colRowSums2

matrixStats: Benchmark report

---

colSums2() and rowSums2() benchmarks

This report benchmark the performance of colSums2() and rowSums2() against alternative methods.

Alternative methods

• apply() + sum()
• colSums() and rowSums()
• .colSums() and .rowSums()

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 3201658 171.0    5709258 305.0  5709258 305.0
Vcells 6523933  49.8   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3199978 170.9    5709258 305.0  5709258 305.0
Vcells 6519323  49.8   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() 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	colSums2	0.001133	0.001386	0.0018295	0.0017945	0.0019915	0.012452
2	.colSums	0.001634	0.001849	0.0022563	0.0021175	0.0024265	0.014209
3	colSums	0.002971	0.003299	0.0038329	0.0036710	0.0042280	0.012916
4	apply+sum	0.019245	0.020566	0.0221572	0.0209930	0.0216235	0.113238

	expr	min	lq	mean	median	uq	max
1	colSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.colSums	1.442189	1.334055	1.233285	1.179994	1.218428	1.141102
3	colSums	2.622242	2.380231	2.095098	2.045695	2.123023	1.037263
4	apply+sum	16.985878	14.838384	12.111228	11.698523	10.857896	9.093961

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() 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	rowSums2	0.001095	0.0012995	0.0016709	0.0017025	0.0018095	0.009964
2	.rowSums	0.002430	0.0026155	0.0028790	0.0027960	0.0028985	0.012598
3	rowSums	0.003668	0.0038785	0.0049349	0.0043325	0.0046560	0.036676
4	apply+sum	0.019635	0.0203530	0.0211361	0.0205730	0.0208930	0.063792

	expr	min	lq	mean	median	uq	max
1	rowSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowSums	2.219178	2.012697	1.723006	1.642291	1.601824	1.264352
3	rowSums	3.349772	2.984609	2.953468	2.544787	2.573087	3.680851
4	apply+sum	17.931507	15.662178	12.649524	12.083994	11.546283	6.402248

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on integer+10x10 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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	rowSums2	1.095	1.2995	1.67090	1.7025	1.8095	9.964
1	colSums2	1.133	1.3860	1.82948	1.7945	1.9915	12.452

	expr	min	lq	mean	median	uq	max
2	rowSums2	1.000000	1.000000	1.000000	1.000000	1.00000	1.000000
1	colSums2	1.034703	1.066564	1.094907	1.054038	1.10058	1.249699

Figure: Benchmarking of colSums2() and rowSums2() 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 3198558 170.9    5709258 305.0  5709258 305.0
Vcells 6136144  46.9   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3198549 170.9    5709258 305.0  5709258 305.0
Vcells 6141182  46.9   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on integer+100x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	.colSums	0.009088	0.0096250	0.0102482	0.010057	0.010423	0.017808
1	colSums2	0.010865	0.0112420	0.0123764	0.011790	0.012167	0.042313
3	colSums	0.010617	0.0112615	0.0125573	0.011954	0.012511	0.053884
4	apply+sum	0.139616	0.1417465	0.1512812	0.143997	0.149455	0.295777

	expr	min	lq	mean	median	uq	max
2	.colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colSums2	1.195533	1.168000	1.207671	1.172318	1.167322	2.376067
3	colSums	1.168244	1.170026	1.225325	1.188625	1.200326	3.025831
4	apply+sum	15.362676	14.726909	14.761794	14.318087	14.338962	16.609221

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() 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	rowSums2	0.011012	0.0113750	0.0118727	0.0117670	0.0121055	0.022659
2	.rowSums	0.038415	0.0386555	0.0390685	0.0389325	0.0391145	0.043933
3	rowSums	0.040014	0.0402805	0.0407892	0.0405575	0.0410045	0.050169
4	apply+sum	0.140327	0.1423640	0.1474834	0.1432280	0.1459140	0.278656

	expr	min	lq	mean	median	uq	max
1	rowSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowSums	3.488467	3.398286	3.290620	3.308617	3.231135	1.938876
3	rowSums	3.633672	3.541143	3.435555	3.446715	3.387262	2.214087
4	apply+sum	12.743098	12.515517	12.422081	12.172007	12.053529	12.297807

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on integer+100x100 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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
2	rowSums2	11.012	11.375	11.87268	11.767	12.1055	22.659
1	colSums2	10.865	11.242	12.37640	11.790	12.1670	42.313

	expr	min	lq	mean	median	uq	max
2	rowSums2	1.0000000	1.0000000	1.000000	1.000000	1.00000	1.000000
1	colSums2	0.9866509	0.9883077	1.042427	1.001955	1.00508	1.867382

Figure: Benchmarking of colSums2() and rowSums2() 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 3199312 170.9    5709258 305.0  5709258 305.0
Vcells 6139930  46.9   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3199303 170.9    5709258 305.0  5709258 305.0
Vcells 6144968  46.9   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on integer+1000x10 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	.colSums	0.009446	0.0097855	0.0101568	0.0100190	0.0104385	0.014322
3	colSums	0.010809	0.0112885	0.0120771	0.0117345	0.0124155	0.025192
1	colSums2	0.012286	0.0126580	0.0131754	0.0130460	0.0134280	0.025906
4	apply+sum	0.068304	0.0723625	0.0743246	0.0731125	0.0739570	0.159002

	expr	min	lq	mean	median	uq	max
2	.colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colSums	1.144294	1.153595	1.189057	1.171225	1.189395	1.758972
1	colSums2	1.300656	1.293547	1.297198	1.302126	1.286392	1.808826
4	apply+sum	7.230997	7.394870	7.317678	7.297385	7.085022	11.101941

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() 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	rowSums2	0.012446	0.0128310	0.0134742	0.0133635	0.0136605	0.028865
4	apply+sum	0.068625	0.0722675	0.0738879	0.0728035	0.0740535	0.134151
2	.rowSums	0.136008	0.1363105	0.1367446	0.1365245	0.1367660	0.141067
3	rowSums	0.137386	0.1379325	0.1391066	0.1384995	0.1387965	0.179794

	expr	min	lq	mean	median	uq	max
1	rowSums2	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
4	apply+sum	5.51382	5.632258	5.483638	5.447936	5.420995	4.647532
2	.rowSums	10.92785	10.623529	10.148591	10.216223	10.011786	4.887130
3	rowSums	11.03857	10.749942	10.323894	10.364014	10.160426	6.228789

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on integer+1000x10 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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	colSums2	12.286	12.658	13.17544	13.0460	13.4280	25.906
2	rowSums2	12.446	12.831	13.47424	13.3635	13.6605	28.865

	expr	min	lq	mean	median	uq	max
1	colSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums2	1.013023	1.013667	1.022679	1.024337	1.017315	1.114221

Figure: Benchmarking of colSums2() and rowSums2() 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 3199548 170.9    5709258 305.0  5709258 305.0
Vcells 6140754  46.9   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3199539 170.9    5709258 305.0  5709258 305.0
Vcells 6145792  46.9   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on integer+10x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	.colSums	0.008985	0.0094030	0.0100757	0.0097305	0.0102510	0.021481
1	colSums2	0.009497	0.0098950	0.0107956	0.0104910	0.0107835	0.041372
3	colSums	0.010312	0.0111085	0.0120459	0.0118685	0.0126520	0.021115
4	apply+sum	0.790246	0.8194405	0.8353282	0.8273970	0.8535335	0.913296

	expr	min	lq	mean	median	uq	max
2	.colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colSums2	1.056984	1.052324	1.071458	1.078156	1.051946	1.9259811
3	colSums	1.147691	1.181378	1.195545	1.219721	1.234221	0.9829617
4	apply+sum	87.951697	87.146709	82.905562	85.031293	83.263438	42.5164564

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() 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	rowSums2	0.009553	0.0102075	0.0118775	0.0109715	0.0115800	0.020964
2	.rowSums	0.029425	0.0301750	0.0322925	0.0308090	0.0318405	0.050280
3	rowSums	0.031115	0.0321655	0.0361350	0.0331650	0.0347210	0.074309
4	apply+sum	0.787927	0.8261915	0.8764775	0.8407450	0.8615420	1.556184

	expr	min	lq	mean	median	uq	max
1	rowSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowSums	3.080184	2.956160	2.718799	2.808094	2.749611	2.398397
3	rowSums	3.257092	3.151163	3.042305	3.022832	2.998359	3.544600
4	apply+sum	82.479535	80.939652	73.793094	76.629905	74.399136	74.231254

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on integer+10x1000 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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
1	colSums2	9.497	9.8950	10.79565	10.4910	10.7835	41.372
2	rowSums2	9.553	10.2075	11.87750	10.9715	11.5800	20.964

	expr	min	lq	mean	median	uq	max
1	colSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowSums2	1.005897	1.031582	1.100212	1.045801	1.073863	0.5067195

Figure: Benchmarking of colSums2() and rowSums2() 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 3199764 170.9    5709258 305.0  5709258 305.0
Vcells 6141335  46.9   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3199758 170.9    5709258 305.0  5709258 305.0
Vcells 6191378  47.3   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on integer+100x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	.colSums	0.075066	0.0760865	0.0774762	0.0773605	0.0783695	0.081952
3	colSums	0.076752	0.0789050	0.0803119	0.0797965	0.0816615	0.095806
1	colSums2	0.097593	0.0996930	0.1011929	0.1005225	0.1022280	0.116382
4	apply+sum	1.260123	1.2886950	1.3070798	1.2987775	1.3199810	1.644839

	expr	min	lq	mean	median	uq	max
2	.colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colSums	1.022460	1.037043	1.036601	1.031489	1.042006	1.169050
1	colSums2	1.300096	1.310259	1.306115	1.299403	1.304436	1.420124
4	apply+sum	16.786868	16.937236	16.870718	16.788639	16.843045	20.070761

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() 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	rowSums2	0.103800	0.1053270	0.1092996	0.1071950	0.1096705	0.148805
2	.rowSums	0.271360	0.2784505	0.2844060	0.2793865	0.2869725	0.333830
3	rowSums	0.273222	0.2805355	0.2893872	0.2838145	0.2966940	0.335396
4	apply+sum	1.268050	1.3040830	1.3567052	1.3238910	1.3614360	1.864570

	expr	min	lq	mean	median	uq	max
1	rowSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowSums	2.614258	2.643676	2.602078	2.606339	2.616679	2.243406
3	rowSums	2.632197	2.663472	2.647652	2.647647	2.705322	2.253930
4	apply+sum	12.216281	12.381279	12.412723	12.350305	12.413876	12.530291

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on integer+100x1000 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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	colSums2	97.593	99.693	101.1929	100.5225	102.2280	116.382
2	rowSums2	103.800	105.327	109.2996	107.1950	109.6705	148.805

	expr	min	lq	mean	median	uq	max
1	colSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums2	1.063601	1.056514	1.080111	1.066378	1.072803	1.278591

Figure: Benchmarking of colSums2() and rowSums2() 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 3199986 170.9    5709258 305.0  5709258 305.0
Vcells 6142029  46.9   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3199980 170.9    5709258 305.0  5709258 305.0
Vcells 6192072  47.3   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on integer+1000x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	.colSums	0.079430	0.0798290	0.0813971	0.0801585	0.0809085	0.112326
3	colSums	0.081147	0.0818435	0.0842066	0.0825235	0.0837060	0.156258
1	colSums2	0.104633	0.1050895	0.1062148	0.1054080	0.1060155	0.120491
4	apply+sum	0.543041	0.5528160	0.5673520	0.5599640	0.5732780	0.678743

	expr	min	lq	mean	median	uq	max
2	.colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colSums	1.021616	1.025235	1.034516	1.029504	1.034576	1.391112
1	colSums2	1.317298	1.316433	1.304897	1.314995	1.310313	1.072690
4	apply+sum	6.836724	6.925002	6.970173	6.985710	7.085510	6.042617

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() 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	rowSums2	0.113665	0.115278	0.1194711	0.1168145	0.1192320	0.212273
2	.rowSums	0.371720	0.372579	0.3741557	0.3729915	0.3743710	0.383320
3	rowSums	0.373165	0.374549	0.3768583	0.3751790	0.3763885	0.419846
4	apply+sum	0.565210	0.582445	0.5967722	0.5906430	0.6047140	0.704273

	expr	min	lq	mean	median	uq	max
1	rowSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowSums	3.270312	3.232004	3.131767	3.193024	3.139853	1.805788
3	rowSums	3.283025	3.249093	3.154388	3.211750	3.156774	1.977859
4	apply+sum	4.972595	5.052525	4.995116	5.056247	5.071743	3.317770

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on integer+1000x100 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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	colSums2	104.633	105.0895	106.2148	105.4080	106.0155	120.491
2	rowSums2	113.665	115.2780	119.4711	116.8145	119.2320	212.273

	expr	min	lq	mean	median	uq	max
1	colSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums2	1.086321	1.096951	1.124806	1.108213	1.124666	1.761733

Figure: Benchmarking of colSums2() and rowSums2() 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 3200236 171.0    5709258 305.0  5709258 305.0
Vcells 6257960  47.8   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3200218 171.0    5709258 305.0  5709258 305.0
Vcells 6258083  47.8   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() 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	colSums2	0.001125	0.0014110	0.0018459	0.0018445	0.0019880	0.010989
2	.colSums	0.001679	0.0019315	0.0022807	0.0021570	0.0024595	0.010393
3	colSums	0.002981	0.0033455	0.0039332	0.0037940	0.0043160	0.012126
4	apply+sum	0.020129	0.0214210	0.0226355	0.0217665	0.0222410	0.084903

	expr	min	lq	mean	median	uq	max
1	colSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	.colSums	1.492444	1.368887	1.235563	1.169423	1.237173	0.9457639
3	colSums	2.649778	2.371014	2.130780	2.056926	2.171026	1.1034671
4	apply+sum	17.892444	15.181432	12.262471	11.800759	11.187626	7.7261807

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() on double+10x10 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	.rowSums	0.001290	0.0014620	0.0017062	0.0016700	0.0017970	0.007053
1	rowSums2	0.001122	0.0013765	0.0017429	0.0016945	0.0018500	0.011662
3	rowSums	0.002451	0.0028405	0.0036099	0.0032490	0.0034915	0.043073
4	apply+sum	0.019615	0.0205245	0.0215753	0.0208405	0.0213040	0.077098

	expr	min	lq	mean	median	uq	max
2	.rowSums	1.0000000	1.0000000	1.000000	1.000000	1.000000	1.000000
1	rowSums2	0.8697674	0.9415185	1.021497	1.014671	1.029494	1.653481
3	rowSums	1.9000000	1.9428865	2.115704	1.945509	1.942961	6.107047
4	apply+sum	15.2054264	14.0386457	12.644882	12.479341	11.855314	10.931235

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on double+10x10 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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	rowSums2	1.122	1.3765	1.74293	1.6945	1.850	11.662
1	colSums2	1.125	1.4110	1.84592	1.8445	1.988	10.989

	expr	min	lq	mean	median	uq	max
2	rowSums2	1.000000	1.000000	1.00000	1.000000	1.000000	1.0000000
1	colSums2	1.002674	1.025064	1.05909	1.088522	1.074595	0.9422912

Figure: Benchmarking of colSums2() and rowSums2() 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 3200456 171.0    5709258 305.0  5709258 305.0
Vcells 6258975  47.8   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3200447 171.0    5709258 305.0  5709258 305.0
Vcells 6269013  47.9   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on double+100x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	.colSums	0.008481	0.0094420	0.0132974	0.0100500	0.0154795	0.086337
1	colSums2	0.011157	0.0122005	0.0158498	0.0129775	0.0166145	0.114645
3	colSums	0.009952	0.0115635	0.0157287	0.0130725	0.0189935	0.045543
4	apply+sum	0.188215	0.2021885	0.2483000	0.2126865	0.2633415	0.930026

	expr	min	lq	mean	median	uq	max
2	.colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colSums2	1.315529	1.292152	1.191945	1.291294	1.073323	1.3278780
3	colSums	1.173447	1.224688	1.182834	1.300746	1.227010	0.5275027
4	apply+sum	22.192548	21.413736	18.672754	21.162836	17.012274	10.7720444

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() 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	rowSums2	0.011288	0.0117185	0.0124961	0.0121915	0.0125340	0.022394
2	.rowSums	0.019842	0.0202185	0.0215277	0.0203770	0.0206760	0.058601
3	rowSums	0.021223	0.0217120	0.0234411	0.0223080	0.0227835	0.069166
4	apply+sum	0.144447	0.1476880	0.1602648	0.1501135	0.1580005	0.296495

	expr	min	lq	mean	median	uq	max
1	rowSums2	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000
2	.rowSums	1.757796	1.725349	1.722759	1.67141	1.649593	2.616817
3	rowSums	1.880138	1.852797	1.875876	1.82980	1.817736	3.088595
4	apply+sum	12.796510	12.602978	12.825207	12.31296	12.605752	13.239930

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on double+100x100 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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	rowSums2	11.288	11.7185	12.49608	12.1915	12.5340	22.394
1	colSums2	11.157	12.2005	15.84983	12.9775	16.6145	114.645

	expr	min	lq	mean	median	uq	max
2	rowSums2	1.0000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colSums2	0.9883948	1.041132	1.268384	1.064471	1.325554	5.119452

Figure: Benchmarking of colSums2() and rowSums2() 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 3200682 171.0    5709258 305.0  5709258 305.0
Vcells 6260076  47.8   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3200673 171.0    5709258 305.0  5709258 305.0
Vcells 6270114  47.9   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on double+1000x10 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	.colSums	0.009503	0.0098125	0.0105194	0.010159	0.0104955	0.035338
3	colSums	0.010987	0.0114515	0.0123062	0.012100	0.0126045	0.032284
1	colSums2	0.012387	0.0129675	0.0134194	0.013206	0.0134815	0.029126
4	apply+sum	0.116068	0.1181620	0.1211660	0.119921	0.1214535	0.215851

	expr	min	lq	mean	median	uq	max
2	.colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
3	colSums	1.156161	1.167032	1.169860	1.191062	1.200943	0.9135775
1	colSums2	1.303483	1.321529	1.275679	1.299931	1.284503	0.8242119
4	apply+sum	12.213827	12.041987	11.518340	11.804410	11.571959	6.1081838

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() 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	rowSums2	0.012830	0.0136550	0.0165813	0.0142095	0.0168490	0.047365
2	.rowSums	0.020297	0.0207485	0.0244345	0.0213450	0.0229405	0.068766
3	rowSums	0.021788	0.0227020	0.0259081	0.0237630	0.0249980	0.068706
4	apply+sum	0.076028	0.0779730	0.1006974	0.0808515	0.0965640	0.340523

	expr	min	lq	mean	median	uq	max
1	rowSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowSums	1.581995	1.519480	1.473618	1.502164	1.361535	1.451831
3	rowSums	1.698207	1.662541	1.562492	1.672332	1.483649	1.450565
4	apply+sum	5.925799	5.710216	6.072958	5.689961	5.731141	7.189338

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on double+1000x10 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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	colSums2	12.387	12.9675	13.41938	13.2060	13.4815	29.126
2	rowSums2	12.830	13.6550	16.58128	14.2095	16.8490	47.365

	expr	min	lq	mean	median	uq	max
1	colSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
2	rowSums2	1.035763	1.053017	1.235622	1.075988	1.249787	1.62621

Figure: Benchmarking of colSums2() and rowSums2() 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 3200903 171.0    5709258 305.0  5709258 305.0
Vcells 6260222  47.8   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3200894 171.0    5709258 305.0  5709258 305.0
Vcells 6270260  47.9   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on double+10x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	.colSums	0.009544	0.0101400	0.0112099	0.0104835	0.0111480	0.041096
3	colSums	0.011269	0.0121530	0.0136204	0.0127455	0.0141990	0.029504
1	colSums2	0.012166	0.0131125	0.0141914	0.0136505	0.0143675	0.029272
4	apply+sum	0.795365	0.8482970	0.8751510	0.8732355	0.8967020	1.100357

	expr	min	lq	mean	median	uq	max
2	.colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
3	colSums	1.180742	1.198521	1.215033	1.215768	1.273681	0.7179288
1	colSums2	1.274728	1.293146	1.265970	1.302094	1.288796	0.7122834
4	apply+sum	83.336651	83.658481	78.069473	83.296180	80.436132	26.7752823

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() 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	rowSums2	0.012403	0.013533	0.0149897	0.0140480	0.0149865	0.027415
2	.rowSums	0.018839	0.019744	0.0205335	0.0203010	0.0208505	0.027395
3	rowSums	0.019771	0.021400	0.0228646	0.0221615	0.0233765	0.036762
4	apply+sum	0.790190	0.864529	0.8862299	0.8822150	0.9011230	1.170931

	expr	min	lq	mean	median	uq	max
1	rowSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	.rowSums	1.518907	1.458952	1.369838	1.445117	1.391285	0.9992705
3	rowSums	1.594050	1.581320	1.525353	1.577556	1.559837	1.3409447
4	apply+sum	63.709586	63.883027	59.122432	62.800043	60.128983	42.7113259

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on double+10x1000 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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	colSums2	12.166	13.1125	14.19140	13.6505	14.3675	29.272
2	rowSums2	12.403	13.5330	14.98974	14.0480	14.9865	27.415

	expr	min	lq	mean	median	uq	max
1	colSums2	1.000000	1.000000	1.000000	1.00000	1.000000	1.0000000
2	rowSums2	1.019481	1.032069	1.056255	1.02912	1.043083	0.9365605

Figure: Benchmarking of colSums2() and rowSums2() 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 3201123 171.0    5709258 305.0  5709258 305.0
Vcells 6261502  47.8   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3201117 171.0    5709258 305.0  5709258 305.0
Vcells 6361545  48.6   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on double+100x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	.colSums	0.065601	0.0736915	0.0872300	0.078597	0.0878205	0.173754
3	colSums	0.068018	0.0759150	0.0879334	0.079849	0.0918110	0.157878
1	colSums2	0.099009	0.1106685	0.1192029	0.114086	0.1225600	0.180922
4	apply+sum	1.708269	1.9040275	2.3091431	2.021267	2.1891150	15.793835

	expr	min	lq	mean	median	uq	max
2	.colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
3	colSums	1.036844	1.030173	1.008064	1.015929	1.045439	0.9086294
1	colSums2	1.509260	1.501781	1.366536	1.451531	1.395574	1.0412537
4	apply+sum	26.040289	25.837817	26.471895	25.716847	24.927152	90.8976772

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() 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	rowSums2	0.105944	0.1089645	0.1166479	0.1118535	0.1148025	0.232180
2	.rowSums	0.171737	0.1853725	0.1955227	0.1904550	0.1948960	0.276480
3	rowSums	0.179382	0.1836980	0.1924062	0.1909760	0.1927005	0.261175
4	apply+sum	1.309472	1.3688630	1.5825294	1.3847540	1.4315730	15.149573

	expr	min	lq	mean	median	uq	max
1	rowSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowSums	1.621017	1.701219	1.676179	1.702718	1.697663	1.190800
3	rowSums	1.693178	1.685852	1.649461	1.707376	1.678539	1.124882
4	apply+sum	12.360039	12.562468	13.566721	12.380069	12.469876	65.249259

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on double+100x1000 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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	rowSums2	105.944	108.9645	116.6479	111.8535	114.8025	232.180
1	colSums2	99.009	110.6685	119.2029	114.0860	122.5600	180.922

	expr	min	lq	mean	median	uq	max
2	rowSums2	1.0000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colSums2	0.9345409	1.015638	1.021904	1.019959	1.067573	0.7792316

Figure: Benchmarking of colSums2() and rowSums2() 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 3201358 171.0    5709258 305.0  5709258 305.0
Vcells 6261664  47.8   22343563 170.5 56666022 432.4
> colStats <- microbenchmark(colSums2 = colSums2(X, na.rm = FALSE), .colSums = .colSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), colSums = colSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")
> X <- t(X)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 3201349 171.0    5709258 305.0  5709258 305.0
Vcells 6361702  48.6   22343563 170.5 56666022 432.4
> rowStats <- microbenchmark(rowSums2 = rowSums2(X, na.rm = FALSE), .rowSums = .rowSums(X, m = nrow(X), 
+     n = ncol(X), na.rm = FALSE), rowSums = rowSums(X, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, 
+     FUN = sum, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on double+1000x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	.colSums	0.077325	0.0781145	0.0829543	0.0792585	0.086708	0.107558
3	colSums	0.079287	0.0801690	0.0853740	0.0817965	0.089144	0.115481
1	colSums2	0.104967	0.1055385	0.1126490	0.1071660	0.120821	0.143122
4	apply+sum	0.591068	0.6022620	0.7025576	0.6200985	0.652509	7.226098

	expr	min	lq	mean	median	uq	max
2	.colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colSums	1.025373	1.026301	1.029168	1.032022	1.028094	1.073663
1	colSums2	1.357478	1.351074	1.357964	1.352107	1.393424	1.330650
4	apply+sum	7.643944	7.709990	8.469209	7.823748	7.525361	67.183269

Table: Benchmarking of rowSums2(), .rowSums(), rowSums() and apply+sum() 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	rowSums2	0.108654	0.1114860	0.1175326	0.1143585	0.1189145	0.167381
2	.rowSums	0.187866	0.1929845	0.2002140	0.1954265	0.2032060	0.244465
3	rowSums	0.189519	0.1935800	0.2028186	0.1980415	0.2049410	0.279904
4	apply+sum	0.626604	0.6531920	0.7608880	0.6655130	0.7198745	7.481785

	expr	min	lq	mean	median	uq	max
1	rowSums2	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	.rowSums	1.729030	1.731020	1.703477	1.708894	1.708841	1.460530
3	rowSums	1.744243	1.736361	1.725638	1.731760	1.723432	1.672257
4	apply+sum	5.766967	5.858960	6.473849	5.819532	6.053715	44.699129

Figure: Benchmarking of colSums2(), .colSums(), colSums() and apply+sum() on double+1000x100 data as well as rowSums2(), .rowSums(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colSums2() and rowSums2() 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	colSums2	104.967	105.5385	112.6489	107.1660	120.8210	143.122
2	rowSums2	108.654	111.4860	117.5326	114.3585	118.9145	167.381

	expr	min	lq	mean	median	uq	max
1	colSums2	1.000000	1.000000	1.000000	1.000000	1.0000000	1.000000
2	rowSums2	1.035125	1.056354	1.043352	1.067116	0.9842205	1.169499

Figure: Benchmarking of colSums2() and rowSums2() on double+1000x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

Appendix

Session information

R version 3.6.1 Patched (2019-08-27 r77078)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.3 LTS

Matrix products: default
BLAS:   /home/hb/software/R-devel/R-3-6-branch/lib/R/lib/libRblas.so
LAPACK: /home/hb/software/R-devel/R-3-6-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_1.4-6    matrixStats_0.55.0-9000 ggplot2_3.2.1          
[4] knitr_1.24              R.devices_2.16.0        R.utils_2.9.0          
[7] R.oo_1.22.0             R.methodsS3_1.7.1       history_0.0.0-9002     

loaded via a namespace (and not attached):
 [1] Biobase_2.45.0       bit64_0.9-7          splines_3.6.1       
 [4] network_1.15         assertthat_0.2.1     highr_0.8           
 [7] stats4_3.6.1         blob_1.2.0           robustbase_0.93-5   
[10] pillar_1.4.2         RSQLite_2.1.2        backports_1.1.4     
[13] lattice_0.20-38      glue_1.3.1           digest_0.6.20       
[16] colorspace_1.4-1     sandwich_2.5-1       Matrix_1.2-17       
[19] XML_3.98-1.20        lpSolve_5.6.13.3     pkgconfig_2.0.2     
[22] genefilter_1.66.0    purrr_0.3.2          ergm_3.10.4         
[25] xtable_1.8-4         mvtnorm_1.0-11       scales_1.0.0        
[28] tibble_2.1.3         annotate_1.62.0      IRanges_2.18.2      
[31] TH.data_1.0-10       withr_2.1.2          BiocGenerics_0.30.0 
[34] lazyeval_0.2.2       mime_0.7             survival_2.44-1.1   
[37] magrittr_1.5         crayon_1.3.4         statnet.common_4.3.0
[40] memoise_1.1.0        laeken_0.5.0         R.cache_0.13.0      
[43] MASS_7.3-51.4        R.rsp_0.43.1         tools_3.6.1         
[46] multcomp_1.4-10      S4Vectors_0.22.1     trust_0.1-7         
[49] munsell_0.5.0        AnnotationDbi_1.46.1 compiler_3.6.1      
[52] rlang_0.4.0          grid_3.6.1           RCurl_1.95-4.12     
[55] cwhmisc_6.6          rappdirs_0.3.1       labeling_0.3        
[58] bitops_1.0-6         base64enc_0.1-3      boot_1.3-23         
[61] gtable_0.3.0         codetools_0.2-16     DBI_1.0.0           
[64] markdown_1.1         R6_2.4.0             zoo_1.8-6           
[67] dplyr_0.8.3          bit_1.1-14           zeallot_0.1.0       
[70] parallel_3.6.1       Rcpp_1.0.2           vctrs_0.2.0         
[73] DEoptimR_1.0-8       tidyselect_0.2.5     xfun_0.9            
[76] coda_0.19-3         

Total processing time was 27.73 secs.

Reproducibility

To reproduce this report, do:

html <- matrixStats:::benchmark('colSums2')

---

Copyright Henrik Bengtsson. Last updated on 2019-09-10 20:53:36 (-0700 UTC). Powered by RSP.

<script> var link = document.createElement('link'); link.rel = 'icon'; link.href = "data:image/png;base64,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" document.getElementsByTagName('head')[0].appendChild(link); </script>