matrixStats.benchmarks

---

colMeans2() and rowMeans2() benchmarks

This report benchmark the performance of colMeans2() and rowMeans2() against alternative methods.

Alternative methods

• apply() + mean()
• .colMeans() and .rowMeans()
• colMeans() and rowMeans()

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"]]
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5257921 280.9    7916910 422.9  7916910 422.9
Vcells 10186358  77.8   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5244238 280.1    7916910 422.9  7916910 422.9
Vcells 10141213  77.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.002973	0.0034290	0.0045484	0.0039170	0.0047025	0.029963
1	colMeans2	0.002586	0.0032390	0.0050546	0.0041735	0.0057835	0.018183
3	colMeans	0.005649	0.0064000	0.0090514	0.0076845	0.0109675	0.031805
4	apply+mean	0.070580	0.0741775	0.0921426	0.0758115	0.1013795	0.227490

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.0000000	1.0000000	1.000000	1.000000	1.000000	1.0000000
1	colMeans2	0.8698285	0.9445903	1.111296	1.065484	1.229878	0.6068484
3	colMeans	1.9001009	1.8664334	1.990018	1.961833	2.332270	1.0614758
4	apply+mean	23.7403296	21.6324001	20.258283	19.354481	21.558639	7.5923639

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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	0.002397	0.0028305	0.0035766	0.0035125	0.0040885	0.013986
2	.rowMeans	0.004117	0.0045145	0.0048169	0.0047230	0.0049440	0.012413
3	rowMeans	0.006448	0.0070385	0.0080657	0.0075265	0.0080655	0.037236
4	apply+mean	0.067079	0.0704940	0.0738455	0.0736710	0.0749055	0.139345

	expr	min	lq	mean	median	uq	max
1	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	.rowMeans	1.717564	1.594948	1.346761	1.344626	1.209245	0.8875304
3	rowMeans	2.690029	2.486663	2.255097	2.142776	1.972728	2.6623767
4	apply+mean	27.984564	24.905140	20.646601	20.973950	18.321022	9.9631775

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on integer+10x10 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() on integer+10x10 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	rowMeans2	2.397	2.8305	3.57664	3.5125	4.0885	13.986
1	colMeans2	2.586	3.2390	5.05461	4.1735	5.7835	18.183

	expr	min	lq	mean	median	uq	max
2	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colMeans2	1.078849	1.144321	1.413229	1.188185	1.414578	1.300086

Figure: Benchmarking of colMeans2() and rowMeans2() on integer+10x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x100 integer matrix

> X <- data[["100x100"]]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5242821 280.0    7916910 422.9  7916910 422.9
Vcells 9758020  74.5   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5242815 280.0    7916910 422.9  7916910 422.9
Vcells 9763063  74.5   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.012112	0.0137450	0.0153540	0.0146185	0.0167420	0.024553
3	colMeans	0.014325	0.0166630	0.0186439	0.0177450	0.0200885	0.041164
1	colMeans2	0.017660	0.0194840	0.0213776	0.0202760	0.0221845	0.039210
4	apply+mean	0.425152	0.4657975	0.5208733	0.4980050	0.5700890	0.776684

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colMeans	1.182711	1.212295	1.214276	1.213873	1.199887	1.676537
1	colMeans2	1.458058	1.417534	1.392317	1.387010	1.325081	1.596953
4	apply+mean	35.101717	33.888505	33.924361	34.066765	34.051428	31.632957

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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	0.016040	0.0187000	0.0217922	0.0211005	0.0254330	0.037635
2	.rowMeans	0.034859	0.0369595	0.0433372	0.0415920	0.0463565	0.069092
3	rowMeans	0.036387	0.0418075	0.0465615	0.0447965	0.0486725	0.070743
4	apply+mean	0.400687	0.4319210	0.4969557	0.4827705	0.5341675	0.750953

	expr	min	lq	mean	median	uq	max
1	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowMeans	2.173254	1.976444	1.988656	1.971138	1.822691	1.835844
3	rowMeans	2.268516	2.235695	2.136609	2.123007	1.913754	1.879713
4	apply+mean	24.980486	23.097380	22.804260	22.879576	21.002929	19.953580

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on integer+100x100 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() 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	17.66	19.484	21.37758	20.2760	22.1845	39.210
2	rowMeans2	16.04	18.700	21.79223	21.1005	25.4330	37.635

	expr	min	lq	mean	median	uq	max
1	colMeans2	1.0000000	1.0000000	1.000000	1.000000	1.000000	1.0000000
2	rowMeans2	0.9082673	0.9597619	1.019397	1.040664	1.146431	0.9598317

Figure: Benchmarking of colMeans2() and rowMeans2() on integer+100x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x10 integer matrix

> X <- data[["1000x10"]]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5243591 280.1    7916910 422.9  7916910 422.9
Vcells 9761813  74.5   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5243579 280.1    7916910 422.9  7916910 422.9
Vcells 9766846  74.6   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.013293	0.0148880	0.0159000	0.0159465	0.0166735	0.022332
3	colMeans	0.016158	0.0181375	0.0198784	0.0192400	0.0205770	0.048743
1	colMeans2	0.021063	0.0237615	0.0250283	0.0249125	0.0260580	0.039862
4	apply+mean	0.140458	0.1559735	0.1652961	0.1653795	0.1731740	0.278474

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colMeans	1.215527	1.218263	1.250215	1.206534	1.234114	2.182653
1	colMeans2	1.584518	1.596017	1.574109	1.562255	1.562839	1.784972
4	apply+mean	10.566313	10.476458	10.395984	10.370896	10.386182	12.469729

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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	0.018759	0.0205295	0.0230194	0.0216080	0.0234490	0.043084
4	apply+mean	0.124769	0.1375675	0.1546701	0.1434855	0.1663905	0.291129
2	.rowMeans	0.156096	0.1666985	0.1835071	0.1722290	0.2058055	0.232899
3	rowMeans	0.157676	0.1692335	0.1852035	0.1749680	0.1966100	0.259823

	expr	min	lq	mean	median	uq	max
1	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	apply+mean	6.651154	6.700967	6.719132	6.640388	7.095846	6.757242
2	.rowMeans	8.321126	8.119949	7.971860	7.970613	8.776728	5.405696
3	rowMeans	8.405352	8.243430	8.045555	8.097371	8.384579	6.030615

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on integer+1000x10 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() on integer+1000x10 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	rowMeans2	18.759	20.5295	23.01936	21.6080	23.449	43.084
1	colMeans2	21.063	23.7615	25.02834	24.9125	26.058	39.862

	expr	min	lq	mean	median	uq	max
2	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colMeans2	1.122821	1.157432	1.087274	1.152929	1.111263	0.9252159

Figure: Benchmarking of colMeans2() and rowMeans2() on integer+1000x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

10x1000 integer matrix

> X <- data[["10x1000"]]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5243807 280.1    7916910 422.9  7916910 422.9
Vcells 9762612  74.5   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5243801 280.1    7916910 422.9  7916910 422.9
Vcells 9767655  74.6   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.012298	0.0140630	0.0160430	0.0157990	0.0168240	0.026693
3	colMeans	0.014148	0.0169780	0.0213489	0.0190300	0.0230850	0.068485
1	colMeans2	0.015985	0.0180765	0.0211984	0.0196635	0.0227085	0.044537
4	apply+mean	3.091406	3.2096035	3.6078381	3.4905205	3.7838090	8.776626

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colMeans	1.150431	1.207281	1.330730	1.204507	1.372147	2.565654
1	colMeans2	1.299805	1.285394	1.321347	1.244604	1.349768	1.668490
4	apply+mean	251.374695	228.230356	224.884938	220.933002	224.905433	328.798786

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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	0.015877	0.0166955	0.0218266	0.019220	0.0250790	0.064051
2	.rowMeans	0.026027	0.0264320	0.0293154	0.028028	0.0307880	0.051111
3	rowMeans	0.027577	0.0289865	0.0340805	0.031922	0.0356905	0.064902
4	apply+mean	3.082995	3.2155610	3.5407100	3.380379	3.5959655	8.124443

	expr	min	lq	mean	median	uq	max
1	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	.rowMeans	1.639290	1.583181	1.343103	1.458273	1.227641	0.7979735
3	rowMeans	1.736915	1.736186	1.561421	1.660874	1.423123	1.0132863
4	apply+mean	194.179946	192.600461	162.219804	175.878200	143.385522	126.8433436

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on integer+10x1000 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() 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	15.877	16.6955	21.82662	19.2200	25.0790	64.051
1	colMeans2	15.985	18.0765	21.19842	19.6635	22.7085	44.537

	expr	min	lq	mean	median	uq	max
2	rowMeans2	1.000000	1.000000	1.0000000	1.000000	1.0000000	1.0000000
1	colMeans2	1.006802	1.082717	0.9712186	1.023075	0.9054787	0.6953365

Figure: Benchmarking of colMeans2() and rowMeans2() on integer+10x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x1000 integer matrix

> X <- data[["100x1000"]]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5244037 280.1    7916910 422.9  7916910 422.9
Vcells 9763204  74.5   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5244031 280.1    7916910 422.9  7916910 422.9
Vcells 9813247  74.9   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.088433	0.0888500	0.0948695	0.0903635	0.0930635	0.171991
3	colMeans	0.090535	0.0911145	0.0972239	0.0930615	0.0978145	0.142057
1	colMeans2	0.138827	0.1393920	0.1472027	0.1408785	0.1454935	0.211462
4	apply+mean	3.670839	3.7336380	4.1002198	3.8298305	3.9675910	16.060589

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colMeans	1.023769	1.025487	1.024817	1.029857	1.051051	0.825956
1	colMeans2	1.569855	1.568846	1.551633	1.559020	1.563379	1.229495
4	apply+mean	41.509832	42.021812	43.219574	42.382494	42.633159	93.380404

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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	0.139095	0.1402800	0.1533820	0.1429255	0.1535730	0.238553
2	.rowMeans	0.221425	0.2220600	0.2383832	0.2241970	0.2349075	0.381985
3	rowMeans	0.223737	0.2252015	0.2397627	0.2293305	0.2426695	0.311189
4	apply+mean	3.674989	3.7546890	4.1612656	3.8256255	4.0240735	19.497721

	expr	min	lq	mean	median	uq	max
1	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowMeans	1.591898	1.582977	1.554180	1.568628	1.529615	1.601258
3	rowMeans	1.608519	1.605371	1.563173	1.604546	1.580157	1.304486
4	apply+mean	26.420713	26.765676	27.130078	26.766571	26.203001	81.733288

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on integer+100x1000 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() 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	138.827	139.392	147.2027	140.8785	145.4935	211.462
2	rowMeans2	139.095	140.280	153.3820	142.9255	153.5730	238.553

	expr	min	lq	mean	median	uq	max
1	colMeans2	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000
2	rowMeans2	1.001931	1.006371	1.041978	1.01453	1.055532	1.128113

Figure: Benchmarking of colMeans2() and rowMeans2() on integer+100x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x100 integer matrix

> X <- data[["1000x100"]]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5244261 280.1    7916910 422.9  7916910 422.9
Vcells 9763984  74.5   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5244255 280.1    7916910 422.9  7916910 422.9
Vcells 9814027  74.9   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.081000	0.0813370	0.0899024	0.0821110	0.0887765	0.148613
3	colMeans	0.083234	0.0838735	0.0939699	0.0853420	0.1025235	0.146897
1	colMeans2	0.132713	0.1333555	0.1479354	0.1344490	0.1597905	0.230693
4	apply+mean	0.893604	0.9013625	1.0113902	0.9101045	1.0796735	1.572386

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
3	colMeans	1.027580	1.031185	1.045244	1.039349	1.154850	0.9884532
1	colMeans2	1.638432	1.639543	1.645511	1.637406	1.799919	1.5523070
4	apply+mean	11.032148	11.081826	11.249871	11.083832	12.161704	10.5804068

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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	0.134541	0.1355810	0.1441573	0.1367815	0.139278	0.225825
2	.rowMeans	0.320036	0.3209695	0.3420746	0.3212865	0.327356	0.541075
3	rowMeans	0.322424	0.3235910	0.3447219	0.3245065	0.345412	0.514245
4	apply+mean	0.893850	0.9032450	0.9761286	0.9099045	0.950947	1.584763

	expr	min	lq	mean	median	uq	max
1	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowMeans	2.378725	2.367363	2.372926	2.348903	2.350378	2.395993
3	rowMeans	2.396474	2.386699	2.391290	2.372444	2.480018	2.277184
4	apply+mean	6.643700	6.662032	6.771274	6.652248	6.827690	7.017660

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on integer+1000x100 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() 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	132.713	133.3555	147.9354	134.4490	159.7905	230.693
2	rowMeans2	134.541	135.5810	144.1573	136.7815	139.2780	225.825

	expr	min	lq	mean	median	uq	max
1	colMeans2	1.000000	1.000000	1.0000000	1.000000	1.0000000	1.0000000
2	rowMeans2	1.013774	1.016689	0.9744613	1.017349	0.8716288	0.9788984

Figure: Benchmarking of colMeans2() and rowMeans2() 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"]]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5244498 280.1    7916910 422.9  7916910 422.9
Vcells 9879920  75.4   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5244477 280.1    7916910 422.9  7916910 422.9
Vcells 9880038  75.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.002776	0.0031885	0.0037724	0.003554	0.0041845	0.013491
1	colMeans2	0.002411	0.0027805	0.0036513	0.003734	0.0041320	0.015417
3	colMeans	0.005321	0.0058290	0.0067177	0.006375	0.0074175	0.018090
4	apply+mean	0.068469	0.0712470	0.0744565	0.073194	0.0750125	0.155593

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.0000000	1.0000000	1.0000000	1.000000	1.0000000	1.000000
1	colMeans2	0.8685159	0.8720401	0.9679089	1.050647	0.9874537	1.142762
3	colMeans	1.9167867	1.8281324	1.7807697	1.793754	1.7726132	1.340894
4	apply+mean	24.6646254	22.3449898	19.7372720	20.594823	17.9262755	11.533096

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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
2	.rowMeans	0.002426	0.0027290	0.0030330	0.0029570	0.0031865	0.007529
1	rowMeans2	0.002355	0.0028005	0.0035996	0.0036665	0.0041455	0.014751
3	rowMeans	0.004684	0.0051880	0.0060911	0.0059685	0.0063850	0.027172
4	apply+mean	0.066863	0.0715830	0.0743118	0.0737275	0.0747935	0.135378

	expr	min	lq	mean	median	uq	max
2	.rowMeans	1.0000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	rowMeans2	0.9707337	1.026200	1.186844	1.239939	1.300957	1.959224
3	rowMeans	1.9307502	1.901063	2.008305	2.018431	2.003766	3.608979
4	apply+mean	27.5610058	26.230487	24.501395	24.933209	23.471991	17.980874

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on double+10x10 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() 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	2.355	2.8005	3.59965	3.6665	4.1455	14.751
1	colMeans2	2.411	2.7805	3.65132	3.7340	4.1320	15.417

	expr	min	lq	mean	median	uq	max
2	rowMeans2	1.000000	1.0000000	1.000000	1.00000	1.0000000	1.000000
1	colMeans2	1.023779	0.9928584	1.014354	1.01841	0.9967435	1.045149

Figure: Benchmarking of colMeans2() and rowMeans2() on double+10x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x100 double matrix

> X <- data[["100x100"]]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5244711 280.1    7916910 422.9  7916910 422.9
Vcells 9880943  75.4   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5244705 280.1    7916910 422.9  7916910 422.9
Vcells 9890986  75.5   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.009332	0.0103345	0.0116088	0.0112220	0.0122990	0.019403
3	colMeans	0.011515	0.0132410	0.0149219	0.0140775	0.0158050	0.039144
1	colMeans2	0.013225	0.0147725	0.0165608	0.0157235	0.0170340	0.034387
4	apply+mean	0.425886	0.4646460	0.5187741	0.4955835	0.5462565	0.796944

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colMeans	1.233926	1.281242	1.285392	1.254455	1.285064	2.017420
1	colMeans2	1.417167	1.429435	1.426566	1.401132	1.384991	1.772252
4	apply+mean	45.637163	44.960666	44.687890	44.161780	44.414709	41.073236

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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	0.013392	0.0159650	0.0186733	0.0179485	0.0203945	0.032255
2	.rowMeans	0.025087	0.0273720	0.0308377	0.0297420	0.0340090	0.049206
3	rowMeans	0.027367	0.0300700	0.0338425	0.0324820	0.0369925	0.053002
4	apply+mean	0.425429	0.4658275	0.5324830	0.5030520	0.5608050	0.929950

	expr	min	lq	mean	median	uq	max
1	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowMeans	1.873283	1.714500	1.651431	1.657074	1.667557	1.525531
3	rowMeans	2.043534	1.883495	1.812346	1.809733	1.813847	1.643218
4	apply+mean	31.767398	29.178046	28.515725	28.027523	27.497855	28.831189

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on double+100x100 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() 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	13.225	14.7725	16.56076	15.7235	17.0340	34.387
2	rowMeans2	13.392	15.9650	18.67331	17.9485	20.3945	32.255

	expr	min	lq	mean	median	uq	max
1	colMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowMeans2	1.012628	1.080724	1.127564	1.141508	1.197282	0.9379998

Figure: Benchmarking of colMeans2() and rowMeans2() on double+100x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x10 double matrix

> X <- data[["1000x10"]]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5244941 280.2    7916910 422.9  7916910 422.9
Vcells 9882068  75.4   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5244929 280.2    7916910 422.9  7916910 422.9
Vcells 9892101  75.5   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.012477	0.0136280	0.0150680	0.0147515	0.0159775	0.023577
3	colMeans	0.015002	0.0168430	0.0191004	0.0181305	0.0195950	0.050398
1	colMeans2	0.016514	0.0184595	0.0206924	0.0205380	0.0215860	0.035745
4	apply+mean	0.151964	0.1621315	0.1802185	0.1756700	0.1934355	0.320478

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colMeans	1.202372	1.235911	1.267612	1.229062	1.226412	2.137592
1	colMeans2	1.323555	1.354527	1.373266	1.392265	1.351025	1.516096
4	apply+mean	12.179530	11.896940	11.960325	11.908620	12.106744	13.592823

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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	0.016585	0.0176235	0.0193872	0.0190780	0.0201155	0.036711
2	.rowMeans	0.029477	0.0310105	0.0337392	0.0329205	0.0356865	0.046924
3	rowMeans	0.031684	0.0339570	0.0370014	0.0355740	0.0390985	0.057199
4	apply+mean	0.146000	0.1541335	0.1709902	0.1683030	0.1799280	0.286069

	expr	min	lq	mean	median	uq	max
1	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowMeans	1.777329	1.759611	1.740279	1.725574	1.774080	1.278200
3	rowMeans	1.910401	1.926802	1.908546	1.864661	1.943700	1.558089
4	apply+mean	8.803135	8.745907	8.819727	8.821837	8.944744	7.792460

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on double+1000x10 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() 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	16.585	17.6235	19.38724	19.078	20.1155	36.711
1	colMeans2	16.514	18.4595	20.69241	20.538	21.5860	35.745

	expr	min	lq	mean	median	uq	max
2	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colMeans2	0.995719	1.047437	1.067321	1.076528	1.073103	0.9736864

Figure: Benchmarking of colMeans2() and rowMeans2() on double+1000x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

10x1000 double matrix

> X <- data[["10x1000"]]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5245157 280.2    7916910 422.9  7916910 422.9
Vcells 9882205  75.4   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5245151 280.2    7916910 422.9  7916910 422.9
Vcells 9892248  75.5   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.009626	0.0101775	0.0120814	0.0115835	0.0129925	0.025867
3	colMeans	0.011458	0.0125005	0.0162083	0.0145280	0.0173725	0.053365
1	colMeans2	0.014976	0.0158040	0.0186646	0.0171290	0.0194755	0.048990
4	apply+mean	3.086088	3.1559520	3.4518498	3.2799350	3.4935680	8.668491

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colMeans	1.190318	1.228249	1.341593	1.254198	1.337118	2.063053
1	colMeans2	1.555786	1.552837	1.544902	1.478741	1.498980	1.893919
4	apply+mean	320.599211	310.091083	285.715807	283.155782	268.891129	335.117756

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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	0.015009	0.0156130	0.0190411	0.0169375	0.0195240	0.057578
2	.rowMeans	0.025013	0.0254050	0.0275033	0.0263310	0.0281965	0.048781
3	rowMeans	0.026902	0.0278405	0.0311949	0.0295030	0.0322195	0.060199
4	apply+mean	3.069896	3.1930860	3.4740875	3.2728830	3.4834760	8.631549

	expr	min	lq	mean	median	uq	max
1	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowMeans	1.666533	1.627170	1.444414	1.554598	1.444197	0.847216
3	rowMeans	1.792391	1.783162	1.638293	1.741874	1.650251	1.045521
4	apply+mean	204.537011	204.514571	182.451844	193.232945	178.420201	149.910539

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on double+10x1000 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() 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	15.009	15.613	19.04112	16.9375	19.5240	57.578
1	colMeans2	14.976	15.804	18.66459	17.1290	19.4755	48.990

	expr	min	lq	mean	median	uq	max
2	rowMeans2	1.0000000	1.000000	1.0000000	1.000000	1.0000000	1.0000000
1	colMeans2	0.9978013	1.012233	0.9802254	1.011306	0.9975159	0.8508458

Figure: Benchmarking of colMeans2() and rowMeans2() on double+10x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x1000 double matrix

> X <- data[["100x1000"]]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5245387 280.2    7916910 422.9  7916910 422.9
Vcells 9883526  75.5   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5245381 280.2    7916910 422.9  7916910 422.9
Vcells 9983569  76.2   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.063167	0.0638000	0.0737637	0.0674975	0.0780270	0.159803
3	colMeans	0.065253	0.0660820	0.0755566	0.0693515	0.0817260	0.142163
1	colMeans2	0.104201	0.1049505	0.1161812	0.1094545	0.1198685	0.179286
4	apply+mean	3.785904	3.8621365	4.3237555	4.0004650	4.1720880	18.349369

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
3	colMeans	1.033024	1.035768	1.024306	1.027468	1.047407	0.8896141
1	colMeans2	1.649611	1.644992	1.575044	1.621608	1.536244	1.1219189
4	apply+mean	59.934839	60.535055	58.616272	59.268343	53.469799	114.8249345

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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	0.109627	0.1129705	0.1305515	0.1195635	0.1331590	0.369935
2	.rowMeans	0.210915	0.2115055	0.2348538	0.2157175	0.2437235	0.409699
3	rowMeans	0.213266	0.2142505	0.2363136	0.2187875	0.2484610	0.382752
4	apply+mean	3.804420	3.8713845	4.3908263	3.9796310	4.1606125	20.476403

	expr	min	lq	mean	median	uq	max
1	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowMeans	1.923933	1.872219	1.798936	1.804209	1.830319	1.107489
3	rowMeans	1.945378	1.896517	1.810118	1.829885	1.865897	1.034647
4	apply+mean	34.703312	34.268986	33.632908	33.284665	31.245447	55.351354

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on double+100x1000 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() 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	104.201	104.9505	116.1812	109.4545	119.8685	179.286
2	rowMeans2	109.627	112.9705	130.5515	119.5635	133.1590	369.935

	expr	min	lq	mean	median	uq	max
1	colMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2	1.052072	1.076417	1.123689	1.092358	1.110876	2.063379

Figure: Benchmarking of colMeans2() and rowMeans2() on double+100x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x100 double matrix

> X <- data[["1000x100"]]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5245611 280.2    7916910 422.9  7916910 422.9
Vcells 9883675  75.5   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMeans2 = colMeans2(X, na.rm = FALSE), .colMeans = .colMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colMeans = colMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 2L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5245605 280.2    7916910 422.9  7916910 422.9
Vcells 9983718  76.2   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMeans2 = rowMeans2(X, na.rm = FALSE), .rowMeans = .rowMeans(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowMeans = rowMeans(X, na.rm = FALSE), `apply+mean` = apply(X, MARGIN = 1L, 
+     FUN = mean, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() 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
2	.colMeans	0.078013	0.078512	0.0923077	0.080634	0.1019995	0.159646
3	colMeans	0.080225	0.081468	0.0958915	0.089389	0.1062610	0.151461
1	colMeans2	0.106241	0.108208	0.1251798	0.118682	0.1379010	0.185638
4	apply+mean	0.999177	1.033953	1.2982486	1.196687	1.3312755	8.843105

	expr	min	lq	mean	median	uq	max
2	.colMeans	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
3	colMeans	1.028354	1.037650	1.038824	1.108577	1.041780	0.9487303
1	colMeans2	1.361837	1.378235	1.356114	1.471861	1.351977	1.1628102
4	apply+mean	12.807827	13.169356	14.064354	14.840979	13.051785	55.3919610

Table: Benchmarking of rowMeans2(), .rowMeans(), rowMeans() and apply+mean() 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	0.110857	0.1143645	0.1347093	0.1257395	0.1430745	0.249168
2	.rowMeans	0.214976	0.2158880	0.2522388	0.2332150	0.2794755	0.396863
3	rowMeans	0.217536	0.2199010	0.2625061	0.2465075	0.2836185	0.410389
4	apply+mean	1.028700	1.0431015	1.3299909	1.1217340	1.3888410	8.272883

	expr	min	lq	mean	median	uq	max
1	rowMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowMeans	1.939219	1.887719	1.872468	1.854747	1.953356	1.592753
3	rowMeans	1.962312	1.922808	1.948686	1.960462	1.982313	1.647037
4	apply+mean	9.279522	9.120850	9.873043	8.921095	9.707118	33.202028

Figure: Benchmarking of colMeans2(), .colMeans(), colMeans() and apply+mean() on double+1000x100 data as well as rowMeans2(), .rowMeans(), rowMeans() and apply+mean() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMeans2() and rowMeans2() 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	106.241	108.2080	125.1798	118.6820	137.9010	185.638
2	rowMeans2	110.857	114.3645	134.7093	125.7395	143.0745	249.168

	expr	min	lq	mean	median	uq	max
1	colMeans2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2	1.043448	1.056895	1.076126	1.059466	1.037516	1.342225

Figure: Benchmarking of colMeans2() and rowMeans2() 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 28 secs.

Reproducibility

To reproduce this report, do:

html <- matrixStats:::benchmark('colMeans2')

---

Copyright Henrik Bengtsson. Last updated on 2021-08-25 22:19:37 (+0200 UTC). Powered by RSP.