matrixStats.benchmarks

colMeans2() and rowMeans2() benchmarks on subsetted computation

This report benchmark the performance of colMeans2() and rowMeans2() 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  5252835 280.6    7916910 422.9  7916910 422.9
Vcells 10106300  77.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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  5240507 279.9    7916910 422.9  7916910 422.9
Vcells 10065440  76.8   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.002294	0.0024120	0.0043432	0.0025045	0.0025780	0.170474
2	colMeans2(X, rows, cols)	0.002652	0.0027815	0.0031773	0.0028690	0.0029965	0.012348
3	colMeans2(X[rows, cols])	0.003272	0.0034580	0.0039271	0.0035605	0.0036925	0.017949

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
2	colMeans2(X, rows, cols)	1.156059	1.153192	0.7315624	1.145538	1.162335	0.0724333
3	colMeans2(X[rows, cols])	1.426330	1.433665	0.9042105	1.421641	1.432312	0.1052888

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.002446	0.0025150	0.0026478	0.0025895	0.0026900	0.005326
2	rowMeans2(X, cols, rows)	0.002823	0.0029035	0.0047946	0.0029715	0.0030365	0.182578
3	rowMeans2(X[cols, rows])	0.003349	0.0035715	0.0037700	0.0036675	0.0037710	0.009410

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.154129	1.154473	1.810824	1.147519	1.128810	34.280511
3	rowMeans2(X[cols, rows])	1.369174	1.420079	1.423869	1.416297	1.401859	1.766804

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on integer+10x10 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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
1	colMeans2_X_S	2.294	2.412	4.34317	2.5045	2.578	170.474
2	rowMeans2_X_S	2.446	2.515	2.64775	2.5895	2.690	5.326

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.00000	1.000000	1.0000000	1.000000	1.000000	1.0000000
2	rowMeans2_X_S	1.06626	1.042703	0.6096354	1.033939	1.043445	0.0312423

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_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 5238997 279.8    7916910 422.9  7916910 422.9
Vcells 9733104  74.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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 5238991 279.8    7916910 422.9  7916910 422.9
Vcells 9738187  74.3   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.013164	0.0134095	0.0138760	0.0135765	0.0139710	0.022677
2	colMeans2(X, rows, cols)	0.016860	0.0170785	0.0178900	0.0172790	0.0179740	0.031211
3	colMeans2(X[rows, cols])	0.024562	0.0249020	0.0262237	0.0251175	0.0261345	0.067093

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colMeans2(X, rows, cols)	1.280766	1.273612	1.289278	1.272714	1.286522	1.376328
3	colMeans2(X[rows, cols])	1.865846	1.857042	1.889855	1.850072	1.870625	2.958636

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.013154	0.0134565	0.0139070	0.013717	0.0141730	0.022962
2	rowMeans2(X, cols, rows)	0.016143	0.0164220	0.0171836	0.016903	0.0171945	0.038360
3	rowMeans2(X[cols, rows])	0.024383	0.0248665	0.0257033	0.025244	0.0259545	0.041047

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.227231	1.220377	1.235610	1.232266	1.213187	1.670586
3	rowMeans2(X[cols, rows])	1.853657	1.847917	1.848229	1.840344	1.831264	1.787606

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on integer+100x100 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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
1	colMeans2_X_S	13.164	13.4095	13.87601	13.5765	13.971	22.677
2	rowMeans2_X_S	13.154	13.4565	13.90698	13.7170	14.173	22.962

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.0000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2_X_S	0.9992404	1.003505	1.002232	1.010349	1.014458	1.012568

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_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 5239739 279.9    7916910 422.9  7916910 422.9
Vcells 9737142  74.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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 5239733 279.9    7916910 422.9  7916910 422.9
Vcells 9742225  74.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.013525	0.0138060	0.0141308	0.0140575	0.0143785	0.017029
2	colMeans2(X, rows, cols)	0.017440	0.0177395	0.0183813	0.0179230	0.0184440	0.033282
3	colMeans2(X[rows, cols])	0.025793	0.0260755	0.0268971	0.0263030	0.0268290	0.056145

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colMeans2(X, rows, cols)	1.289464	1.284912	1.300793	1.274978	1.282749	1.954431
3	colMeans2(X[rows, cols])	1.907061	1.888708	1.903439	1.871101	1.865911	3.297023

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.013471	0.013852	0.0152808	0.0140745	0.0145335	0.027956
2	rowMeans2(X, cols, rows)	0.017151	0.017873	0.0195282	0.0182540	0.0191740	0.037533
3	rowMeans2(X[cols, rows])	0.027436	0.027982	0.0332071	0.0286495	0.0339845	0.058162

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.273179	1.290283	1.277957	1.296955	1.319297	1.342574
3	rowMeans2(X[cols, rows])	2.036671	2.020069	2.173123	2.035561	2.338356	2.080484

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on integer+1000x10 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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
1	colMeans2_X_S	13.525	13.806	14.13082	14.0575	14.3785	17.029
2	rowMeans2_X_S	13.471	13.852	15.28083	14.0745	14.5335	27.956

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.0000000	1.000000	1.000000	1.000000	1.00000	1.00000
2	rowMeans2_X_S	0.9960074	1.003332	1.081383	1.001209	1.01078	1.64167

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_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 5239943 279.9    7916910 422.9  7916910 422.9
Vcells 9738013  74.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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 5239937 279.9    7916910 422.9  7916910 422.9
Vcells 9743096  74.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.014418	0.0150675	0.0157001	0.0155160	0.0157495	0.032616
2	colMeans2(X, rows, cols)	0.020543	0.0209690	0.0217421	0.0214040	0.0219145	0.036509
3	colMeans2(X[rows, cols])	0.028129	0.0289605	0.0300164	0.0296105	0.0300815	0.045692

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colMeans2(X, rows, cols)	1.424816	1.391671	1.384839	1.379479	1.391441	1.119359
3	colMeans2(X[rows, cols])	1.950964	1.922051	1.911863	1.908385	1.909997	1.400907

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.014402	0.014888	0.0152724	0.0150735	0.0152415	0.023331
2	rowMeans2(X, cols, rows)	0.019989	0.020267	0.0208216	0.0204160	0.0205635	0.058464
3	rowMeans2(X[cols, rows])	0.026293	0.026972	0.0275260	0.0272370	0.0275170	0.037526

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.387932	1.361298	1.363350	1.354430	1.349178	2.505851
3	rowMeans2(X[cols, rows])	1.825649	1.811660	1.802339	1.806946	1.805400	1.608418

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on integer+10x1000 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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	rowMeans2_X_S	14.402	14.8880	15.27238	15.0735	15.2415	23.331
1	colMeans2_X_S	14.418	15.0675	15.70009	15.5160	15.7495	32.616

	expr	min	lq	mean	median	uq	max
2	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.00000	1.000000
1	colMeans2_X_S	1.001111	1.012057	1.028005	1.029356	1.03333	1.397968

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_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 5240156 279.9    7916910 422.9  7916910 422.9
Vcells 9760690  74.5   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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 5240150 279.9    7916910 422.9  7916910 422.9
Vcells 9810773  74.9   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.083261	0.0862270	0.0973144	0.0912025	0.1077500	0.174421
2	colMeans2(X, rows, cols)	0.108624	0.1124885	0.1247937	0.1209180	0.1402155	0.206528
3	colMeans2(X[rows, cols])	0.166429	0.1731815	0.1969018	0.1919690	0.2160765	0.331763

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colMeans2(X, rows, cols)	1.304620	1.304562	1.282377	1.325819	1.301304	1.184078
3	colMeans2(X[rows, cols])	1.998883	2.008437	2.023357	2.104865	2.005350	1.902082

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.083326	0.0889110	0.0985526	0.0968415	0.1074245	0.150370
2	rowMeans2(X, cols, rows)	0.104037	0.1077665	0.1187866	0.1114585	0.1296540	0.194352
3	rowMeans2(X[cols, rows])	0.162011	0.1700415	0.1904480	0.1888250	0.2101220	0.263337

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.248554	1.212072	1.205311	1.150937	1.206931	1.292492
3	rowMeans2(X[cols, rows])	1.944303	1.912491	1.932451	1.949836	1.955997	1.751260

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on integer+100x1000 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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	colMeans2_X_S	83.261	86.227	97.31441	91.2025	107.7500	174.421
2	rowMeans2_X_S	83.326	88.911	98.55260	96.8415	107.4245	150.370

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.0000000	1.0000000
2	rowMeans2_X_S	1.000781	1.031127	1.012724	1.061829	0.9969791	0.8621095

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_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 5240366 279.9    7916910 422.9  7916910 422.9
Vcells 9761473  74.5   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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 5240360 279.9    7916910 422.9  7916910 422.9
Vcells 9811556  74.9   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.078729	0.0838265	0.0904083	0.0867000	0.0957645	0.129325
2	colMeans2(X, rows, cols)	0.096257	0.1027300	0.1125829	0.1092610	0.1177165	0.141283
3	colMeans2(X[rows, cols])	0.154079	0.1642395	0.1815537	0.1714335	0.1961330	0.266746

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colMeans2(X, rows, cols)	1.222637	1.225507	1.245272	1.260219	1.229229	1.092465
3	colMeans2(X[rows, cols])	1.957081	1.959279	2.008154	1.977318	2.048076	2.062602

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.079396	0.0852955	0.0950694	0.0905880	0.1035715	0.136084
2	rowMeans2(X, cols, rows)	0.097017	0.1060405	0.1176491	0.1104115	0.1280920	0.226063
3	rowMeans2(X[cols, rows])	0.163346	0.1743350	0.1912017	0.1810720	0.2131840	0.246745

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.221938	1.243213	1.237508	1.218831	1.236749	1.661202
3	rowMeans2(X[cols, rows])	2.057358	2.043894	2.011181	1.998852	2.058327	1.813182

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on integer+1000x100 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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	colMeans2_X_S	78.729	83.8265	90.40829	86.700	95.7645	129.325
2	rowMeans2_X_S	79.396	85.2955	95.06935	90.588	103.5715	136.084

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2_X_S	1.008472	1.017524	1.051556	1.044844	1.081523	1.052264

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_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 5240584 279.9    7916910 422.9  7916910 422.9
Vcells 9852586  75.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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 5240569 279.9    7916910 422.9  7916910 422.9
Vcells 9852754  75.2   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.002482	0.002542	0.0028578	0.0026080	0.0027160	0.022842
2	colMeans2(X, rows, cols)	0.002821	0.002966	0.0030825	0.0030130	0.0031275	0.005631
3	colMeans2(X[rows, cols])	0.003461	0.003654	0.0038629	0.0037685	0.0038890	0.010141

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colMeans2(X, rows, cols)	1.136583	1.166798	1.078598	1.155291	1.151510	0.2465196
3	colMeans2(X[rows, cols])	1.394440	1.437451	1.351673	1.444977	1.431885	0.4439629

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.002389	0.0024620	0.0026118	0.0025515	0.0026835	0.005730
2	rowMeans2(X, cols, rows)	0.002704	0.0028515	0.0032448	0.0029215	0.0030530	0.028284
3	rowMeans2(X[cols, rows])	0.003285	0.0035380	0.0036989	0.0036060	0.0037600	0.008246

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.131854	1.158205	1.242346	1.145013	1.137693	4.936126
3	rowMeans2(X[cols, rows])	1.375052	1.437043	1.416224	1.413286	1.401155	1.439093

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on double+10x10 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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	rowMeans2_X_S	2.389	2.462	2.61184	2.5515	2.6835	5.730
1	colMeans2_X_S	2.482	2.542	2.85785	2.6080	2.7160	22.842

	expr	min	lq	mean	median	uq	max
2	rowMeans2_X_S	1.000000	1.000000	1.00000	1.000000	1.000000	1.000000
1	colMeans2_X_S	1.038928	1.032494	1.09419	1.022144	1.012111	3.986387

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_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 5240782 279.9    7916910 422.9  7916910 422.9
Vcells 9858541  75.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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 5240776 279.9    7916910 422.9  7916910 422.9
Vcells 9868624  75.3   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.010334	0.0106855	0.0132859	0.0111200	0.0138825	0.023145
2	colMeans2(X, rows, cols)	0.014852	0.0152460	0.0186061	0.0156625	0.0166375	0.039959
3	colMeans2(X[rows, cols])	0.026213	0.0266285	0.0313546	0.0274930	0.0280825	0.097558

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colMeans2(X, rows, cols)	1.437198	1.426793	1.400439	1.408498	1.198451	1.726464
3	colMeans2(X[rows, cols])	2.536578	2.492022	2.359983	2.472392	2.022870	4.215079

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.010809	0.0114975	0.0132233	0.011789	0.0124035	0.037322
2	rowMeans2(X, cols, rows)	0.015564	0.0161410	0.0185190	0.016456	0.0172920	0.059948
3	rowMeans2(X[cols, rows])	0.027363	0.0280370	0.0313376	0.028601	0.0298250	0.077409

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.439911	1.40387	1.400480	1.395878	1.394123	1.606238
3	rowMeans2(X[cols, rows])	2.531502	2.43853	2.369871	2.426075	2.404563	2.074085

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on double+100x100 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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
1	colMeans2_X_S	10.334	10.6855	13.28593	11.120	13.8825	23.145
2	rowMeans2_X_S	10.809	11.4975	13.22333	11.789	12.4035	37.322

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.0000000	1.000000	1.000000	1.00000
2	rowMeans2_X_S	1.045965	1.075991	0.9952882	1.060162	0.893463	1.61253

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_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 5240981 279.9    7916910 422.9  7916910 422.9
Vcells 9859959  75.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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 5240975 279.9    7916910 422.9  7916910 422.9
Vcells 9870042  75.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.010889	0.011438	0.0118635	0.0116950	0.0118645	0.026262
2	colMeans2(X, rows, cols)	0.014526	0.015511	0.0160361	0.0158090	0.0161345	0.030255
3	colMeans2(X[rows, cols])	0.027583	0.027943	0.0289620	0.0283485	0.0291410	0.060104

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colMeans2(X, rows, cols)	1.334007	1.356094	1.351726	1.351774	1.359897	1.152045
3	colMeans2(X[rows, cols])	2.533107	2.442997	2.441274	2.423985	2.456151	2.288630

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.010785	0.011305	0.0120096	0.0116590	0.0119500	0.025670
2	rowMeans2(X, cols, rows)	0.014448	0.015365	0.0160376	0.0156750	0.0159625	0.026501
3	rowMeans2(X[cols, rows])	0.028587	0.029910	0.0316715	0.0307145	0.0310805	0.089509

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.339638	1.359133	1.335405	1.344455	1.335774	1.032372
3	rowMeans2(X[cols, rows])	2.650626	2.645732	2.637193	2.634403	2.600879	3.486911

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on double+1000x10 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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
2	rowMeans2_X_S	10.785	11.305	12.00955	11.659	11.9500	25.670
1	colMeans2_X_S	10.889	11.438	11.86346	11.695	11.8645	26.262

	expr	min	lq	mean	median	uq	max
2	rowMeans2_X_S	1.000000	1.000000	1.0000000	1.000000	1.0000000	1.000000
1	colMeans2_X_S	1.009643	1.011765	0.9878355	1.003088	0.9928452	1.023062

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_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 5241185 280.0    7916910 422.9  7916910 422.9
Vcells 9860094  75.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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 5241179 280.0    7916910 422.9  7916910 422.9
Vcells 9870177  75.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.014546	0.0152375	0.0156257	0.0153605	0.0155955	0.033381
2	colMeans2(X, rows, cols)	0.021314	0.0222540	0.0227571	0.0224390	0.0226465	0.040444
3	colMeans2(X[rows, cols])	0.033985	0.0353185	0.0357392	0.0354915	0.0357630	0.043462

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colMeans2(X, rows, cols)	1.465283	1.460476	1.456389	1.460825	1.452118	1.211587
3	colMeans2(X[rows, cols])	2.336381	2.317867	2.287208	2.310569	2.293162	1.301998

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.013485	0.0141825	0.0147644	0.0144765	0.0147680	0.028325
2	rowMeans2(X, cols, rows)	0.019153	0.0199835	0.0211735	0.0204265	0.0206745	0.061564
3	rowMeans2(X[cols, rows])	0.028680	0.0300200	0.0312564	0.0304030	0.0311810	0.045597

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.420319	1.409025	1.434093	1.411011	1.399953	2.173486
3	rowMeans2(X[cols, rows])	2.126808	2.116693	2.117015	2.100162	2.111390	1.609779

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on double+10x1000 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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	rowMeans2_X_S	13.485	14.1825	14.76439	14.4765	14.7680	28.325
1	colMeans2_X_S	14.546	15.2375	15.62570	15.3605	15.5955	33.381

	expr	min	lq	mean	median	uq	max
2	rowMeans2_X_S	1.00000	1.000000	1.000000	1.000000	1.000000	1.0000
1	colMeans2_X_S	1.07868	1.074387	1.058337	1.061065	1.056033	1.1785

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_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 5241398 280.0    7916910 422.9  7916910 422.9
Vcells 9905566  75.6   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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  5241392 280.0    7916910 422.9  7916910 422.9
Vcells 10005649  76.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.062106	0.065537	0.0725551	0.0688215	0.0795305	0.112613
2	colMeans2(X, rows, cols)	0.096242	0.101563	0.1130196	0.1063970	0.1281875	0.215617
3	colMeans2(X[rows, cols])	0.177671	0.190052	0.2093418	0.2050015	0.2291870	0.257966

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colMeans2(X, rows, cols)	1.549641	1.549705	1.557708	1.545985	1.611803	1.914672
3	colMeans2(X[rows, cols])	2.860770	2.899919	2.885282	2.978742	2.881750	2.290730

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.065601	0.0686490	0.0772700	0.0720610	0.0807190	0.170813
2	rowMeans2(X, cols, rows)	0.115845	0.1203110	0.1293163	0.1243680	0.1337185	0.246570
3	rowMeans2(X[cols, rows])	0.187867	0.1977165	0.2149916	0.2057115	0.2280220	0.354159

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.765903	1.752553	1.673564	1.725871	1.656593	1.443508
3	rowMeans2(X[cols, rows])	2.863783	2.880107	2.782342	2.854686	2.824886	2.073373

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on double+100x1000 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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	colMeans2_X_S	62.106	65.537	72.55507	68.8215	79.5305	112.613
2	rowMeans2_X_S	65.601	68.649	77.27002	72.0610	80.7190	170.813

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2_X_S	1.056275	1.047485	1.064984	1.047071	1.014944	1.516814

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_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 5241608 280.0    7916910 422.9  7916910 422.9
Vcells 9905707  75.6   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2_X_S = colMeans2(X_S, na.rm = FALSE), `colMeans2(X, rows, cols)` = colMeans2(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colMeans2(X[rows, cols])` = colMeans2(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  5241602 280.0    7916910 422.9  7916910 422.9
Vcells 10005790  76.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2_X_S = rowMeans2(X_S, na.rm = FALSE), `rowMeans2(X, cols, rows)` = rowMeans2(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowMeans2(X[cols, rows])` = rowMeans2(X[cols, rows], 
+     na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(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	colMeans2_X_S	0.063284	0.0674915	0.0728476	0.0697015	0.0740970	0.127155
2	colMeans2(X, rows, cols)	0.084262	0.0899585	0.0978887	0.0933055	0.1058415	0.143755
3	colMeans2(X[rows, cols])	0.173459	0.1859725	0.2027362	0.1939840	0.2198335	0.326939

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colMeans2(X, rows, cols)	1.331490	1.332886	1.343746	1.338644	1.428418	1.130549
3	colMeans2(X[rows, cols])	2.740961	2.755495	2.783018	2.783068	2.966834	2.571185

Table: Benchmarking of rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(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	rowMeans2_X_S	0.066900	0.0702755	0.0770423	0.0729595	0.0826495	0.134242
2	rowMeans2(X, cols, rows)	0.109936	0.1160740	0.1268164	0.1186395	0.1314250	0.301125
3	rowMeans2(X[cols, rows])	0.186958	0.1997405	0.2180105	0.2069390	0.2358975	0.306135

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2(X, cols, rows)	1.643288	1.651699	1.646062	1.626101	1.590149	2.243150
3	rowMeans2(X[cols, rows])	2.794589	2.842249	2.829752	2.836354	2.854192	2.280471

Figure: Benchmarking of colMeans2_X_S(), colMeans2(X, rows, cols)() and colMeans2(X[rows, cols])() on double+1000x100 data as well as rowMeans2_X_S(), rowMeans2(X, cols, rows)() and rowMeans2(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2_X_S() and rowMeans2_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
1	colMeans2_X_S	63.284	67.4915	72.84759	69.7015	74.0970	127.155
2	rowMeans2_X_S	66.900	70.2755	77.04228	72.9595	82.6495	134.242

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2_X_S	1.057139	1.04125	1.057582	1.046742	1.115423	1.055735

Figure: Benchmarking of colMeans2_X_S() and rowMeans2_X_S() on double+1000x100 data (original and transposed). 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_0.60.0     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         glue_1.4.2             
[16] digest_0.6.27           XVector_0.32.0          colorspace_2.0-2       
[19] Matrix_1.3-4            XML_3.99-0.7            pkgconfig_2.0.3        
[22] zlibbioc_1.38.0         genefilter_1.74.0       purrr_0.3.4            
[25] ergm_4.1.2              xtable_1.8-4            scales_1.1.1           
[28] tibble_3.1.4            annotate_1.70.0         KEGGREST_1.32.0        
[31] farver_2.1.0            generics_0.1.0          IRanges_2.26.0         
[34] ellipsis_0.3.2          cachem_1.0.6            withr_2.4.2            
[37] BiocGenerics_0.38.0     mime_0.11               survival_3.2-13        
[40] magrittr_2.0.1          crayon_1.4.1            statnet.common_4.5.0   
[43] memoise_2.0.0           laeken_0.5.1            fansi_0.5.0            
[46] R.cache_0.15.0          MASS_7.3-54             R.rsp_0.44.0           
[49] progressr_0.8.0         tools_4.1.1             lifecycle_1.0.0        
[52] S4Vectors_0.30.0        trust_0.1-8             munsell_0.5.0          
[55] tabby_0.0.1-9001        AnnotationDbi_1.54.1    Biostrings_2.60.2      
[58] compiler_4.1.1          GenomeInfoDb_1.28.1     rlang_0.4.11           
[61] grid_4.1.1              RCurl_1.98-1.4          cwhmisc_6.6            
[64] rstudioapi_0.13         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 24.22 secs.

Reproducibility

To reproduce this report, do:

html <- matrixStats:::benchmark('colRowMeans2_subset')