matrixStats.benchmarks

---

colDiffs() and rowDiffs() benchmarks

This report benchmark the performance of colDiffs() and rowDiffs() against alternative methods.

Alternative methods

• apply() + diff()
• apply() + diff2()
• diff()

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  5234885 279.6    7916910 422.9  7916910 422.9
Vcells 10018469  76.5   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5220903 278.9    7916910 422.9  7916910 422.9
Vcells 9972243  76.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.002410	0.0032175	0.0040717	0.0039040	0.0042950	0.018556
4	diff	0.008681	0.0099505	0.0123318	0.0112145	0.0121575	0.086460
3	apply+diff2	0.051363	0.0550640	0.0618324	0.0579920	0.0634745	0.191141
2	apply+diff	0.089033	0.0937465	0.1039585	0.0975035	0.1069685	0.187278

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	diff	3.602075	3.092618	3.028634	2.872567	2.830617	4.659409
3	apply+diff2	21.312448	17.113908	15.185821	14.854508	14.778696	10.300765
2	apply+diff	36.943154	29.136441	25.531849	24.975282	24.905355	10.092585

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.002251	0.0032095	0.0040860	0.0038905	0.0044525	0.011247
4	diff + t	0.010882	0.0133260	0.0155085	0.0148845	0.0168720	0.034290
3	apply+diff2	0.049278	0.0543650	0.0612715	0.0579860	0.0642190	0.169976
2	apply+diff	0.088994	0.0955335	0.1053381	0.1009290	0.1104345	0.193562

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	diff + t	4.834296	4.152049	3.795472	3.825858	3.789332	3.048813
3	apply+diff2	21.891604	16.938775	14.995279	14.904511	14.423133	15.113008
2	apply+diff	39.535318	29.765851	25.779924	25.942424	24.802807	17.210100

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on integer+10x10 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

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

	expr	min	lq	mean	median	uq	max
2	rowDiffs	2.251	3.2095	4.08605	3.8905	4.4525	11.247
1	colDiffs	2.410	3.2175	4.07172	3.9040	4.2950	18.556

	expr	min	lq	mean	median	uq	max
2	rowDiffs	1.000000	1.000000	1.0000000	1.00000	1.0000000	1.000000
1	colDiffs	1.070635	1.002493	0.9964929	1.00347	0.9646266	1.649862

Figure: Benchmarking of colDiffs() and rowDiffs() 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 5219531 278.8    7916910 422.9  7916910 422.9
Vcells 9589103  73.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5219525 278.8    7916910 422.9  7916910 422.9
Vcells 9594146  73.2   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.010233	0.0120500	0.0143566	0.0129230	0.0157705	0.036864
4	diff	0.087554	0.0909240	0.1065238	0.0999280	0.1161120	0.189905
3	apply+diff2	0.274204	0.2875025	0.3443083	0.3087115	0.3727895	0.608045
2	apply+diff	0.665882	0.6867605	0.7869371	0.7301075	0.8513480	1.293287

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.00000	1.000000	1.00000	1.000000	1.000000
4	diff	8.556044	7.54556	7.419829	7.73257	7.362607	5.151503
3	apply+diff2	26.796052	23.85913	23.982513	23.88853	23.638407	16.494276
2	apply+diff	65.072022	56.99257	54.813460	56.49675	53.983577	35.082655

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.010478	0.0124790	0.0148709	0.0141820	0.0163615	0.029732
4	diff + t	0.101260	0.1061495	0.1226143	0.1162905	0.1309370	0.203687
3	apply+diff2	0.275292	0.2883290	0.3386836	0.3113205	0.3673170	0.728100
2	apply+diff	0.662150	0.6785940	0.7995663	0.7387975	0.8693675	1.547435

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.000000	1.000000	1.000000	1.000000	1.00000	1.000000
4	diff + t	9.664058	8.506251	8.245246	8.199866	8.00275	6.850767
3	apply+diff2	26.273335	23.105137	22.774908	21.951805	22.45008	24.488766
2	apply+diff	63.194312	54.378876	53.767139	52.094028	53.13495	52.046112

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on integer+100x100 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colDiffs() and rowDiffs() 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	colDiffs	10.233	12.050	14.35664	12.923	15.7705	36.864
2	rowDiffs	10.478	12.479	14.87091	14.182	16.3615	29.732

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowDiffs	1.023942	1.035602	1.035821	1.097423	1.037475	0.8065321

Figure: Benchmarking of colDiffs() and rowDiffs() 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 5220319 278.8    7916910 422.9  7916910 422.9
Vcells 9592922  73.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5220307 278.8    7916910 422.9  7916910 422.9
Vcells 9597955  73.3   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.011760	0.0140630	0.0159658	0.0152555	0.0168710	0.037522
4	diff	0.103828	0.1132165	0.1227273	0.1170555	0.1220665	0.245433
3	apply+diff2	0.143026	0.1567360	0.1734666	0.1625070	0.1759850	0.380707
2	apply+diff	0.293264	0.3146275	0.3432318	0.3312005	0.3574070	0.620222

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	diff	8.828912	8.050665	7.686904	7.673003	7.235285	6.541043
3	apply+diff2	12.162075	11.145275	10.864916	10.652355	10.431213	10.146234
2	apply+diff	24.937415	22.372716	21.497990	21.710236	21.184696	16.529556

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.011410	0.0133540	0.0156139	0.0145475	0.0163070	0.039514
4	diff + t	0.112171	0.1280085	0.1407737	0.1369440	0.1466525	0.235174
3	apply+diff2	0.130754	0.1534225	0.1698552	0.1654425	0.1776625	0.379489
2	apply+diff	0.269161	0.3030970	0.3387973	0.3320595	0.3625425	0.504731

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	diff + t	9.830938	9.585779	9.015911	9.413576	8.993224	5.951663
3	apply+diff2	11.459597	11.488880	10.878448	11.372573	10.894861	9.603912
2	apply+diff	23.589921	22.697094	21.698417	22.825881	22.232323	12.773473

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on integer+1000x10 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

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

	expr	min	lq	mean	median	uq	max
2	rowDiffs	11.41	13.354	15.61392	14.5475	16.307	39.514
1	colDiffs	11.76	14.063	15.96576	15.2555	16.871	37.522

	expr	min	lq	mean	median	uq	max
2	rowDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colDiffs	1.030675	1.053093	1.022534	1.048668	1.034586	0.9495875

Figure: Benchmarking of colDiffs() and rowDiffs() 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 5220555 278.9    7916910 422.9  7916910 422.9
Vcells 9593747  73.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5220549 278.9    7916910 422.9  7916910 422.9
Vcells 9598790  73.3   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.010385	0.0121705	0.0170979	0.0155630	0.0203250	0.062486
4	diff	0.086857	0.0961335	0.1046856	0.1014835	0.1096045	0.152155
3	apply+diff2	1.845579	2.0226450	2.1602227	2.0687535	2.1450155	3.594670
2	apply+diff	4.717286	5.1664400	5.5673794	5.3721440	5.6495685	10.634311

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.000000	1.00000	1.000000	1.000000	1.000000
4	diff	8.363698	7.898895	6.12273	6.520819	5.392595	2.435025
3	apply+diff2	177.715840	166.192432	126.34455	132.927681	105.535818	57.527606
2	apply+diff	454.240347	424.505156	325.61830	345.186918	277.961550	170.187098

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.009787	0.011020	0.0158380	0.0155095	0.0183135	0.048899
4	diff + t	0.097048	0.110901	0.1259177	0.1227690	0.1325280	0.235955
3	apply+diff2	1.808700	2.002831	2.1504046	2.0630150	2.1220855	8.365290
2	apply+diff	4.638639	5.136468	5.6383491	5.3584645	5.7189120	12.578408

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
4	diff + t	9.916011	10.06361	7.950343	7.915729	7.236629	4.825354
3	apply+diff2	184.806376	181.74510	135.774840	133.016216	115.875474	171.072824
2	apply+diff	473.959232	466.10417	356.000885	345.495632	312.278483	257.232418

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on integer+10x1000 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colDiffs() and rowDiffs() 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	rowDiffs	9.787	11.0200	15.83802	15.5095	18.3135	48.899
1	colDiffs	10.385	12.1705	17.09787	15.5630	20.3250	62.486

	expr	min	lq	mean	median	uq	max
2	rowDiffs	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000
1	colDiffs	1.061101	1.104401	1.079546	1.00345	1.109837	1.277858

Figure: Benchmarking of colDiffs() and rowDiffs() 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 5220789 278.9    7916910 422.9  7916910 422.9
Vcells 9594345  73.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5220783 278.9    7916910 422.9  7916910 422.9
Vcells 9644388  73.6   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.089649	0.1018180	0.1172249	0.1081050	0.116504	0.247380
4	diff	0.815440	0.9071745	1.1495363	0.9351265	0.972877	20.430142
3	apply+diff2	2.644608	2.8736100	3.0427602	2.9520865	3.047727	4.831793
2	apply+diff	6.575066	7.2131575	8.1939990	7.3440485	7.752326	28.913440

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
4	diff	9.095918	8.909766	9.806244	8.650169	8.350589	82.58607
3	apply+diff2	29.499582	28.223006	25.956594	27.307585	26.159844	19.53187
2	apply+diff	73.342324	70.843638	69.899793	67.934402	66.541282	116.87865

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.085926	0.0946745	0.2806362	0.102276	0.109824	17.608900
4	diff + t	0.956782	1.0473905	1.4834830	1.086612	1.161292	20.341576
3	apply+diff2	2.733862	2.9199090	3.0264864	2.980916	3.088248	3.999677
2	apply+diff	6.646363	7.2116670	7.9014230	7.328157	7.532244	28.131772

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.00000	1.00000	1.000000	1.00000	1.00000	1.0000000
4	diff + t	11.13495	11.06307	5.286143	10.62431	10.57412	1.1551872
3	apply+diff2	31.81647	30.84156	10.784377	29.14580	28.11997	0.2271395
2	apply+diff	77.34985	76.17328	28.155395	71.65080	68.58468	1.5975883

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on integer+100x1000 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colDiffs() and rowDiffs() 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
2	rowDiffs	85.926	94.6745	280.6362	102.276	109.824	17608.90
1	colDiffs	89.649	101.8180	117.2249	108.105	116.504	247.38

	expr	min	lq	mean	median	uq	max
2	rowDiffs	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
1	colDiffs	1.043328	1.075453	0.4177114	1.056993	1.060825	0.0140486

Figure: Benchmarking of colDiffs() and rowDiffs() 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 5221013 278.9    7916910 422.9  7916910 422.9
Vcells 9595023  73.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5221007 278.9    7916910 422.9  7916910 422.9
Vcells 9645066  73.6   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.082110	0.0902385	0.0992295	0.0943880	0.105174	0.208432
4	diff	0.808117	0.8352460	0.9882278	0.8880655	0.974726	8.515980
3	apply+diff2	1.010645	1.0632850	1.2178040	1.1018805	1.204460	8.988312
2	apply+diff	2.259232	2.3836330	2.7445288	2.4882985	2.723021	11.129779

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.000000	1.000000	1.00000	1.000000	1.00000
4	diff	9.841883	9.255983	9.959008	9.40867	9.267747	40.85735
3	apply+diff2	12.308428	11.783053	12.272595	11.67395	11.452065	43.12347
2	apply+diff	27.514700	26.414812	27.658384	26.36245	25.890624	53.39765

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.087239	0.094856	0.1045206	0.0991225	0.112059	0.194481
4	diff + t	0.944986	1.006004	1.1652765	1.0588235	1.154020	8.434238
3	apply+diff2	1.020452	1.077211	1.1790055	1.1315640	1.224992	1.785453
2	apply+diff	2.280171	2.378445	2.8924438	2.4819050	2.941225	12.003412

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.00000	1.00000	1.00000	1.00000	1.00000	1.000000
4	diff + t	10.83215	10.60559	11.14877	10.68197	10.29833	43.367928
3	apply+diff2	11.69720	11.35628	11.28012	11.41581	10.93167	9.180604
2	apply+diff	26.13706	25.07427	27.67343	25.03877	26.24711	61.720230

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on integer+1000x100 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colDiffs() and rowDiffs() 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	colDiffs	82.110	90.2385	99.22954	94.3880	105.174	208.432
2	rowDiffs	87.239	94.8560	104.52062	99.1225	112.059	194.481

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.00000	1.000000	1.00000	1.000000	1.0000000
2	rowDiffs	1.062465	1.05117	1.053322	1.05016	1.065463	0.9330669

Figure: Benchmarking of colDiffs() and rowDiffs() 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 5221245 278.9    7916910 422.9  7916910 422.9
Vcells 9710895  74.1   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5221230 278.9    7916910 422.9  7916910 422.9
Vcells 9711023  74.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.002882	0.0038320	0.0050783	0.0045370	0.0051030	0.019096
4	diff	0.009565	0.0107785	0.0140648	0.0127385	0.0136985	0.071148
3	apply+diff2	0.060373	0.0623130	0.0704627	0.0643260	0.0674945	0.160876
2	apply+diff	0.098355	0.1067560	0.1182098	0.1094600	0.1136130	0.252843

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	diff	3.318876	2.812761	2.769608	2.807692	2.684401	3.725807
3	apply+diff2	20.948300	16.261221	13.875398	14.178091	13.226435	8.424592
2	apply+diff	34.127342	27.859081	23.277665	24.126075	22.263962	13.240626

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.002457	0.0034230	0.0042402	0.0040435	0.0044610	0.012912
4	diff + t	0.011420	0.0140535	0.0165485	0.0163245	0.0175085	0.037819
3	apply+diff2	0.054587	0.0581645	0.0633354	0.0620350	0.0635035	0.166219
2	apply+diff	0.094534	0.0997970	0.1091344	0.1075865	0.1092550	0.240004

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000
4	diff + t	4.647945	4.105609	3.902738	4.03722	3.924793	2.928981
3	apply+diff2	22.216931	16.992258	14.936822	15.34191	14.235261	12.873219
2	apply+diff	38.475377	29.154835	25.737907	26.60727	24.491145	18.587670

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on double+10x10 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colDiffs() and rowDiffs() 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	rowDiffs	2.457	3.423	4.24022	4.0435	4.461	12.912
1	colDiffs	2.882	3.832	5.07825	4.5370	5.103	19.096

	expr	min	lq	mean	median	uq	max
2	rowDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colDiffs	1.172975	1.119486	1.197638	1.122048	1.143914	1.478934

Figure: Benchmarking of colDiffs() and rowDiffs() 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 5221464 278.9    7916910 422.9  7916910 422.9
Vcells 9711901  74.1   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5221458 278.9    7916910 422.9  7916910 422.9
Vcells 9721944  74.2   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.008937	0.0115895	0.0140659	0.0132655	0.0158875	0.034496
4	diff	0.050511	0.0554920	0.0638905	0.0616950	0.0677750	0.159708
3	apply+diff2	0.277555	0.2939775	0.3352664	0.3135475	0.3634770	0.578353
2	apply+diff	0.607888	0.6584725	0.7502030	0.7012380	0.8136910	1.165844

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	diff	5.651897	4.788127	4.542222	4.650786	4.265932	4.629754
3	apply+diff2	31.056842	25.365848	23.835389	23.636312	22.878175	16.765799
2	apply+diff	68.019246	56.816299	53.334836	52.861784	51.215799	33.796498

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.008128	0.0100500	0.0121818	0.0116100	0.0131995	0.025898
4	diff + t	0.066060	0.0693595	0.0789167	0.0747990	0.0843110	0.122213
3	apply+diff2	0.268388	0.2806290	0.3194527	0.2984125	0.3504460	0.469339
2	apply+diff	0.601671	0.6159255	0.7099022	0.6555765	0.7829940	1.215689

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	diff + t	8.127461	6.901443	6.478235	6.442636	6.387439	4.719013
3	apply+diff2	33.020177	27.923284	26.223721	25.703058	26.549945	18.122596
2	apply+diff	74.024483	61.286119	58.275549	56.466538	59.319974	46.941424

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on double+100x100 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colDiffs() and rowDiffs() 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	rowDiffs	8.128	10.0500	12.18182	11.6100	13.1995	25.898
1	colDiffs	8.937	11.5895	14.06591	13.2655	15.8875	34.496

	expr	min	lq	mean	median	uq	max
2	rowDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colDiffs	1.099533	1.153184	1.154664	1.142593	1.203644	1.331995

Figure: Benchmarking of colDiffs() and rowDiffs() 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 5221694 278.9    7916910 422.9  7916910 422.9
Vcells 9713001  74.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5221682 278.9    7916910 422.9  7916910 422.9
Vcells 9723034  74.2   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.009700	0.0116410	0.0140630	0.0129225	0.0146775	0.038085
4	diff	0.061551	0.0675830	0.0760927	0.0713770	0.0823970	0.183753
3	apply+diff2	0.138074	0.1532545	0.1715054	0.1605680	0.1773945	0.388456
2	apply+diff	0.223189	0.2453330	0.2867097	0.2697225	0.3103900	0.484405

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
4	diff	6.345464	5.805601	5.410822	5.523467	5.613831	4.824813
3	apply+diff2	14.234433	13.165063	12.195464	12.425460	12.086152	10.199711
2	apply+diff	23.009175	21.074908	20.387445	20.872316	21.147334	12.719050

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.009308	0.0115990	0.0146869	0.0125940	0.0157475	0.047693
4	diff + t	0.079128	0.0885355	0.1029702	0.0962810	0.1138945	0.238251
3	apply+diff2	0.135614	0.1502960	0.1745175	0.1626475	0.1935285	0.356424
2	apply+diff	0.222073	0.2418025	0.3042227	0.2782955	0.3219105	0.830262

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000
4	diff + t	8.501074	7.633029	7.011026	7.64499	7.232545	4.995513
3	apply+diff2	14.569617	12.957669	11.882535	12.91468	12.289475	7.473298
2	apply+diff	23.858294	20.846840	20.713893	22.09747	20.442007	17.408467

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on double+1000x10 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colDiffs() and rowDiffs() 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	rowDiffs	9.308	11.599	14.68689	12.5940	15.7475	47.693
1	colDiffs	9.700	11.641	14.06305	12.9225	14.6775	38.085

	expr	min	lq	mean	median	uq	max
2	rowDiffs	1.000000	1.000000	1.000000	1.000000	1.0000000	1.0000000
1	colDiffs	1.042114	1.003621	0.957524	1.026084	0.9320527	0.7985449

Figure: Benchmarking of colDiffs() and rowDiffs() 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 5221910 278.9    7916910 422.9  7916910 422.9
Vcells 9713138  74.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5221904 278.9    7916910 422.9  7916910 422.9
Vcells 9723181  74.2   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.009808	0.0116055	0.0188031	0.0180945	0.023668	0.036956
4	diff	0.056120	0.0619115	0.0766055	0.0700460	0.084563	0.145850
3	apply+diff2	1.848987	2.0830720	2.4352643	2.1809515	2.423824	5.236078
2	apply+diff	4.714430	5.3118405	6.2771351	5.4645685	5.986134	14.150990

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
4	diff	5.72186	5.334669	4.074083	3.871121	3.572883	3.946585
3	apply+diff2	188.51825	179.490069	129.513841	120.531183	102.409350	141.684111
2	apply+diff	480.67190	457.700271	333.834765	302.001630	252.920969	382.914547

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.008273	0.0118360	0.0181096	0.0173575	0.0225460	0.061652
4	diff + t	0.071806	0.0810575	0.0987393	0.0948085	0.1088825	0.231444
3	apply+diff2	1.831776	2.0396465	5.9064930	2.1343270	2.2788475	363.466332
2	apply+diff	4.716785	5.2535175	5.9440702	5.5665380	6.0398345	12.269958

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
4	diff + t	8.67956	6.848386	5.452328	5.462106	4.829349	3.754039
3	apply+diff2	221.41617	172.325659	326.153131	122.962812	101.075468	5895.450788
2	apply+diff	570.14203	443.859201	328.228124	320.699294	267.889404	199.019626

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on double+10x1000 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colDiffs() and rowDiffs() 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	rowDiffs	8.273	11.8360	18.10957	17.3575	22.546	61.652
1	colDiffs	9.808	11.6055	18.80312	18.0945	23.668	36.956

	expr	min	lq	mean	median	uq	max
2	rowDiffs	1.000000	1.0000000	1.000000	1.00000	1.000000	1.0000000
1	colDiffs	1.185543	0.9805255	1.038297	1.04246	1.049765	0.5994291

Figure: Benchmarking of colDiffs() and rowDiffs() 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 5222140 278.9    7916910 422.9  7916910 422.9
Vcells 9714417  74.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5222134 278.9    7916910 422.9  7916910 422.9
Vcells 9814460  74.9   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.075012	0.0985830	0.2902086	0.1139370	0.1237880	17.79806
4	diff	0.493519	0.5568505	1.0134305	0.6001095	0.6565765	20.83392
3	apply+diff2	2.712056	2.9899120	3.5664476	3.1660935	3.3969795	19.59518
2	apply+diff	6.142650	6.8205220	8.0680224	7.3424865	8.2348650	27.38095

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.000000	1.000000	1.000000	1.00000	1.000000
4	diff	6.579201	5.648545	3.492076	5.267029	5.30404	1.170573
3	apply+diff2	36.154962	30.328880	12.289255	27.788107	27.44191	1.100973
2	apply+diff	81.888898	69.185580	27.800770	64.443390	66.52394	1.538423

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.069182	0.0904975	0.1012337	0.1013925	0.111012	0.139194
4	diff + t	0.642505	0.7206050	0.8094761	0.7759935	0.862320	1.347559
3	apply+diff2	2.647845	2.9005910	3.8183691	3.1300825	3.479826	19.288616
2	apply+diff	6.249945	6.8263040	8.1394325	7.1884865	8.074667	25.079023

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.00000	1.000000	1.00000	1.000000	1.000000	1.000000
4	diff + t	9.28717	7.962706	7.99611	7.653362	7.767809	9.681157
3	apply+diff2	38.27361	32.051615	37.71834	30.870947	31.346390	138.573617
2	apply+diff	90.34062	75.430857	80.40237	70.897616	72.736884	180.173161

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on double+100x1000 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colDiffs() and rowDiffs() 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	rowDiffs	69.182	90.4975	101.2337	101.3925	111.012	139.194
1	colDiffs	75.012	98.5830	290.2086	113.9370	123.788	17798.058

	expr	min	lq	mean	median	uq	max
2	rowDiffs	1.00000	1.000000	1.000000	1.000000	1.000000	1.0000
1	colDiffs	1.08427	1.089345	2.866718	1.123722	1.115087	127.8651

Figure: Benchmarking of colDiffs() and rowDiffs() 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 5222370 279.0    7916910 422.9  7916910 422.9
Vcells 9714576  74.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs = colDiffs(X), `apply+diff` = apply(X, MARGIN = 2L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 2L, FUN = diff2), diff = diff(X), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5222358 279.0    7916910 422.9  7916910 422.9
Vcells 9814609  74.9   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs = rowDiffs(X), `apply+diff` = apply(X, MARGIN = 1L, FUN = diff), 
+     `apply+diff2` = apply(X, MARGIN = 1L, FUN = diff2), `diff + t` = diff(t(X)), unit = "ms")

Table: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() 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	colDiffs	0.073464	0.086826	0.1017142	0.0995620	0.1129340	0.200830
4	diff	0.506183	0.567059	0.7713862	0.5982465	0.6454205	8.016259
3	apply+diff2	1.085764	1.203712	1.4604403	1.2906665	1.3534965	8.590198
2	apply+diff	1.944802	2.112408	2.5438341	2.2889845	2.4551580	9.843505

	expr	min	lq	mean	median	uq	max
1	colDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
4	diff	6.890218	6.530982	7.583863	6.008783	5.715024	39.91565
3	apply+diff2	14.779538	13.863497	14.358280	12.963445	11.984845	42.77348
2	apply+diff	26.472857	24.329210	25.009636	22.990544	21.739760	49.01412

Table: Benchmarking of rowDiffs(), apply+diff(), apply+diff2() and diff + t() 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	rowDiffs	0.075482	0.0845145	0.0957152	0.091523	0.102683	0.162823
4	diff + t	0.672398	0.7471305	0.8871564	0.790105	0.857436	7.560255
3	apply+diff2	1.112277	1.2630005	1.4263114	1.334588	1.417919	7.707001
2	apply+diff	1.935607	2.2474060	2.8169005	2.412302	2.532773	10.904855

	expr	min	lq	mean	median	uq	max
1	rowDiffs	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
4	diff + t	8.908058	8.840264	9.268713	8.632857	8.350321	46.43235
3	apply+diff2	14.735659	14.944187	14.901623	14.581990	13.808703	47.33361
2	apply+diff	25.643292	26.591958	29.430032	26.357331	24.665948	66.97368

Figure: Benchmarking of colDiffs(), apply+diff(), apply+diff2() and diff() on double+1000x100 data as well as rowDiffs(), apply+diff(), apply+diff2() and diff + t() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colDiffs() and rowDiffs() 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	rowDiffs	75.482	84.5145	95.71517	91.523	102.683	162.823
1	colDiffs	73.464	86.8260	101.71416	99.562	112.934	200.830

	expr	min	lq	mean	median	uq	max
2	rowDiffs	1.0000000	1.00000	1.000000	1.000000	1.000000	1.000000
1	colDiffs	0.9732651	1.02735	1.062675	1.087836	1.099832	1.233425

Figure: Benchmarking of colDiffs() and rowDiffs() 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 34.71 secs.

Reproducibility

To reproduce this report, do:

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

---

Copyright Henrik Bengtsson. Last updated on 2021-08-25 22:16:12 (+0200 UTC). Powered by RSP.