.1808675	0.2056625	0.1959665	0.2229560	0.315525
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.179475	0.1848660	0.2146579	0.1998310	0.2377745	0.357214
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.242626	0.2504795	0.2873048	0.2788640	0.3075145	0.407485

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.020376	1.022107	1.043739	1.019720	1.066464	1.132126
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.379411	1.384878	1.396972	1.423019	1.379261	1.291451

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x1000+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x1000+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.147238	0.165022	0.1962571	0.1859460	0.2129895	0.332220
1	x_OP_y_x_y_S	0.173023	0.199325	0.2330480	0.2252210	0.2837100	0.312995
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.240177	0.276577	0.3360721	0.3193585	0.4119960	0.442113

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	x_OP_y_x_y_S	1.175125	1.207869	1.187462	1.211217	1.332037	0.9421317
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.631216	1.676001	1.712407	1.717480	1.934349	1.3307838

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+100x1000+div data. Outliers are displayed as crosses. Times are in milliseconds.

1000x100 vector

> x <- data[["1000x100"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x100+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.180843	0.1880885	0.2119209	0.1963410	0.2285125	0.309191
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.197919	0.2087795	0.2311080	0.2192845	0.2499530	0.327220
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.245555	0.2594810	0.2907359	0.2715310	0.3132960	0.485465

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.094424	1.110007	1.090539	1.116855	1.093826	1.058310
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.357835	1.379569	1.371907	1.382956	1.371023	1.570114

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x100+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x100+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.168634	0.1736620	0.1981503	0.1887375	0.2126645	0.291207
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.174055	0.1816790	0.2036345	0.1951350	0.2209715	0.344913
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.233615	0.2404315	0.2707032	0.2536875	0.2950505	0.395924

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.032147	1.046164	1.027677	1.033896	1.039062	1.184425
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.385338	1.384480	1.366151	1.344129	1.387399	1.359596

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x100+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x100+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.174094	0.1816295	0.2007525	0.1892840	0.2187415	0.277477
1	x_OP_y_x_y_S	0.170311	0.1800660	0.2057349	0.2025870	0.2219100	0.273469
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.235386	0.2468720	0.2791036	0.2617945	0.2986935	0.429722

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.0000000	1.0000000	1.000000	1.000000	1.000000	1.0000000
1	x_OP_y_x_y_S	0.9782704	0.9913918	1.024819	1.070281	1.014485	0.9855556
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.3520627	1.3592065	1.390287	1.383078	1.365509	1.5486761

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x100+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x100+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.147386	0.1504230	0.1807286	0.1790535	0.2000180	0.302061
1	x_OP_y_x_y_S	0.168148	0.1802880	0.2065650	0.2051900	0.2230125	0.291066
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.233129	0.2396775	0.2860645	0.2701700	0.3192560	0.521570

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	x_OP_y_x_y_S	1.140868	1.198540	1.142957	1.145970	1.114962	0.9636001
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.581758	1.593357	1.582840	1.508879	1.596136	1.7267042

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on integer+1000x100+div data. 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 vector

> x <- data[["10x10"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x10+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.004007	0.0042435	0.0046154	0.004309	0.0044575	0.026006
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.004415	0.0047690	0.0051865	0.004905	0.0051000	0.019105
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.005172	0.0057000	0.0059797	0.005832	0.0059885	0.011790

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.101822	1.123836	1.123759	1.138315	1.144139	0.7346382
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.290741	1.343231	1.295611	1.353446	1.343466	0.4533569

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x10+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x10+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.003924	0.0042500	0.0046433	0.0043730	0.0045035	0.019671
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.004600	0.0048475	0.0052752	0.0049585	0.0051495	0.027715
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.005362	0.0057805	0.0059265	0.0058670	0.0060230	0.008279

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.172273	1.140588	1.136079	1.133890	1.143444	1.4089268
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.366463	1.360118	1.276343	1.341642	1.337404	0.4208734

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x10+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x10+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.004143	0.0044090	0.0046136	0.0045275	0.0046970	0.008978
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.004739	0.0050370	0.0052290	0.0051570	0.0052710	0.009482
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.005533	0.0059855	0.0065448	0.0060825	0.0062325	0.031994

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.143857	1.142436	1.133396	1.139039	1.122206	1.056137
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.335506	1.357564	1.418604	1.343457	1.326911	3.563600

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x10+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x10+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.004190	0.0044060	0.0048055	0.0045145	0.0046615	0.028665
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.004829	0.0050485	0.0052809	0.0052135	0.0053430	0.010504
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.005602	0.0059510	0.0061455	0.0060545	0.0062095	0.012102

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.152506	1.145824	1.098922	1.154834	1.146198	0.3664399
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.336993	1.350658	1.278839	1.341123	1.332082	0.4221873

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x10+div data. Outliers are displayed as crosses. Times are in milliseconds.

100x100 vector

> x <- data[["100x100"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x100+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.018921	0.0200560	0.0208662	0.0205905	0.0209585	0.047946
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.020011	0.0212110	0.0220968	0.0216275	0.0220920	0.035997
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.034958	0.0372825	0.0384677	0.0381305	0.0385890	0.053472

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.057608	1.057589	1.058975	1.050363	1.054083	0.7507821
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.847577	1.858920	1.843544	1.851849	1.841210	1.1152547

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x100+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x100+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.019884	0.020962	0.0219376	0.0215945	0.0222645	0.036253
1	x_OP_y_x_y_S	0.019865	0.021299	0.0228464	0.0223555	0.0230205	0.037736
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.035511	0.037379	0.0396596	0.0393995	0.0400780	0.078540

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.0000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	x_OP_y_x_y_S	0.9990445	1.016077	1.041427	1.035241	1.033955	1.040907
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.7859083	1.783179	1.807835	1.824515	1.800085	2.166441

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x100+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x100+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.016755	0.0179470	0.0197936	0.0188665	0.0215640	0.043107
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.017068	0.0184285	0.0195337	0.0189670	0.0195440	0.033667
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.031789	0.0347395	0.0367148	0.0362525	0.0386495	0.050281

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.0000000	1.000000	1.0000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.018681	1.026829	0.9868685	1.005327	0.9063254	0.781010
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.897284	1.935672	1.8548853	1.921528	1.7923159	1.166423

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x100+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x100+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.020658	0.0211830	0.0219019	0.021704	0.0226035	0.028553
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.021397	0.0224315	0.0238046	0.022957	0.0237585	0.056920
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.035703	0.0375210	0.0394449	0.038391	0.0402820	0.058048

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.035773	1.058939	1.086873	1.057731	1.051098	1.993486
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.728289	1.771279	1.800982	1.768844	1.782113	2.032991

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x100+div data. Outliers are displayed as crosses. Times are in milliseconds.

1000x10 vector

> x <- data[["1000x10"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x10+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.017948	0.0190455	0.0196756	0.019635	0.0199815	0.026537
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.023462	0.0247585	0.0258214	0.025403	0.0261505	0.038817
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.035203	0.0380690	0.0400503	0.039363	0.0401285	0.074188

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.307221	1.299966	1.312355	1.293761	1.308736	1.462750
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.961388	1.998845	2.035529	2.004736	2.008283	2.795644

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x10+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x10+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.019271	0.0205785	0.0219720	0.0214275	0.0222635	0.036208
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.023510	0.0245950	0.0256708	0.0257655	0.0261990	0.031045
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.038267	0.0393525	0.0418808	0.0416540	0.0421725	0.079506

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.00000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.219968	1.195179	1.168343	1.20245	1.176769	0.8574072
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.985730	1.912311	1.906099	1.94395	1.894244	2.1958131

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x10+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x10+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.015671	0.016612	0.0177387	0.0172290	0.0185950	0.032977
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.020567	0.021725	0.0228417	0.0224510	0.0233540	0.038974
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.033540	0.035821	0.0380847	0.0378045	0.0382185	0.072395

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.312424	1.307789	1.287680	1.303094	1.255929	1.181854
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	2.140259	2.156333	2.146989	2.194237	2.055311	2.195318

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x10+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x10+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.017455	0.0183825	0.0194611	0.0188340	0.0194310	0.049882
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.022128	0.0226470	0.0238869	0.0231430	0.0235045	0.070870
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.033198	0.0342730	0.0360986	0.0352025	0.0362200	0.055394

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.267717	1.231987	1.227422	1.228788	1.209639	1.420753
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.901919	1.864436	1.854915	1.869093	1.864032	1.110501

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x10+div data. Outliers are displayed as crosses. Times are in milliseconds.

10x1000 vector

> x <- data[["10x1000"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x1000+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.019165	0.0200490	0.0213936	0.0210015	0.0217435	0.040746
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.022015	0.0224620	0.0247404	0.0241405	0.0252130	0.046853
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.037404	0.0387665	0.0416224	0.0412025	0.0425965	0.087210

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.148709	1.120355	1.156442	1.149466	1.159565	1.149880
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.951683	1.933588	1.945554	1.961884	1.959045	2.140333

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x1000+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x1000+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.021663	0.0230445	0.0244748	0.0235825	0.0239975	0.057882
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.023220	0.0243250	0.0258979	0.0251470	0.0256180	0.058005
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.040878	0.0426105	0.0453638	0.0440105	0.0445710	0.090679

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.071874	1.055566	1.058149	1.066342	1.067528	1.002125
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.886996	1.849053	1.853495	1.866236	1.857319	1.566618

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x1000+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x1000+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.018644	0.0199625	0.0212947	0.0206335	0.0216025	0.049270
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.019222	0.0206365	0.0216590	0.0214130	0.0221410	0.036155
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.036794	0.0391210	0.0412337	0.0408730	0.0419065	0.065178

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.031002	1.033763	1.017109	1.037778	1.024928	0.7338137
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.973504	1.959725	1.936335	1.980905	1.939891	1.3228740

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x1000+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x1000+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.020488	0.0217060	0.0223496	0.0220785	0.022611	0.027915
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.023204	0.0247455	0.0257294	0.0253595	0.025990	0.038179
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.038994	0.0408990	0.0430385	0.0418220	0.043412	0.089404

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.132565	1.140030	1.151225	1.148606	1.149441	1.367688
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.903260	1.884226	1.925697	1.894241	1.919950	3.202723

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+10x1000+div data. Outliers are displayed as crosses. Times are in milliseconds.

100x1000 vector

> x <- data[["100x1000"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x1000+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.111212	0.1171325	0.1428754	0.1264610	0.1461250	0.246682
1	x_OP_y_x_y_S	0.112956	0.1240785	0.1563260	0.1325065	0.1806025	0.256439
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.221258	0.2361565	0.3000092	0.2565715	0.3174245	0.476788

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	x_OP_y_x_y_S	1.015682	1.059300	1.094142	1.047805	1.235945	1.039553
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.989515	2.016148	2.099795	2.028859	2.172281	1.932804

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x1000+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x1000+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.099193	0.1084515	0.1216147	0.1185555	0.129359	0.181530
1	x_OP_y_x_y_S	0.113058	0.1194550	0.1381332	0.1342135	0.149449	0.199032
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.211295	0.2368895	0.2643763	0.2567880	0.282842	0.464185

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
1	x_OP_y_x_y_S	1.139778	1.10146	1.135827	1.132073	1.155304	1.096414
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	2.130140	2.18429	2.173885	2.165973	2.186489	2.557071

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x1000+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x1000+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.090981	0.099252	0.1115135	0.1066405	0.1221585	0.249941
1	x_OP_y_x_y_S	0.092870	0.103781	0.1157310	0.1112165	0.1258340	0.184918
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.198043	0.217417	0.2398771	0.2366935	0.2543620	0.320824

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	x_OP_y_x_y_S	1.020763	1.045631	1.037821	1.042911	1.030088	0.7398466
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	2.176751	2.190555	2.151104	2.219546	2.082229	1.2835989

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x1000+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x1000+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.110208	0.122033	0.1376094	0.1345310	0.1491830	0.198168
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.108664	0.124796	0.1410654	0.1374735	0.1560285	0.303923
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.206159	0.224439	0.2553685	0.2498510	0.2713090	0.352952

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.0000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.9859901	1.022641	1.025115	1.021872	1.045887	1.533663
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.8706355	1.839166	1.855750	1.857200	1.818632	1.781075

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+100x1000+div data. Outliers are displayed as crosses. Times are in milliseconds.

1000x100 vector

> x <- data[["1000x100"]]
> y <- x[, 1L]
> xrows <- sample.int(nrow(x), size = nrow(x) * 0.7)
> xcols <- sample.int(ncol(x), size = ncol(x) * 0.7)
> x_S <- x[xrows, xcols]
> yidxs <- xrows
> y_S <- y[yidxs]

> OP
[1] "+"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x100+add data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.100184	0.1099710	0.1215489	0.1169825	0.1297705	0.177145
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.103078	0.1124000	0.1255489	0.1238145	0.1360605	0.166470
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.199365	0.2211095	0.2454439	0.2370905	0.2626015	0.443893

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.028887	1.022088	1.032908	1.058402	1.048470	0.9397386
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.989988	2.010616	2.019301	2.026718	2.023584	2.5058173

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x100+add data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "-"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x100+sub data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.098278	0.1097645	0.1248652	0.1199835	0.1330065	0.253161
1	x_OP_y_x_y_S	0.105373	0.1130865	0.1319000	0.1284740	0.1397205	0.217976
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.201413	0.2194830	0.2518723	0.2488705	0.2701825	0.379458

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	x_OP_y_x_y_S	1.072193	1.030265	1.056340	1.070764	1.050479	0.8610173
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	2.049421	1.999581	2.017154	2.074206	2.031348	1.4988802

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x100+sub data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "*"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x100+mul data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.100973	0.1060295	0.1231831	0.1167420	0.1307630	0.353595
1	x_OP_y_x_y_S	0.097044	0.1136845	0.1300198	0.1257265	0.1354095	0.289740
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.210217	0.2282090	0.2652190	0.2484985	0.2884675	0.630044

	expr	min	lq	mean	median	uq	max
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.0000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	x_OP_y_x_y_S	0.9610886	1.072197	1.055501	1.076960	1.035534	0.819412
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	2.0819130	2.152316	2.153048	2.128613	2.206033	1.781824

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x100+mul data. Outliers are displayed as crosses. Times are in milliseconds.

> OP
[1] "/"
> stats <- microbenchmark(x_OP_y_x_y_S = x_OP_y(x_S, y_S, OP = OP, na.rm = FALSE), `x_OP_y(x, y, OP, xrows, xcols, yidxs)` = x_OP_y(x, 
+     y, OP = OP, xrows = xrows, xcols = xcols, yidxs = yidxs, na.rm = FALSE), `x_OP_y(x[xrows, xcols], y[yidxs], OP)` = x_OP_y(x[xrows, 
+     xcols], y[yidxs], OP = OP, na.rm = FALSE), unit = "ms")

Table: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x100+div data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	0.110896	0.127263	0.1412336	0.1370605	0.1531265	0.264155
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	0.115811	0.122829	0.1476885	0.1412225	0.1533625	0.333660
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	0.212906	0.230869	0.2707054	0.2538215	0.2939905	0.496476

	expr	min	lq	mean	median	uq	max
1	x_OP_y_x_y_S	1.000000	1.0000000	1.000000	1.000000	1.000000	1.000000
2	x_OP_y(x, y, OP, xrows, xcols, yidxs)	1.044321	0.9651588	1.045704	1.030366	1.001541	1.263122
3	x_OP_y(x[xrows, xcols], y[yidxs], OP)	1.919871	1.8141094	1.916721	1.851894	1.919919	1.879487

Figure: Benchmarking of x_OP_y_x_y_S(), x_OP_y(x, y, OP, xrows, xcols, yidxs)() and x_OP_y(x[xrows, xcols], y[yidxs], OP)() on double+1000x100+div data. Outliers are displayed as crosses. Times are in milliseconds.

Appendix

Session information

R version 4.1.1 Patched (2021-08-10 r80727)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.5 LTS

Matrix products: default
BLAS:   /home/hb/software/R-devel/R-4-1-branch/lib/R/lib/libRblas.so
LAPACK: /home/hb/software/R-devel/R-4-1-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_1.4-7   matrixStats_0.60.1     ggplot2_3.3.5         
[4] knitr_1.33             R.devices_2.17.0       R.utils_2.10.1        
[7] R.oo_1.24.0            R.methodsS3_1.8.1-9001 history_0.0.1-9000    

loaded via a namespace (and not attached):
 [1] Biobase_2.52.0          httr_1.4.2              splines_4.1.1          
 [4] bit64_4.0.5             network_1.17.1          assertthat_0.2.1       
 [7] highr_0.9               stats4_4.1.1            blob_1.2.2             
[10] GenomeInfoDbData_1.2.6  robustbase_0.93-8       pillar_1.6.2           
[13] RSQLite_2.2.8           lattice_0.20-44         limma_3.48.3           
[16] glue_1.4.2              digest_0.6.27           XVector_0.32.0         
[19] colorspace_2.0-2        Matrix_1.3-4            XML_3.99-0.7           
[22] pkgconfig_2.0.3         zlibbioc_1.38.0         genefilter_1.74.0      
[25] purrr_0.3.4             ergm_4.1.2              xtable_1.8-4           
[28] scales_1.1.1            tibble_3.1.4            annotate_1.70.0        
[31] KEGGREST_1.32.0         farver_2.1.0            generics_0.1.0         
[34] IRanges_2.26.0          ellipsis_0.3.2          cachem_1.0.6           
[37] withr_2.4.2             BiocGenerics_0.38.0     mime_0.11              
[40] survival_3.2-13         magrittr_2.0.1          crayon_1.4.1           
[43] statnet.common_4.5.0    memoise_2.0.0           laeken_0.5.1           
[46] fansi_0.5.0             R.cache_0.15.0          MASS_7.3-54            
[49] R.rsp_0.44.0            progressr_0.8.0         tools_4.1.1            
[52] lifecycle_1.0.0         S4Vectors_0.30.0        trust_0.1-8            
[55] munsell_0.5.0           tabby_0.0.1-9001        AnnotationDbi_1.54.1   
[58] Biostrings_2.60.2       compiler_4.1.1          GenomeInfoDb_1.28.1    
[61] rlang_0.4.11            grid_4.1.1              RCurl_1.98-1.4         
[64] cwhmisc_6.6             rappdirs_0.3.3          startup_0.15.0         
[67] labeling_0.4.2          bitops_1.0-7            base64enc_0.1-3        
[70] boot_1.3-28             gtable_0.3.0            DBI_1.1.1              
[73] markdown_1.1            R6_2.5.1                lpSolveAPI_5.5.2.0-17.7
[76] rle_0.9.2               dplyr_1.0.7             fastmap_1.1.0          
[79] bit_4.0.4               utf8_1.2.2              parallel_4.1.1         
[82] Rcpp_1.0.7              vctrs_0.3.8             png_0.1-7              
[85] DEoptimR_1.0-9          tidyselect_1.1.1        xfun_0.25              
[88] coda_0.19-4            

Total processing time was 45.43 secs.

Reproducibility

To reproduce this report, do:

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