matrixStats.benchmarks

---

colMins() and rowMins() benchmarks

This report benchmark the performance of colMins() and rowMins() against alternative methods.

Alternative methods

• apply() + min()
• lapply() + pmin()
• lapply() + pmin.int()

See also StackOverflow:colMins?.

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  5260520 281.0    7916910 422.9  7916910 422.9
Vcells 10252251  78.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5260328 281.0    7916910 422.9  7916910 422.9
Vcells 10252337  78.3   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	colMins	0.002378	0.0028950	0.0035735	0.0034215	0.0038475	0.016397
4	lapply+pmin.int	0.018538	0.0201950	0.0217514	0.0208230	0.0221110	0.066468
3	lapply+pmin	0.030322	0.0325930	0.0340012	0.0334305	0.0347640	0.054249
2	apply+min	0.031949	0.0346265	0.0376096	0.0359045	0.0387025	0.099151

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	7.795627	6.97582	6.086870	6.085927	5.746849	4.053668
3	lapply+pmin	12.751051	11.25838	9.514829	9.770715	9.035478	3.308471
2	apply+min	13.435240	11.96079	10.524589	10.493789	10.059129	6.046899

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.002335	0.0028945	0.0035748	0.0034300	0.0039830	0.015571
4	lapply+pmin.int	0.018434	0.0191985	0.0203179	0.0198125	0.0209715	0.035603
3	lapply+pmin	0.029785	0.0314305	0.0337356	0.0329995	0.0338185	0.084097
2	apply+min	0.032904	0.0349735	0.0374044	0.0362240	0.0378260	0.092718

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	7.894647	6.632752	5.683564	5.776239	5.265252	2.286494
3	lapply+pmin	12.755889	10.858697	9.436919	9.620845	8.490711	5.400873
2	apply+min	14.091649	12.082743	10.463203	10.560933	9.496862	5.954531

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on integer+10x10 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMins() and rowMins() 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	colMins	2.378	2.8950	3.57350	3.4215	3.8475	16.397
2	rowMins	2.335	2.8945	3.57485	3.4300	3.9830	15.571

	expr	min	lq	mean	median	uq	max
1	colMins	1.0000000	1.0000000	1.000000	1.000000	1.000000	1.0000000
2	rowMins	0.9819176	0.9998273	1.000378	1.002484	1.035218	0.9496249

Figure: Benchmarking of colMins() and rowMins() 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 5258907 280.9    7916910 422.9  7916910 422.9
Vcells 9869084  75.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5258901 280.9    7916910 422.9  7916910 422.9
Vcells 9874191  75.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	colMins	0.019092	0.0208085	0.0250486	0.0231260	0.0260610	0.055158
4	lapply+pmin.int	0.162959	0.1747140	0.1993368	0.1883025	0.2079535	0.380376
2	apply+min	0.163356	0.1783380	0.2085342	0.2009140	0.2189795	0.382003
3	lapply+pmin	0.216376	0.2347825	0.2672773	0.2579365	0.2734245	0.525389

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	8.535460	8.39628	7.957991	8.142459	7.979490	6.896117
2	apply+min	8.556254	8.57044	8.325172	8.687797	8.402575	6.925614
3	lapply+pmin	11.333333	11.28301	10.670336	11.153529	10.491712	9.525164

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.021189	0.0241980	0.0273833	0.0275665	0.0292965	0.045254
4	lapply+pmin.int	0.140113	0.1514660	0.1689246	0.1665185	0.1780995	0.278341
2	apply+min	0.167289	0.1788965	0.2029122	0.1932630	0.2155980	0.359967
3	lapply+pmin	0.192182	0.2099030	0.2338272	0.2327380	0.2460980	0.327113

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	6.612535	6.259443	6.168889	6.040611	6.079207	6.150639
2	apply+min	7.895087	7.393028	7.410069	7.010792	7.359173	7.954369
3	lapply+pmin	9.069895	8.674395	8.539039	8.442784	8.400253	7.228378

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on integer+100x100 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMins() and rowMins() 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	colMins	19.092	20.8085	25.04863	23.1260	26.0610	55.158
2	rowMins	21.189	24.1980	27.38331	27.5665	29.2965	45.254

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.00000	1.000000	1.000000	1.000000	1.0000000
2	rowMins	1.109837	1.16289	1.093206	1.192013	1.124151	0.8204431

Figure: Benchmarking of colMins() and rowMins() 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 5259677 280.9    7916910 422.9  7916910 422.9
Vcells 9872876  75.4   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5259665 280.9    7916910 422.9  7916910 422.9
Vcells 9877973  75.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	colMins	0.015204	0.017997	0.0211080	0.019834	0.0228425	0.051227
2	apply+min	0.083499	0.095359	0.1079905	0.105140	0.1160725	0.178032
4	lapply+pmin.int	1.013110	1.084493	1.1475571	1.130878	1.1718020	1.857980
3	lapply+pmin	1.499132	1.592197	1.7595750	1.688099	1.7522765	7.066660

	expr	min	lq	mean	median	uq	max
1	colMins	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
2	apply+min	5.49191	5.298605	5.116095	5.300998	5.081427	3.475355
4	lapply+pmin.int	66.63444	60.259682	54.366005	57.017167	51.299201	36.269545
3	lapply+pmin	98.60116	88.470134	83.360612	85.111374	76.711240	137.947957

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.018456	0.021255	0.0262313	0.0244300	0.028759	0.066980
2	apply+min	0.085855	0.099826	0.1168604	0.1114285	0.127582	0.267035
4	lapply+pmin.int	0.968886	1.052808	1.1880199	1.1138560	1.168856	6.644591
3	lapply+pmin	1.483855	1.588378	1.6993377	1.6673565	1.782304	2.461269

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	apply+min	4.651875	4.696589	4.454992	4.561134	4.436246	3.986787
4	lapply+pmin.int	52.497074	49.532251	45.290098	45.593778	40.643138	99.202613
3	lapply+pmin	80.399599	74.729617	64.782725	68.250368	61.973782	36.746327

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on integer+1000x10 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMins() and rowMins() 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	colMins	15.204	17.997	21.10799	19.834	22.8425	51.227
2	rowMins	18.456	21.255	26.23134	24.430	28.7590	66.980

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
2	rowMins	1.213891	1.18103	1.242721	1.231723	1.259013	1.307514

Figure: Benchmarking of colMins() and rowMins() 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 5259893 281.0    7916910 422.9  7916910 422.9
Vcells 9873763  75.4   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5259887 281.0    7916910 422.9  7916910 422.9
Vcells 9878870  75.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	colMins	0.033876	0.0401855	0.0473595	0.0455700	0.0518715	0.073317
4	lapply+pmin.int	0.088449	0.1019290	0.1224060	0.1136535	0.1308815	0.247705
3	lapply+pmin	0.100228	0.1143980	0.1288952	0.1229210	0.1378105	0.240051
2	apply+min	0.966494	1.0888590	1.2126394	1.1751415	1.2805590	1.900782

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	2.610963	2.536462	2.584612	2.494042	2.523187	3.378548
3	lapply+pmin	2.958673	2.846748	2.721633	2.697411	2.656767	3.274152
2	apply+min	28.530346	27.095818	25.604976	25.787613	24.687140	25.925529

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.034944	0.0400575	0.0449176	0.0436105	0.0479255	0.084419
4	lapply+pmin.int	0.067814	0.0753215	0.0857738	0.0837045	0.0914050	0.124539
3	lapply+pmin	0.074270	0.0863680	0.0975050	0.0936825	0.1012875	0.194381
2	apply+min	0.949017	1.0237465	1.1477589	1.1101285	1.2040955	1.927425

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	1.940648	1.880335	1.909584	1.919366	1.907231	1.475248
3	lapply+pmin	2.125401	2.156101	2.170755	2.148164	2.113437	2.302574
2	apply+min	27.158225	25.556924	25.552571	25.455533	25.124318	22.831649

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on integer+10x1000 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMins() and rowMins() 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	rowMins	34.944	40.0575	44.91755	43.6105	47.9255	84.419
1	colMins	33.876	40.1855	47.35952	45.5700	51.8715	73.317

	expr	min	lq	mean	median	uq	max
2	rowMins	1.0000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colMins	0.9694368	1.003195	1.054366	1.044932	1.082336	0.8684893

Figure: Benchmarking of colMins() and rowMins() 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 5260135 281.0    7916910 422.9  7916910 422.9
Vcells 9874394  75.4   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5260117 281.0    7916910 422.9  7916910 422.9
Vcells 9924481  75.8   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	colMins	0.173151	0.1837955	0.2023116	0.1893960	0.2087765	0.354524
4	lapply+pmin.int	0.704513	0.7437555	0.8466216	0.7748180	0.8585805	1.739525
3	lapply+pmin	0.746105	0.7785445	0.8749611	0.8182535	0.8979790	1.474941
2	apply+min	1.497374	1.5675940	1.9411797	1.6268280	1.8791445	17.475682

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	4.068778	4.046647	4.184741	4.090994	4.112438	4.906650
3	lapply+pmin	4.308985	4.235928	4.324819	4.320331	4.301150	4.160342
2	apply+min	8.647793	8.529012	9.594999	8.589558	9.000747	49.293368

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.189489	0.1975455	0.2131814	0.2053355	0.2170535	0.320012
4	lapply+pmin.int	0.468702	0.4916095	0.5383268	0.5107885	0.5656335	0.823576
3	lapply+pmin	0.525118	0.5532100	0.6033865	0.5771825	0.6289020	0.843668
2	apply+min	1.480868	1.5417425	1.8505077	1.5998610	1.7850685	17.184817

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	2.473505	2.488589	2.525206	2.487580	2.605964	2.573578
3	lapply+pmin	2.771232	2.800418	2.830391	2.810924	2.897451	2.636364
2	apply+min	7.815061	7.804493	8.680439	7.791449	8.224094	53.700539

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on integer+100x1000 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMins() and rowMins() 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	colMins	173.151	183.7955	202.3116	189.3960	208.7765	354.524
2	rowMins	189.489	197.5455	213.1814	205.3355	217.0535	320.012

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.00000	1.000000	1.0000000
2	rowMins	1.094357	1.074811	1.053728	1.08416	1.039645	0.9026526

Figure: Benchmarking of colMins() and rowMins() 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 5260353 281.0    7916910 422.9  7916910 422.9
Vcells 9875181  75.4   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5260341 281.0    7916910 422.9  7916910 422.9
Vcells 9925278  75.8   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	colMins	0.136101	0.1503335	0.1573457	0.1537545	0.161151	0.223754
2	apply+min	0.685892	0.7400800	0.7827800	0.7549980	0.783667	1.135333
4	lapply+pmin.int	1.542049	1.7352045	1.8195725	1.7817705	1.845368	3.037280
3	lapply+pmin	2.064401	2.2727160	2.5680116	2.3579955	2.448529	18.779174

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	apply+min	5.039581	4.922921	4.974906	4.910412	4.862936	5.074023
4	lapply+pmin.int	11.330181	11.542367	11.564172	11.588412	11.451170	13.574193
3	lapply+pmin	15.168154	15.117828	16.320827	15.336107	15.194004	83.927769

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.168151	0.1827725	0.1953922	0.1896045	0.1994725	0.282647
2	apply+min	0.671201	0.7403345	0.8154171	0.7745820	0.8492855	1.305584
4	lapply+pmin.int	1.316031	1.4738265	1.5586782	1.5223385	1.5858250	2.465012
3	lapply+pmin	1.793270	1.9826560	2.2973394	2.0594770	2.1693160	19.665516

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	apply+min	3.991656	4.050579	4.173233	4.085251	4.257657	4.619133
4	lapply+pmin.int	7.826483	8.063721	7.977178	8.029021	7.950093	8.721168
3	lapply+pmin	10.664641	10.847671	11.757581	10.861963	10.875264	69.576242

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on integer+1000x100 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMins() and rowMins() 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	colMins	136.101	150.3335	157.3457	153.7545	161.1510	223.754
2	rowMins	168.151	182.7725	195.3922	189.6045	199.4725	282.647

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
2	rowMins	1.235487	1.21578	1.241802	1.233164	1.237799	1.263204

Figure: Benchmarking of colMins() and rowMins() 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 5260578 281.0    7916910 422.9  7916910 422.9
Vcells 9991121  76.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5260563 281.0    7916910 422.9  7916910 422.9
Vcells 9991313  76.3   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	colMins	0.002559	0.0031140	0.0040405	0.0039365	0.004402	0.016816
4	lapply+pmin.int	0.020488	0.0215480	0.0243013	0.0222940	0.023612	0.066716
3	lapply+pmin	0.033412	0.0346730	0.0366451	0.0360750	0.036895	0.063086
2	apply+min	0.036128	0.0383625	0.0420743	0.0393625	0.040843	0.096656

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	8.006252	6.919717	6.014389	5.663407	5.363925	3.967412
3	lapply+pmin	13.056663	11.134554	9.069387	9.164232	8.381417	3.751546
2	apply+min	14.118015	12.319364	10.413062	9.999365	9.278283	5.747859

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.004152	0.0055465	0.0079116	0.0074035	0.0095270	0.022584
4	lapply+pmin.int	0.030853	0.0366285	0.0393853	0.0391410	0.0413905	0.064172
3	lapply+pmin	0.051094	0.0604975	0.0642979	0.0639605	0.0668780	0.144046
2	apply+min	0.057760	0.0670640	0.0731486	0.0711605	0.0776280	0.133239

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	7.430877	6.603894	4.978191	5.286824	4.344547	2.841481
3	lapply+pmin	12.305877	10.907329	8.127081	8.639225	7.019838	6.378232
2	apply+min	13.911368	12.091229	9.245791	9.611738	8.148210	5.899708

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on double+10x10 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMins() and rowMins() 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
1	colMins	2.559	3.1140	4.04053	3.9365	4.402	16.816
2	rowMins	4.152	5.5465	7.91156	7.4035	9.527	22.584

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.00000	1.00000	1.000000	1.000000	1.000000
2	rowMins	1.622509	1.78115	1.95805	1.880732	2.164243	1.343007

Figure: Benchmarking of colMins() and rowMins() 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 5260797 281.0    7916910 422.9  7916910 422.9
Vcells 9992189  76.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5260791 281.0    7916910 422.9  7916910 422.9
Vcells 10002296  76.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on double+100x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colMins	0.014245	0.0161605	0.0187732	0.0176050	0.0197225	0.038535
2	apply+min	0.169881	0.1887915	0.2171170	0.2124155	0.2254880	0.369419
4	lapply+pmin.int	0.178860	0.1962625	0.2260470	0.2163410	0.2466915	0.376520
3	lapply+pmin	0.234848	0.2671490	0.2966658	0.2866140	0.3142840	0.441471

	expr	min	lq	mean	median	uq	max
1	colMins	1.00000	1.00000	1.00000	1.00000	1.00000	1.000000
2	apply+min	11.92566	11.68228	11.56527	12.06563	11.43303	9.586584
4	lapply+pmin.int	12.55598	12.14458	12.04095	12.28861	12.50813	9.770858
3	lapply+pmin	16.48635	16.53099	15.80263	16.28026	15.93530	11.456364

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.025081	0.0284690	0.0331240	0.0325100	0.0364080	0.052407
4	lapply+pmin.int	0.144802	0.1667395	0.1900562	0.1842575	0.2066830	0.294848
2	apply+min	0.170805	0.1862275	0.2179104	0.2091380	0.2355375	0.406395
3	lapply+pmin	0.197326	0.2304875	0.2581022	0.2524005	0.2823025	0.359785

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	5.773374	5.856879	5.737728	5.667718	5.676857	5.626119
2	apply+min	6.810135	6.541413	6.578637	6.433036	6.469389	7.754594
3	lapply+pmin	7.867549	8.096087	7.792011	7.763780	7.753859	6.865209

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on double+100x100 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMins() and rowMins() 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	colMins	14.245	16.1605	18.77319	17.605	19.7225	38.535
2	rowMins	25.081	28.4690	33.12395	32.510	36.4080	52.407

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMins	1.760688	1.761641	1.764428	1.846634	1.846013	1.359984

Figure: Benchmarking of colMins() and rowMins() 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 5261027 281.0    7916910 422.9  7916910 422.9
Vcells 9993356  76.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5261015 281.0    7916910 422.9  7916910 422.9
Vcells 10003453  76.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on double+1000x10 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colMins	0.010171	0.0127705	0.0151684	0.0138865	0.0156765	0.037325
2	apply+min	0.094885	0.1024595	0.1135641	0.1085090	0.1193030	0.203318
4	lapply+pmin.int	0.983094	1.0817250	1.1584047	1.1319085	1.1808190	1.937538
3	lapply+pmin	1.491201	1.6045695	1.7541474	1.6675530	1.7674035	6.749631

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	apply+min	9.328974	8.023139	7.486877	7.813992	7.610308	5.447234
4	lapply+pmin.int	96.656573	84.704984	76.369505	81.511432	75.324148	51.909926
3	lapply+pmin	146.613017	125.646568	115.644699	120.084470	112.742226	180.834052

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.021171	0.0235035	0.0263587	0.0249765	0.027178	0.049838
2	apply+min	0.091676	0.1012805	0.1140803	0.1080480	0.118503	0.202495
4	lapply+pmin.int	0.985767	1.0501180	1.1722938	1.1027315	1.154998	6.566454
3	lapply+pmin	1.473725	1.5633845	1.6783980	1.6493110	1.751914	2.322945

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	apply+min	4.330263	4.309167	4.327991	4.325986	4.360255	4.063064
4	lapply+pmin.int	46.562137	44.679218	44.474628	44.150762	42.497535	131.755969
3	lapply+pmin	69.610552	66.517093	63.675272	66.034512	64.460740	46.609916

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on double+1000x10 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

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

	expr	min	lq	mean	median	uq	max
1	colMins	10.171	12.7705	15.16842	13.8865	15.6765	37.325
2	rowMins	21.171	23.5035	26.35871	24.9765	27.1780	49.838

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMins	2.081506	1.840453	1.737736	1.798617	1.733678	1.335244

Figure: Benchmarking of colMins() and rowMins() 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 5261243 281.0    7916910 422.9  7916910 422.9
Vcells 9993493  76.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5261237 281.0    7916910 422.9  7916910 422.9
Vcells 10003600  76.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on double+10x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colMins	0.025067	0.0296085	0.0346065	0.031992	0.0376200	0.062044
4	lapply+pmin.int	0.097867	0.1034770	0.1156185	0.108960	0.1237030	0.202144
3	lapply+pmin	0.107130	0.1137945	0.1311495	0.118408	0.1443055	0.251911
2	apply+min	0.932091	0.9844785	1.0660914	1.020206	1.1126960	1.471314

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	3.904217	3.494841	3.340946	3.405851	3.288224	3.258075
3	lapply+pmin	4.273746	3.843305	3.789734	3.701175	3.835872	4.060199
2	apply+min	37.183987	33.249861	30.806084	31.889394	29.577246	23.714042

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.042748	0.0472885	0.0545116	0.0511535	0.0580690	0.102756
4	lapply+pmin.int	0.083369	0.0894075	0.1006704	0.0946780	0.1087155	0.170449
3	lapply+pmin	0.095375	0.1024595	0.1160465	0.1102000	0.1241225	0.193722
2	apply+min	1.079917	1.1519850	1.2618139	1.1978620	1.3460985	1.812822

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	1.950243	1.890682	1.846771	1.850861	1.872178	1.658774
3	lapply+pmin	2.231098	2.166690	2.128841	2.154300	2.137500	1.885262
2	apply+min	25.262398	24.360785	23.147626	23.417010	23.181017	17.642006

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on double+10x1000 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMins() and rowMins() 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
1	colMins	25.067	29.6085	34.60652	31.9920	37.620	62.044
2	rowMins	42.748	47.2885	54.51159	51.1535	58.069	102.756

	expr	min	lq	mean	median	uq	max
1	colMins	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMins	1.70535	1.597126	1.575183	1.598947	1.543567	1.656179

Figure: Benchmarking of colMins() and rowMins() 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 5261485 281.0    7916910 422.9  7916910 422.9
Vcells 9994877  76.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5261467 281.0    7916910 422.9  7916910 422.9
Vcells 10094964  77.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on double+100x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colMins	0.127265	0.1349055	0.1486744	0.1405395	0.1616905	0.222179
4	lapply+pmin.int	0.828202	0.8813465	1.0957530	0.9127240	1.0373565	13.710847
3	lapply+pmin	0.881749	0.9218375	1.0259031	0.9729400	1.1231360	1.514968
2	apply+min	1.456139	1.5760315	1.8804393	1.6519590	1.8856640	15.256789

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	lapply+pmin.int	6.507696	6.533066	7.370152	6.494430	6.415692	61.710814
3	lapply+pmin	6.928449	6.833209	6.900334	6.922894	6.946209	6.818682
2	apply+min	11.441787	11.682485	12.648036	11.754411	11.662182	68.668907

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.217670	0.2241645	0.2422804	0.2308085	0.2469435	0.330212
4	lapply+pmin.int	0.549340	0.5712525	0.6252280	0.5996425	0.6684070	0.894148
3	lapply+pmin	0.609355	0.6427005	0.8184364	0.6692155	0.7685150	13.080882
2	apply+min	1.462504	1.5650950	1.8555314	1.6297425	1.8673170	15.008537

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
4	lapply+pmin.int	2.523729	2.548363	2.580596	2.598009	2.706720	2.70780
3	lapply+pmin	2.799444	2.867093	3.378054	2.899440	3.112109	39.61359
2	apply+min	6.718905	6.981904	7.658610	7.061016	7.561718	45.45122

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on double+100x1000 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMins() and rowMins() 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	colMins	127.265	134.9055	148.6744	140.5395	161.6905	222.179
2	rowMins	217.670	224.1645	242.2804	230.8085	246.9435	330.212

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.000000	1.00000	1.000000
2	rowMins	1.710368	1.661641	1.629604	1.642303	1.52726	1.486243

Figure: Benchmarking of colMins() and rowMins() 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 5261697 281.1    7916910 422.9  7916910 422.9
Vcells 9995006  76.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colMins = colMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 2L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(nrow(X)), function(i) X[i, 
+     ])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(nrow(X)), function(i) X[i, ])), unit = "ms")
> X <- t(X)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5261691 281.1    7916910 422.9  7916910 422.9
Vcells 10095113  77.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowMins = rowMins(X, na.rm = FALSE), `apply+min` = apply(X, MARGIN = 1L, 
+     FUN = min, na.rm = FALSE), `lapply+pmin` = do.call(pmin, lapply(seq_len(ncol(X)), function(i) X[, 
+     i])), `lapply+pmin.int` = do.call(pmin.int, lapply(seq_len(ncol(X)), function(i) X[, i])), unit = "ms")

Table: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on double+1000x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colMins	0.089180	0.0981350	0.1067731	0.1015805	0.109393	0.164375
2	apply+min	0.733633	0.7785275	1.0328662	0.8040005	0.899850	19.219543
4	lapply+pmin.int	1.681829	1.8202495	1.9878918	1.9013125	2.084604	3.414818
3	lapply+pmin	2.181463	2.3605590	2.6874884	2.4375815	2.667784	19.271342

	expr	min	lq	mean	median	uq	max
1	colMins	1.00000	1.00000	1.000000	1.00000	1.000000	1.00000
2	apply+min	8.22643	7.93323	9.673467	7.91491	8.225846	116.92498
4	lapply+pmin.int	18.85881	18.54842	18.617907	18.71730	19.056100	20.77456
3	lapply+pmin	24.46135	24.05420	25.170087	23.99655	24.387159	117.24010

Table: Benchmarking of rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() 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	rowMins	0.194110	0.2055925	0.2247365	0.2121715	0.2282480	0.386047
2	apply+min	0.734856	0.7807880	1.0438740	0.8124760	0.9384715	17.674332
4	lapply+pmin.int	1.395158	1.4954840	1.6229063	1.5350885	1.7286160	2.635179
3	lapply+pmin	1.881986	1.9766720	2.3272536	2.0611610	2.1800355	20.420684

	expr	min	lq	mean	median	uq	max
1	rowMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	apply+min	3.785771	3.797745	4.644879	3.829336	4.111631	45.782850
4	lapply+pmin.int	7.187461	7.274020	7.221373	7.235131	7.573411	6.826057
3	lapply+pmin	9.695461	9.614514	10.355476	9.714599	9.551170	52.896886

Figure: Benchmarking of colMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on double+1000x100 data as well as rowMins(), apply+min(), lapply+pmin() and lapply+pmin.int() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colMins() and rowMins() 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	colMins	89.18	98.1350	106.7731	101.5805	109.393	164.375
2	rowMins	194.11	205.5925	224.7365	212.1715	228.248	386.047

	expr	min	lq	mean	median	uq	max
1	colMins	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowMins	2.176609	2.094997	2.104804	2.088703	2.086495	2.348575

Figure: Benchmarking of colMins() and rowMins() 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 29.38 secs.

Reproducibility

To reproduce this report, do:

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

---

Copyright Henrik Bengtsson. Last updated on 2021-08-25 22:21:53 (+0200 UTC). Powered by RSP.