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  5239254 279.9    8529671 455.6  8529671 455.6
Vcells 10067968  76.9   31876688 243.2 60562128 462.1
> 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  5231316 279.4    8529671 455.6  8529671 455.6
Vcells 10042070  76.7   31876688 243.2 60562128 462.1
> 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	0.001847	0.0019260	0.0029480	0.0019790	0.0020720	0.090764
2	colMeans2(X, rows, cols)	0.002228	0.0023255	0.0024960	0.0023960	0.0024840	0.004669
3	colMeans2(X[rows, cols])	0.002640	0.0028950	0.0032197	0.0030255	0.0031415	0.007819

	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)	1.206281	1.207425	0.8466872	1.210713	1.198842	0.0514411
3	colMeans2(X[rows, cols])	1.429345	1.503115	1.0921722	1.528802	1.516168	0.0861465

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	0.001876	0.0019715	0.0020668	0.0020075	0.0020920	0.004386
2	rowMeans2(X, cols, rows)	0.002286	0.0023715	0.0033519	0.0024455	0.0025430	0.090527
3	rowMeans2(X[cols, rows])	0.002684	0.0029440	0.0031055	0.0030285	0.0031615	0.005362

	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)	1.218550	1.202891	1.621821	1.218182	1.215583	20.639991
3	rowMeans2(X[cols, rows])	1.430704	1.493279	1.502610	1.508593	1.511233	1.222526

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	1.847	1.9260	2.94796	1.9790	2.072	90.764
2	rowMeans2_X_S	1.876	1.9715	2.06677	2.0075	2.092	4.386

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
2	rowMeans2_X_S	1.015701	1.023624	0.7010848	1.014401	1.009653	0.0483231

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 5229718 279.3    8529671 455.6  8529671 455.6
Vcells 9709517  74.1   31876688 243.2 60562128 462.1
> 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 5229694 279.3    8529671 455.6  8529671 455.6
Vcells 9714570  74.2   31876688 243.2 60562128 462.1
> 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	0.011503	0.0117600	0.0119683	0.0118845	0.0120145	0.014024
2	colMeans2(X, rows, cols)	0.013068	0.0136855	0.0138894	0.0138315	0.0140065	0.017305
3	colMeans2(X[rows, cols])	0.022900	0.0232510	0.0245251	0.0234165	0.0239040	0.058453

	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)	1.136051	1.163733	1.160514	1.163827	1.165800	1.233956
3	colMeans2(X[rows, cols])	1.990785	1.977126	2.049174	1.970339	1.989596	4.168069

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
2	rowMeans2(X, cols, rows)	0.012217	0.0128180	0.0137361	0.0130705	0.0132990	0.065091
1	rowMeans2_X_S	0.012601	0.0131290	0.0135982	0.0133280	0.0135975	0.025668
3	rowMeans2(X[cols, rows])	0.023290	0.0243555	0.0250001	0.0246105	0.0250065	0.042236

	expr	min	lq	mean	median	uq	max
2	rowMeans2(X, cols, rows)	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
1	rowMeans2_X_S	1.031432	1.024263	0.9899586	1.019701	1.022445	0.3943402
3	rowMeans2(X[cols, rows])	1.906360	1.900101	1.8200274	1.882904	1.880329	0.6488762

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	11.503	11.760	11.96829	11.8845	12.0145	14.024
2	rowMeans2_X_S	12.601	13.129	13.59816	13.3280	13.5975	25.668

	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	1.095453	1.116412	1.136182	1.121461	1.131757	1.830291

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 5230460 279.4    8529671 455.6  8529671 455.6
Vcells 9713555  74.2   31876688 243.2 60562128 462.1
> 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 5230436 279.4    8529671 455.6  8529671 455.6
Vcells 9718608  74.2   31876688 243.2 60562128 462.1
> 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	0.010211	0.0103110	0.0105182	0.0104000	0.010501	0.018538
2	colMeans2(X, rows, cols)	0.014302	0.0147365	0.0152329	0.0151135	0.015476	0.020525
3	colMeans2(X[rows, cols])	0.022029	0.0223415	0.0231530	0.0224825	0.022771	0.073242

	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)	1.400646	1.429202	1.448240	1.453221	1.473764	1.107185
3	colMeans2(X[rows, cols])	2.157379	2.166764	2.201224	2.161779	2.168460	3.950912

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	0.011739	0.011967	0.0124367	0.0122410	0.0124825	0.026305
2	rowMeans2(X, cols, rows)	0.014218	0.014746	0.0151680	0.0149640	0.0154550	0.020347
3	rowMeans2(X[cols, rows])	0.024703	0.025286	0.0263237	0.0259705	0.0263695	0.064014

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowMeans2(X, cols, rows)	1.211176	1.232222	1.219617	1.222449	1.238133	0.7735031
3	rowMeans2(X[cols, rows])	2.104353	2.112977	2.116609	2.121599	2.112518	2.4335297

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	10.211	10.311	10.51822	10.400	10.5010	18.538
2	rowMeans2_X_S	11.739	11.967	12.43672	12.241	12.4825	26.305

	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	1.149642	1.160605	1.182398	1.177019	1.188696	1.418977

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 5230664 279.4    8529671 455.6  8529671 455.6
Vcells 9714428  74.2   31876688 243.2 60562128 462.1
> 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 5230640 279.4    8529671 455.6  8529671 455.6
Vcells 9719481  74.2   31876688 243.2 60562128 462.1
> 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	0.015216	0.0154655	0.0162802	0.0158905	0.0161560	0.034637
2	colMeans2(X, rows, cols)	0.016921	0.0173415	0.0181420	0.0176450	0.0182960	0.033459
3	colMeans2(X[rows, cols])	0.028564	0.0289960	0.0300537	0.0299655	0.0305085	0.042075

	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)	1.112053	1.121302	1.114357	1.110412	1.132459	0.9659901
3	colMeans2(X[rows, cols])	1.877234	1.874883	1.846021	1.885749	1.888370	1.2147415

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	0.012894	0.0132290	0.0135355	0.0134410	0.0136680	0.018662
2	rowMeans2(X, cols, rows)	0.015274	0.0157130	0.0164131	0.0159680	0.0162695	0.039937
3	rowMeans2(X[cols, rows])	0.023609	0.0239345	0.0248161	0.0242625	0.0249900	0.039256

	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)	1.184582	1.187769	1.212594	1.188007	1.190335	2.140017
3	rowMeans2(X[cols, rows])	1.831007	1.809245	1.833409	1.805111	1.828358	2.103526

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	12.894	13.2290	13.53550	13.4410	13.668	18.662
1	colMeans2_X_S	15.216	15.4655	16.28025	15.8905	16.156	34.637

	expr	min	lq	mean	median	uq	max
2	rowMeans2_X_S	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
1	colMeans2_X_S	1.180084	1.16906	1.202782	1.182241	1.182031	1.856018

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 5230877 279.4    8529671 455.6  8529671 455.6
Vcells 9737105  74.3   31876688 243.2 60562128 462.1
> 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 5230853 279.4    8529671 455.6  8529671 455.6
Vcells 9787158  74.7   31876688 243.2 60562128 462.1
> 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	0.072860	0.076004	0.0825173	0.0791475	0.0871470	0.120449
2	colMeans2(X, rows, cols)	0.082647	0.085856	0.0943815	0.0900305	0.1017870	0.172689
3	colMeans2(X[rows, cols])	0.153570	0.161455	0.1768899	0.1698555	0.1904865	0.219053

	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)	1.134326	1.129625	1.143778	1.137503	1.167992	1.433711
3	colMeans2(X[rows, cols])	2.107741	2.124296	2.143670	2.146063	2.185807	1.818637

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	0.081177	0.089784	0.0992917	0.0962035	0.1078760	0.162566
2	rowMeans2(X, cols, rows)	0.100575	0.107135	0.1175959	0.1116880	0.1270455	0.197445
3	rowMeans2(X[cols, rows])	0.152274	0.168322	0.1830259	0.1753465	0.1962145	0.230493

	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)	1.238959	1.193253	1.184348	1.160956	1.177699	1.214553
3	rowMeans2(X[cols, rows])	1.875827	1.874744	1.843315	1.822662	1.818889	1.417843

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	72.860	76.004	82.51728	79.1475	87.147	120.449
2	rowMeans2_X_S	81.177	89.784	99.29170	96.2035	107.876	162.566

	expr	min	lq	mean	median	uq	max
1	colMeans2_X_S	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMeans2_X_S	1.11415	1.181306	1.203284	1.215496	1.237863	1.349667

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 5231087 279.4    8529671 455.6  8529671 455.6
Vcells 9737889  74.3   31876688 243.2 60562128 462.1
> 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 5231063 279.4    8529671 455.6  8529671 455.6
Vcells 9787942  74.7   31876688 243.2 60562128 462.1
> 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	0.064757	0.067414	0.0741388	0.0698780	0.0782725	0.113074
2	colMeans2(X, rows, cols)	0.073908	0.077577	0.0856458	0.0837435	0.0919200	0.103769
3	colMeans2(X[rows, cols])	0.142508	0.151349	0.1670202	0.1629915	0.1827705	0.242255

	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)	1.141313	1.150755	1.155210	1.198424	1.174359	0.9177088
3	colMeans2(X[rows, cols])	2.200658	2.245068	2.252804	2.332515	2.335054	2.1424465

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	0.073403	0.0818965	0.0904302	0.0881815	0.0957480	0.122820
2	rowMeans2(X, cols, rows)	0.095182	0.1109920	0.1188638	0.1152325	0.1239635	0.215424
3	rowMeans2(X[cols, rows])	0.144924	0.1677190	0.1825400	0.1752200	0.1982145	0.253458

	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)	1.296704	1.355272	1.314426	1.306765	1.294685	1.753981
3	rowMeans2(X[cols, rows])	1.974361	2.047939	2.018574	1.987038	2.070169	2.063654

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	64.757	67.4140	74.13878	69.8780	78.2725	113.074
2	rowMeans2_X_S	73.403	81.8965	90.43019	88.1815	95.7480	122.820

	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	1.133514	1.214829	1.219742	1.261935	1.223265	1.086191

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 5231305 279.4    8529671 455.6  8529671 455.6
Vcells 9828998  75.0   31876688 243.2 60562128 462.1
> 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 5231272 279.4    8529671 455.6  8529671 455.6
Vcells 9829136  75.0   31876688 243.2 60562128 462.1
> 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	0.001937	0.0020190	0.0022909	0.0020575	0.0021515	0.019771
2	colMeans2(X, rows, cols)	0.002324	0.0023885	0.0025168	0.0024235	0.0025670	0.004859
3	colMeans2(X[rows, cols])	0.002773	0.0030760	0.0032504	0.0031420	0.0032655	0.008533

	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)	1.199793	1.183011	1.098594	1.177886	1.193121	0.2457640
3	colMeans2(X[rows, cols])	1.431595	1.523527	1.418821	1.527096	1.517778	0.4315917

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	0.001887	0.0019690	0.0020623	0.0020050	0.0020795	0.004996
2	rowMeans2(X, cols, rows)	0.002290	0.0023945	0.0027032	0.0024895	0.0026095	0.020400
3	rowMeans2(X[cols, rows])	0.002780	0.0030365	0.0031873	0.0030870	0.0032415	0.006183

	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)	1.213566	1.216100	1.310781	1.241646	1.254869	4.083267
3	rowMeans2(X[cols, rows])	1.473238	1.542153	1.545530	1.539651	1.558788	1.237590

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	1.887	1.969	2.06229	2.0050	2.0795	4.996
1	colMeans2_X_S	1.937	2.019	2.29091	2.0575	2.1515	19.771

	expr	min	lq	mean	median	uq	max
2	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colMeans2_X_S	1.026497	1.025394	1.110857	1.026185	1.034624	3.957366

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 5231503 279.4    8529671 455.6  8529671 455.6
Vcells 9834953  75.1   31876688 243.2 60562128 462.1
> 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 5231479 279.4    8529671 455.6  8529671 455.6
Vcells 9845006  75.2   31876688 243.2 60562128 462.1
> 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
2	colMeans2(X, rows, cols)	0.016364	0.0173960	0.0179279	0.0178595	0.0182525	0.026306
1	colMeans2_X_S	0.020819	0.0218145	0.0224671	0.0225680	0.0227660	0.037107
3	colMeans2(X[rows, cols])	0.035665	0.0374555	0.0390382	0.0388450	0.0391190	0.067728

	expr	min	lq	mean	median	uq	max
2	colMeans2(X, rows, cols)	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colMeans2_X_S	1.272244	1.253995	1.253189	1.263641	1.247281	1.410591
3	colMeans2(X[rows, cols])	2.179479	2.153110	2.177510	2.175033	2.143213	2.574622

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
2	rowMeans2(X, cols, rows)	0.016692	0.0173425	0.0180302	0.0178205	0.0181035	0.042894
1	rowMeans2_X_S	0.016760	0.0176815	0.0181757	0.0182380	0.0185115	0.024771
3	rowMeans2(X[cols, rows])	0.031850	0.0333835	0.0345505	0.0346645	0.0349370	0.049181

	expr	min	lq	mean	median	uq	max
2	rowMeans2(X, cols, rows)	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	rowMeans2_X_S	1.004074	1.019547	1.008068	1.023428	1.022537	0.5774934
3	rowMeans2(X[cols, rows])	1.908100	1.924953	1.916252	1.945204	1.929848	1.1465706

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
2	rowMeans2_X_S	16.760	17.6815	18.17570	18.238	18.5115	24.771
1	colMeans2_X_S	20.819	21.8145	22.46708	22.568	22.7660	37.107

	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	1.242184	1.233747	1.236105	1.237416	1.22983	1.498002

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 5231702 279.5    8529671 455.6  8529671 455.6
Vcells 9836359  75.1   31876688 243.2 60562128 462.1
> 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 5231678 279.5    8529671 455.6  8529671 455.6
Vcells 9846412  75.2   31876688 243.2 60562128 462.1
> 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
2	colMeans2(X, rows, cols)	0.017148	0.0183160	0.0186735	0.0187245	0.018959	0.022967
1	colMeans2_X_S	0.019208	0.0200920	0.0206080	0.0207725	0.020959	0.023622
3	colMeans2(X[rows, cols])	0.034766	0.0363295	0.0375867	0.0367140	0.037854	0.067148

	expr	min	lq	mean	median	uq	max
2	colMeans2(X, rows, cols)	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colMeans2_X_S	1.120131	1.096964	1.103597	1.109375	1.105491	1.028519
3	colMeans2(X[rows, cols])	2.027408	1.983484	2.012834	1.960747	1.996624	2.923673

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	0.015285	0.0160695	0.0167367	0.0165120	0.0168140	0.031157
2	rowMeans2(X, cols, rows)	0.017499	0.0184060	0.0188089	0.0188765	0.0190935	0.022353
3	rowMeans2(X[cols, rows])	0.033629	0.0352120	0.0362132	0.0364040	0.0367485	0.071494

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowMeans2(X, cols, rows)	1.144848	1.145400	1.123811	1.143199	1.135571	0.7174311
3	rowMeans2(X[cols, rows])	2.200131	2.191232	2.163702	2.204700	2.185589	2.2946368

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	15.285	16.0695	16.73669	16.5120	16.814	31.157
1	colMeans2_X_S	19.208	20.0920	20.60805	20.7725	20.959	23.622

	expr	min	lq	mean	median	uq	max
2	rowMeans2_X_S	1.000000	1.000000	1.00000	1.000000	1.000000	1.0000000
1	colMeans2_X_S	1.256657	1.250319	1.23131	1.258025	1.246521	0.7581603

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 5231906 279.5    8529671 455.6  8529671 455.6
Vcells 9836494  75.1   31876688 243.2 60562128 462.1
> 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 5231882 279.5    8529671 455.6  8529671 455.6
Vcells 9846547  75.2   31876688 243.2 60562128 462.1
> 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
2	colMeans2(X, rows, cols)	0.023945	0.024541	0.0255356	0.0253150	0.0257285	0.046420
1	colMeans2_X_S	0.025012	0.025952	0.0266161	0.0265230	0.0269200	0.042034
3	colMeans2(X[rows, cols])	0.044219	0.045548	0.0463394	0.0463295	0.0469120	0.052864

	expr	min	lq	mean	median	uq	max
2	colMeans2(X, rows, cols)	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colMeans2_X_S	1.044561	1.057496	1.042313	1.047719	1.046310	0.9055149
3	colMeans2(X[rows, cols])	1.846690	1.855996	1.814699	1.830121	1.823348	1.1388195

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	0.020393	0.021392	0.0219239	0.0219955	0.022301	0.027068
2	rowMeans2(X, cols, rows)	0.022744	0.023988	0.0249798	0.0248580	0.025115	0.051117
3	rowMeans2(X[cols, rows])	0.035782	0.037693	0.0389474	0.0390735	0.039362	0.053669

	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)	1.115285	1.121354	1.139387	1.130140	1.126183	1.888466
3	rowMeans2(X[cols, rows])	1.754622	1.762014	1.776482	1.776431	1.765033	1.982747

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	20.393	21.392	21.92387	21.9955	22.301	27.068
1	colMeans2_X_S	25.012	25.952	26.61607	26.5230	26.920	42.034

	expr	min	lq	mean	median	uq	max
2	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colMeans2_X_S	1.226499	1.213164	1.214022	1.205838	1.207121	1.552904

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 5232119 279.5    8529671 455.6  8529671 455.6
Vcells 9881967  75.4   31876688 243.2 60562128 462.1
> 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 5232095 279.5    8529671 455.6  8529671 455.6
Vcells 9982020  76.2   31876688 243.2 60562128 462.1
> 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
2	colMeans2(X, rows, cols)	0.093012	0.1019725	0.1165956	0.1137045	0.1283545	0.239620
1	colMeans2_X_S	0.124963	0.1393400	0.1593735	0.1571800	0.1755740	0.226523
3	colMeans2(X[rows, cols])	0.224138	0.2515870	0.2842428	0.2785885	0.3118385	0.374830

	expr	min	lq	mean	median	uq	max
2	colMeans2(X, rows, cols)	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colMeans2_X_S	1.343515	1.366447	1.366892	1.382355	1.367884	0.9453426
3	colMeans2(X[rows, cols])	2.409775	2.467204	2.437852	2.450110	2.429510	1.5642684

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	0.107334	0.124809	0.1383345	0.1290975	0.1491210	0.313997
2	rowMeans2(X, cols, rows)	0.112252	0.124729	0.1390577	0.1324275	0.1445240	0.267007
3	rowMeans2(X[cols, rows])	0.216403	0.244320	0.2705480	0.2606555	0.2951735	0.447450

	expr	min	lq	mean	median	uq	max
1	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.0000000	1.0000000
2	rowMeans2(X, cols, rows)	1.045820	0.999359	1.005228	1.025794	0.9691727	0.8503489
3	rowMeans2(X[cols, rows])	2.016164	1.957551	1.955752	2.019059	1.9794228	1.4250136

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
2	rowMeans2_X_S	107.334	124.809	138.3345	129.0975	149.121	313.997
1	colMeans2_X_S	124.963	139.340	159.3735	157.1800	175.574	226.523

	expr	min	lq	mean	median	uq	max
2	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colMeans2_X_S	1.164244	1.116426	1.152088	1.217529	1.177393	0.7214177

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 5232329 279.5    8529671 455.6  8529671 455.6
Vcells 9882108  75.4   31876688 243.2 60562128 462.1
> 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 5232305 279.5    8529671 455.6  8529671 455.6
Vcells 9982161  76.2   31876688 243.2 60562128 462.1
> 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
2	colMeans2(X, rows, cols)	0.087341	0.0955840	0.1081725	0.1041995	0.1177120	0.168556
1	colMeans2_X_S	0.122635	0.1336755	0.1506216	0.1463645	0.1580090	0.250562
3	colMeans2(X[rows, cols])	0.219466	0.2412715	0.2736795	0.2705760	0.2947245	0.434267

	expr	min	lq	mean	median	uq	max
2	colMeans2(X, rows, cols)	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colMeans2_X_S	1.404094	1.398513	1.392421	1.404656	1.342335	1.486521
3	colMeans2(X[rows, cols])	2.512749	2.524183	2.530029	2.596711	2.503776	2.576396

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	0.098873	0.1072580	0.1179329	0.1155910	0.1253360	0.194452
2	rowMeans2(X, cols, rows)	0.106817	0.1176845	0.1293736	0.1248245	0.1341950	0.304632
3	rowMeans2(X[cols, rows])	0.202904	0.2246740	0.2468780	0.2393140	0.2738205	0.323391

	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)	1.080345	1.097209	1.097011	1.079881	1.070682	1.566618
3	rowMeans2(X[cols, rows])	2.052168	2.094706	2.093377	2.070352	2.184692	1.663089

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
2	rowMeans2_X_S	98.873	107.2580	117.9329	115.5910	125.336	194.452
1	colMeans2_X_S	122.635	133.6755	150.6216	146.3645	158.009	250.562

	expr	min	lq	mean	median	uq	max
2	rowMeans2_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colMeans2_X_S	1.240328	1.246299	1.277181	1.266228	1.260683	1.288555

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_0.60.1     ggplot2_3.3.5         
[4] knitr_1.33             R.devices_2.17.0       R.utils_2.10.1        
[7] R.oo_1.24.0            R.methodsS3_1.8.1-9001 history_0.0.1-9000    

loaded via a namespace (and not attached):
 [1] Biobase_2.52.0          httr_1.4.2              splines_4.1.1          
 [4] bit64_4.0.5             network_1.17.1          assertthat_0.2.1       
 [7] highr_0.9               stats4_4.1.1            blob_1.2.2             
[10] GenomeInfoDbData_1.2.6  robustbase_0.93-8       pillar_1.6.2           
[13] RSQLite_2.2.8           lattice_0.20-44         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] rappdirs_0.3.3          startup_0.15.0          labeling_0.4.2         
[67] bitops_1.0-7            base64enc_0.1-3         boot_1.3-28            
[70] gtable_0.3.0            DBI_1.1.1               markdown_1.1           
[73] R6_2.5.1                lpSolveAPI_5.5.2.0-17.7 rle_0.9.2              
[76] dplyr_1.0.7             fastmap_1.1.0           bit_4.0.4              
[79] utf8_1.2.2              parallel_4.1.1          Rcpp_1.0.7             
[82] vctrs_0.3.8             png_0.1-7               DEoptimR_1.0-9         
[85] tidyselect_1.1.1        xfun_0.25               coda_0.19-4            

Total processing time was 23.28 secs.

Reproducibility

To reproduce this report, do:

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