colRowCounts

matrixStats: Benchmark report

---

colCounts() and rowCounts() benchmarks

This report benchmark the performance of colCounts() and rowCounts() against alternative methods.

Alternative methods

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

Data type "logical"

Data

> rmatrix <- function(nrow, ncol, mode = c("logical", "double", "integer", "index"), range = c(-100, 
+     +100), naProb = 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 (naProb > 0) 
+         X[sample(n, size = naProb * 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 matrix

> X <- data[["10x10"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), colSums() and apply+sum() on logical+10x10 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

expr	min	lq	mean	median	uq	max	neval
colCounts	0.005938	0.0067165	0.0080754	0.0082460	0.0088910	0.025245	100
colSums	0.008057	0.0088875	0.0099427	0.0099305	0.0104160	0.020142	100
apply+sum	0.059256	0.0607415	0.0627614	0.0616695	0.0626665	0.118169	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
colSums	1.356854	1.323234	1.231234	1.204281	1.171522	0.797861	100
apply+sum	9.979117	9.043624	7.771964	7.478717	7.048307	4.680887	100
Table: Benchmarking of rowCounts(), rowSums() and apply+sum() on logical+10x10 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr	min	lq	mean	median	uq	max	neval
rowCounts	0.006094	0.0067220	0.0081886	0.0074820	0.0089930	0.026223	100
rowSums	0.007512	0.0083075	0.0094609	0.0097285	0.0103050	0.016426	100
apply+sum	0.059751	0.0616895	0.0633502	0.0623910	0.0633815	0.118581	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000	100
rowSums	1.232688	1.235867	1.155373	1.300254	1.145891	0.6263967	100
apply+sum	9.804890	9.177254	7.736375	8.338813	7.047871	4.5220227	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on logical+10x10 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() on logical+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	neval
2	rowCounts	6.094	6.7220	8.18861	7.482	8.993	26.223	100
1	colCounts	5.938	6.7165	8.07536	8.246	8.891	25.245	100

	expr	min	lq	mean	median	uq	max	neval
2	rowCounts	1.0000000	1.0000000	1.0000000	1.000000	1.0000000	1.0000000	100
1	colCounts	0.9744011	0.9991818	0.9861698	1.102112	0.9886578	0.9627045	100
Figure: Benchmarking of colCounts() and rowCounts() on logical+10x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x100 matrix

> X <- data[["100x100"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

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

expr	min	lq	mean	median	uq	max	neval
colCounts	0.036579	0.0395410	0.0432680	0.0434065	0.0451985	0.070758	100
colSums	0.088357	0.0906760	0.1266016	0.0920435	0.0936095	1.763543	100
apply+sum	0.497600	0.5070345	0.5219988	0.5176640	0.5271675	0.702221	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000	100
colSums	2.415512	2.293215	2.925988	2.12050	2.071075	24.923585	100
apply+sum	13.603434	12.823006	12.064315	11.92596	11.663385	9.924263	100
Table: Benchmarking of rowCounts(), rowSums() and apply+sum() on logical+100x100 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr	min	lq	mean	median	uq	max	neval
rowCounts	0.041875	0.0452355	0.0484104	0.0475000	0.0500030	0.076317	100
rowSums	0.104313	0.1055440	0.1071632	0.1063995	0.1072025	0.126581	100
apply+sum	0.497969	0.5036205	0.5462540	0.5080855	0.5176900	2.208795	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000	100
rowSums	2.491057	2.333212	2.213638	2.23999	2.143921	1.658621	100
apply+sum	11.891797	11.133302	11.283803	10.69654	10.353179	28.942372	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on logical+100x100 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() on logical+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	neval
colCounts	36.579	39.5410	43.26800	43.4065	45.1985	70.758	100
rowCounts	41.875	45.2355	48.41045	47.5000	50.0030	76.317	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowCounts	1.144783	1.144015	1.118851	1.094306	1.106298	1.078564	100
Figure: Benchmarking of colCounts() and rowCounts() on logical+100x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x10 matrix

> X <- data[["1000x10"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

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

expr	min	lq	mean	median	uq	max	neval
colCounts	0.035913	0.038243	0.0408713	0.0402695	0.0428585	0.061141	100
colSums	0.087694	0.088412	0.0901938	0.0891375	0.0898490	0.111044	100
apply+sum	0.222391	0.224935	0.2481667	0.2266585	0.2281415	1.254100	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
colSums	2.441846	2.311848	2.206775	2.213524	2.096410	1.816195	100
apply+sum	6.192493	5.881730	6.071908	5.628540	5.323133	20.511604	100
Table: Benchmarking of rowCounts(), rowSums() and apply+sum() on logical+1000x10 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr	min	lq	mean	median	uq	max	neval
rowCounts	0.042930	0.0471360	0.0497080	0.0487460	0.052275	0.079379	100
rowSums	0.105768	0.1067035	0.1089648	0.1075655	0.108248	0.138420	100
apply+sum	0.222589	0.2263275	0.2503524	0.2281120	0.230878	1.232702	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowSums	2.463732	2.263737	2.192101	2.206653	2.070741	1.743786	100
apply+sum	5.184929	4.801585	5.036466	4.679604	4.416605	15.529321	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on logical+1000x10 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() on logical+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	neval
colCounts	35.913	38.243	40.87129	40.2695	42.8585	61.141	100
rowCounts	42.930	47.136	49.70795	48.7460	52.2750	79.379	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowCounts	1.195389	1.232539	1.216207	1.210494	1.219711	1.298294	100
Figure: Benchmarking of colCounts() and rowCounts() on logical+1000x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

10x1000 matrix

> X <- data[["10x1000"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

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

expr	min	lq	mean	median	uq	max	neval
colCounts	0.039266	0.0425685	0.0499366	0.0475745	0.0562715	0.071372	100
colSums	0.089403	0.0914930	0.0947872	0.0934355	0.0973960	0.116637	100
apply+sum	3.119022	3.1601980	3.3255419	3.2152245	3.3158165	5.537935	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
colSums	2.276855	2.149312	1.898151	1.963983	1.730823	1.634212	100
apply+sum	79.433148	74.237946	66.595294	67.582938	58.925326	77.592543	100
Table: Benchmarking of rowCounts(), rowSums() and apply+sum() on logical+10x1000 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr	min	lq	mean	median	uq	max	neval
rowCounts	0.042912	0.0465335	0.0526830	0.0518285	0.0582990	0.073202	100
rowSums	0.105237	0.1068495	0.1102865	0.1080705	0.1120115	0.128831	100
apply+sum	3.091413	3.1668350	3.3342174	3.2436465	3.3429070	4.604021	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowSums	2.452391	2.296184	2.093398	2.085156	1.921328	1.759938	100
apply+sum	72.040758	68.054950	63.288285	62.584225	57.340726	62.894743	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on logical+10x1000 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() on logical+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	neval
colCounts	39.266	42.5685	49.93659	47.5745	56.2715	71.372	100
rowCounts	42.912	46.5335	52.68301	51.8285	58.2990	73.202	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000	100
rowCounts	1.092854	1.093144	1.054998	1.089418	1.036031	1.02564	100
Figure: Benchmarking of colCounts() and rowCounts() on logical+10x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x1000 matrix

> X <- data[["100x1000"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

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

expr	min	lq	mean	median	uq	max	neval
colCounts	0.355991	0.3622200	0.3782094	0.3772685	0.3867415	0.453541	100
colSums	0.800710	0.8074205	1.4562706	0.8257325	0.8781550	43.417423	100
apply+sum	4.697739	4.7985110	5.3041886	4.8985055	6.0378795	7.193121	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000	100
colSums	2.249242	2.229089	3.850435	2.188713	2.270651	95.72987	100
apply+sum	13.196230	13.247504	14.024477	12.984136	15.612184	15.85991	100
Table: Benchmarking of rowCounts(), rowSums() and apply+sum() on logical+100x1000 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr	min	lq	mean	median	uq	max	neval
rowCounts	0.408146	0.4142640	0.4340272	0.4300310	0.4379550	0.625429	100
rowSums	0.961649	0.9693825	1.1144560	0.9791385	0.9937745	2.859870	100
apply+sum	4.747159	4.8462455	5.8856342	4.9676860	6.3587870	47.848796	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.00000	1.000000	1.00000	1.000000	1.000000	1.000000	100
rowSums	2.35614	2.340011	2.56771	2.276902	2.269125	4.572653	100
apply+sum	11.63103	11.698447	13.56052	11.551925	14.519270	76.505560	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on logical+100x1000 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() on logical+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	neval
colCounts	355.991	362.220	378.2094	377.2685	386.7415	453.541	100
rowCounts	408.146	414.264	434.0272	430.0310	437.9550	625.429	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowCounts	1.146507	1.143681	1.147585	1.139854	1.132423	1.378991	100
Figure: Benchmarking of colCounts() and rowCounts() on logical+100x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x100 matrix

> X <- data[["1000x100"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

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

expr	min	lq	mean	median	uq	max	neval
colCounts	0.350349	0.3558935	0.3673879	0.3658435	0.3735365	0.480174	100
colSums	0.794660	0.7980335	0.9464144	0.8052210	0.8212955	1.832601	100
apply+sum	1.962195	1.9924685	2.2194187	2.0185175	2.1264470	3.155022	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
colSums	2.268195	2.242338	2.576063	2.200998	2.198702	3.816535	100
apply+sum	5.600687	5.598497	6.041077	5.517434	5.692742	6.570581	100
Table: Benchmarking of rowCounts(), rowSums() and apply+sum() on logical+1000x100 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

expr	min	lq	mean	median	uq	max	neval
rowCounts	0.414670	0.419928	0.432341	0.428901	0.437181	0.489343	100
rowSums	0.967677	0.971479	1.522248	0.975393	1.001624	42.870240	100
apply+sum	1.983647	2.016538	2.281843	2.044653	2.720711	3.200574	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowSums	2.333607	2.313442	3.520943	2.274168	2.291098	87.607752	100
apply+sum	4.783676	4.802104	5.277877	4.767191	6.223305	6.540553	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on logical+1000x100 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() on logical+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	neval
colCounts	350.349	355.8935	367.3879	365.8435	373.5365	480.174	100
rowCounts	414.670	419.9280	432.3410	428.9010	437.1810	489.343	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowCounts	1.183591	1.179926	1.176797	1.172362	1.170384	1.019095	100
Figure: Benchmarking of colCounts() and rowCounts() on logical+1000x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

Data type "integer"

Data

> rmatrix <- function(nrow, ncol, mode = c("logical", "double", "integer", "index"), range = c(-100, 
+     +100), naProb = 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 (naProb > 0) 
+         X[sample(n, size = naProb * 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 matrix

> X <- data[["10x10"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.005682	0.0064400	0.0077954	0.0076020	0.0087730	0.023538	100
colSums	0.007743	0.0088525	0.0097964	0.0098845	0.0103940	0.019366	100
apply+sum	0.059633	0.0611430	0.0631214	0.0618195	0.0628015	0.115958	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000	100
colSums	1.362724	1.374612	1.256690	1.300250	1.184772	0.8227547	100
apply+sum	10.495072	9.494255	8.097273	8.132005	7.158498	4.9264169	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.005950	0.0066280	0.0078493	0.0076265	0.0087445	0.024736	100
rowSums	0.007211	0.0082150	0.0094199	0.0094705	0.0100245	0.018920	100
apply+sum	0.058600	0.0607045	0.0626365	0.0614350	0.0626330	0.119455	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000	100
rowSums	1.211933	1.239439	1.200097	1.241788	1.146378	0.7648771	100
apply+sum	9.848740	9.158796	7.979906	8.055464	7.162559	4.8291963	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on integer+10x10 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
colCounts	5.682	6.440	7.79539	7.6020	8.7730	23.538	100
rowCounts	5.950	6.628	7.84928	7.6265	8.7445	24.736	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.0000000	1.000000	100
rowCounts	1.047166	1.029192	1.006913	1.003223	0.9967514	1.050896	100
Figure: Benchmarking of colCounts() and rowCounts() on integer+10x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x100 matrix

> X <- data[["100x100"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.015487	0.0162980	0.0186450	0.0186025	0.0194460	0.043599	100
colSums	0.088638	0.0895070	0.1236637	0.0902290	0.0908410	1.746604	100
apply+sum	0.499101	0.5017345	0.5106780	0.5053695	0.5121805	0.577766	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.00000	1.000000	1.00000	100
colSums	5.723381	5.491901	6.632552	4.85037	4.671449	40.06064	100
apply+sum	32.227094	30.785035	27.389602	27.16675	26.338604	13.25182	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.018273	0.0189520	0.0210456	0.0206605	0.0220100	0.036831	100
rowSums	0.104563	0.1054375	0.1079447	0.1061655	0.1068625	0.139020	100
apply+sum	0.495515	0.5023850	0.5441405	0.5059065	0.5161050	2.212074	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.000000	1.00000	1.000000	100
rowSums	5.722268	5.563397	5.129077	5.138574	4.85518	3.774538	100
apply+sum	27.117332	26.508284	25.855274	24.486653	23.44866	60.060112	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on integer+100x100 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
colCounts	15.487	16.298	18.64496	18.6025	19.446	43.599	100
rowCounts	18.273	18.952	21.04563	20.6605	22.010	36.831	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.00000	1.000000	1.0000000	100
rowCounts	1.179893	1.162842	1.128757	1.11063	1.131852	0.8447671	100
Figure: Benchmarking of colCounts() and rowCounts() on integer+100x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x10 matrix

> X <- data[["1000x10"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.014599	0.015304	0.0170788	0.0173265	0.0180870	0.033042	100
colSums	0.087615	0.088299	0.0895675	0.0889625	0.0894515	0.109900	100
apply+sum	0.222761	0.225360	0.2479239	0.2265370	0.2278000	1.236478	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000	100
colSums	6.001438	5.769668	5.244363	5.134476	4.945624	3.32607	100
apply+sum	15.258648	14.725562	14.516464	13.074597	12.594681	37.42140	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.018388	0.019063	0.0210177	0.0213135	0.0218810	0.046230	100
rowSums	0.105702	0.106704	0.1078388	0.1072280	0.1077245	0.137149	100
apply+sum	0.223146	0.225356	0.2480699	0.2264370	0.2287040	1.223689	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.00000	1.000000	1.00000	1.000000	1.000000	100
rowSums	5.748423	5.59744	5.130844	5.03099	4.923198	2.966667	100
apply+sum	12.135414	11.82164	11.802880	10.62411	10.452173	26.469587	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on integer+1000x10 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
colCounts	14.599	15.304	17.07881	17.3265	18.087	33.042	100
rowCounts	18.388	19.063	21.01774	21.3135	21.881	46.230	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000	100
rowCounts	1.259538	1.245622	1.230633	1.23011	1.209764	1.399128	100
Figure: Benchmarking of colCounts() and rowCounts() on integer+1000x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

10x1000 matrix

> X <- data[["10x1000"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.016090	0.017190	0.022674	0.020080	0.0261005	0.040684	100
colSums	0.089775	0.091518	0.095203	0.093436	0.0988845	0.120075	100
apply+sum	3.083524	3.161154	3.295713	3.209097	3.2908455	4.094655	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
colSums	5.579553	5.323909	4.198774	4.653187	3.788606	2.951406	100
apply+sum	191.642262	183.894968	145.352073	159.815612	126.083619	100.645340	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.017991	0.0193095	0.0250942	0.0237690	0.0295245	0.052373	100
rowSums	0.105131	0.1065075	0.1091586	0.1077885	0.1105690	0.131508	100
apply+sum	3.100086	3.1708005	3.3213949	3.2154925	3.3215920	4.163554	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowSums	5.843533	5.515808	4.349955	4.534835	3.744991	2.510988	100
apply+sum	172.313157	164.209353	132.357075	135.280933	112.502904	79.498100	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on integer+10x1000 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
colCounts	16.090	17.1900	22.6740	20.080	26.1005	40.684	100
rowCounts	17.991	19.3095	25.0942	23.769	29.5245	52.373	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowCounts	1.118148	1.123298	1.106739	1.183715	1.131185	1.287312	100
Figure: Benchmarking of colCounts() and rowCounts() on integer+10x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x1000 matrix

> X <- data[["100x1000"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.102628	0.1047720	0.1154164	0.1142385	0.1261055	0.135459	100
colSums	0.801134	0.8090535	1.0734457	0.8200350	0.8711975	2.691553	100
apply+sum	4.710071	4.7907695	5.5991978	4.8378280	5.0479450	47.196851	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000	100
colSums	7.806193	7.722039	9.300632	7.178272	6.908481	19.86987	100
apply+sum	45.894600	45.725666	48.513009	42.348490	40.029539	348.42167	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.124136	0.1270195	0.1359501	0.1363150	0.142851	0.155579	100
rowSums	0.961554	0.9696845	1.1350187	0.9757305	1.006057	2.769947	100
apply+sum	4.694080	4.8420400	5.4685971	4.9833810	6.396341	7.366662	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.00000	1.000000	1.00000	100
rowSums	7.745972	7.634139	8.348788	7.15791	7.042702	17.80412	100
apply+sum	37.814010	38.120446	40.225026	36.55783	44.776312	47.34998	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on integer+100x1000 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
colCounts	102.628	104.7720	115.4164	114.2385	126.1055	135.459	100
rowCounts	124.136	127.0195	135.9501	136.3150	142.8510	155.579	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.00000	1.000000	1.00000	1.000000	100
rowCounts	1.209572	1.212342	1.17791	1.193249	1.13279	1.148532	100
Figure: Benchmarking of colCounts() and rowCounts() on integer+100x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x100 matrix

> X <- data[["1000x100"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.095672	0.098738	0.1064356	0.1066160	0.111216	0.168510	100
colSums	0.794877	0.799742	1.3252402	0.8058195	0.834963	43.680496	100
apply+sum	1.962361	2.001254	2.2991458	2.0403005	2.882436	3.197454	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.00000	1.000000	1.00000	1.00000	100
colSums	8.308356	8.099637	12.45110	7.558148	7.50758	259.21605	100
apply+sum	20.511341	20.268331	21.60128	19.136907	25.91746	18.97486	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.129602	0.1312545	0.1378455	0.1356315	0.144935	0.158294	100
rowSums	0.967389	0.9709875	1.1009219	0.9746685	0.987880	1.985198	100
apply+sum	1.988194	2.0135090	2.2640658	2.0473515	2.456067	3.048246	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000	100
rowSums	7.464306	7.397746	7.986636	7.186151	6.816021	12.54121	100
apply+sum	15.340766	15.340495	16.424662	15.094956	16.945990	19.25686	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on integer+1000x100 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
colCounts	95.672	98.7380	106.4356	106.6160	111.216	168.510	100
rowCounts	129.602	131.2545	137.8455	135.6315	144.935	158.294	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.00000	1.000000	1.0000000	100
rowCounts	1.354649	1.329321	1.295107	1.27215	1.303185	0.9393745	100
Figure: Benchmarking of colCounts() and rowCounts() 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), naProb = 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 (naProb > 0) 
+         X[sample(n, size = naProb * 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 matrix

> X <- data[["10x10"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.005936	0.0067245	0.0079858	0.0077115	0.008852	0.024094	100
colSums	0.007372	0.0082060	0.0092560	0.0094550	0.009833	0.018231	100
apply+sum	0.055647	0.0572880	0.0590818	0.0583645	0.059017	0.112597	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000	100
colSums	1.241914	1.220314	1.159062	1.226091	1.110822	0.7566614	100
apply+sum	9.374495	8.519295	7.398353	7.568502	6.667081	4.6732382	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.006034	0.0067830	0.0079523	0.0073420	0.0088265	0.023782	100
rowSums	0.006896	0.0076980	0.0087938	0.0088325	0.0094700	0.015982	100
apply+sum	0.055437	0.0573025	0.0592130	0.0580355	0.0592000	0.114266	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.00000	1.000000	1.0000000	100
rowSums	1.142857	1.134896	1.105830	1.20301	1.072906	0.6720209	100
apply+sum	9.187438	8.447958	7.446059	7.90459	6.707075	4.8047263	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on double+10x10 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
2	rowCounts	6.034	6.7830	7.95226	7.3420	8.8265	23.782	100
1	colCounts	5.936	6.7245	7.98580	7.7115	8.8520	24.094	100

	expr	min	lq	mean	median	uq	max	neval
2	rowCounts	1.0000000	1.0000000	1.000000	1.000000	1.000000	1.000000	100
1	colCounts	0.9837587	0.9913755	1.004218	1.050327	1.002889	1.013119	100
Figure: Benchmarking of colCounts() and rowCounts() on double+10x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x100 matrix

> X <- data[["100x100"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.034769	0.0361295	0.0386138	0.0381915	0.0391920	0.064935	100
colSums	0.078908	0.0798225	0.0819926	0.0809350	0.0820820	0.095816	100
apply+sum	0.454868	0.4620650	0.5110225	0.4718755	0.4865935	2.280414	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.00000	1.000000	1.000000	1.000000	100
colSums	2.269493	2.209344	2.12340	2.119189	2.094356	1.475568	100
apply+sum	13.082574	12.789134	13.23418	12.355511	12.415633	35.118411	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.043411	0.044414	0.0467659	0.046210	0.0469590	0.068058	100
rowSums	0.094850	0.096010	0.1142993	0.096805	0.0973495	1.748354	100
apply+sum	0.455404	0.461539	0.4849037	0.464896	0.4720785	2.129474	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.00000	1.000000	1.000000	1.000000	1.000000	1.00000	100
rowSums	2.18493	2.161706	2.444074	2.094893	2.073074	25.68918	100
apply+sum	10.49052	10.391746	10.368746	10.060506	10.052993	31.28911	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on double+100x100 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
colCounts	34.769	36.1295	38.61383	38.1915	39.192	64.935	100
rowCounts	43.411	44.4140	46.76590	46.2100	46.959	68.058	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.0000	1.000000	1.000000	1.000000	1.000000	100
rowCounts	1.248555	1.2293	1.211118	1.209955	1.198178	1.048094	100
Figure: Benchmarking of colCounts() and rowCounts() on double+100x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x10 matrix

> X <- data[["1000x10"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.034373	0.0349355	0.0369176	0.036364	0.0372620	0.052782	100
colSums	0.078095	0.0787855	0.0807346	0.079753	0.0803150	0.121939	100
apply+sum	0.211868	0.2134145	0.2366454	0.214333	0.2164105	1.222975	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
colSums	2.271987	2.255170	2.186888	2.193186	2.155413	2.310238	100
apply+sum	6.163791	6.108815	6.410100	5.894099	5.807807	23.170304	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.044714	0.0457965	0.0491711	0.0479145	0.0496450	0.079371	100
rowSums	0.096121	0.0970720	0.0993996	0.0980085	0.0988460	0.114416	100
apply+sum	0.212571	0.2147905	0.2420802	0.2168910	0.2319675	1.194826	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowSums	2.149685	2.119638	2.021503	2.045487	1.991057	1.441534	100
apply+sum	4.754014	4.690107	4.923218	4.526626	4.672525	15.053685	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on double+1000x10 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
colCounts	34.373	34.9355	36.91759	36.3640	37.262	52.782	100
rowCounts	44.714	45.7965	49.17114	47.9145	49.645	79.371	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowCounts	1.300847	1.310887	1.331916	1.317636	1.332323	1.503751	100
Figure: Benchmarking of colCounts() and rowCounts() on double+1000x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

10x1000 matrix

> X <- data[["10x1000"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.037057	0.0384665	0.0438680	0.0415665	0.0487650	0.061573	100
colSums	0.080396	0.0821550	0.0854611	0.0840240	0.0882245	0.102324	100
apply+sum	2.720326	2.7820100	2.9387916	2.8537620	2.9454385	3.949417	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
colSums	2.169523	2.135755	1.948144	2.021435	1.809177	1.661832	100
apply+sum	73.409234	72.322930	66.991754	68.655335	60.400666	64.142027	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.042923	0.0444465	0.0493494	0.0479680	0.0539025	0.071507	100
rowSums	0.096006	0.0973235	0.1001528	0.0988215	0.1022085	0.115077	100
apply+sum	2.729340	2.7744635	2.9178261	2.8183935	2.9185295	3.876591	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowSums	2.236703	2.189678	2.029464	2.060155	1.896174	1.609311	100
apply+sum	63.586888	62.422542	59.125867	58.755702	54.144604	54.212748	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on double+10x1000 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
colCounts	37.057	38.4665	43.86796	41.5665	48.7650	61.573	100
rowCounts	42.923	44.4465	49.34940	47.9680	53.9025	71.507	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000	100
rowCounts	1.158297	1.15546	1.124953	1.154006	1.105352	1.161337	100
Figure: Benchmarking of colCounts() and rowCounts() on double+10x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x1000 matrix

> X <- data[["100x1000"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.295140	0.2969465	0.3087949	0.3070985	0.3171555	0.357848	100
colSums	0.714855	0.7171400	0.7795327	0.7264470	0.7355420	2.132311	100
apply+sum	4.280087	4.3487345	5.3278862	4.4174840	5.8508735	47.168773	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
colSums	2.422088	2.415048	2.524435	2.365518	2.319184	5.958706	100
apply+sum	14.501887	14.644842	17.253802	14.384583	18.447965	131.812314	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.371308	0.374305	0.3836680	0.382210	0.3900715	0.427003	100
rowSums	0.875715	0.878525	0.9894834	0.884685	0.8998235	2.553067	100
apply+sum	4.332604	4.422945	5.0157825	4.582655	5.9243640	6.633233	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowSums	2.35846	2.347083	2.579009	2.314657	2.306817	5.979038	100
apply+sum	11.66849	11.816420	13.073235	11.989888	15.187893	15.534394	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on double+100x1000 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
colCounts	295.140	296.9465	308.7949	307.0985	317.1555	357.848	100
rowCounts	371.308	374.3050	383.6680	382.2100	390.0715	427.003	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowCounts	1.258074	1.260513	1.242469	1.244584	1.229906	1.193252	100
Figure: Benchmarking of colCounts() and rowCounts() on double+100x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x100 matrix

> X <- data[["1000x100"]]
> value <- 42

> colStats <- microbenchmark(colCounts = colCounts(X, value = value, na.rm = FALSE), colSums = colSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 2L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

> X <- t(X)

> rowStats <- microbenchmark(rowCounts = rowCounts(X, value = value, na.rm = FALSE), rowSums = rowSums(X == 
+     value, na.rm = FALSE), `apply+sum` = apply(X, MARGIN = 1L, FUN = function(x) sum(x == value, 
+     na.rm = FALSE)), unit = "ms")

Table: Benchmarking of colCounts(), 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	neval
colCounts	0.287993	0.2892030	0.2994595	0.2959355	0.3058125	0.351222	100
colSums	0.708476	0.7099995	0.7534923	0.7157475	0.7262620	1.709930	100
apply+sum	1.848825	1.8611890	2.5727953	1.8875580	2.7690475	43.493654	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
colSums	2.460046	2.455021	2.516174	2.418593	2.374860	4.868516	100
apply+sum	6.419687	6.435580	8.591462	6.378275	9.054723	123.835221	100
Table: Benchmarking of rowCounts(), 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	neval
rowCounts	0.377871	0.380023	0.3891483	0.385470	0.3955075	0.418472	100
rowSums	0.881318	0.883883	0.9300696	0.887365	0.9021165	1.861215	100
apply+sum	1.879640	1.907911	2.6588699	1.943134	2.8218000	44.852003	100

expr	min	lq	mean	median	uq	max	neval
rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowSums	2.332325	2.325867	2.390013	2.302034	2.280909	4.447645	100
apply+sum	4.974290	5.020516	6.832536	5.040949	7.134631	107.180416	100
Figure: Benchmarking of colCounts(), colSums() and apply+sum() on double+1000x100 data as well as rowCounts(), rowSums() and apply+sum() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCounts() and rowCounts() 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	neval
colCounts	287.993	289.203	299.4595	295.9355	305.8125	351.222	100
rowCounts	377.871	380.023	389.1483	385.4700	395.5075	418.472	100

expr	min	lq	mean	median	uq	max	neval
colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	100
rowCounts	1.312084	1.314035	1.299502	1.302547	1.293301	1.191474	100
Figure: Benchmarking of colCounts() and rowCounts() on double+1000x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

Appendix

Session information

R version 3.0.2 (2013-09-25)
Platform: x86_64-pc-linux-gnu (64-bit)

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] markdown_0.7.4       microbenchmark_1.4-2 matrixStats_0.14.0  
[4] ggplot2_1.0.1        knitr_1.9            R.devices_2.13.0    
[7] R.utils_2.0.0        R.oo_1.19.0          R.methodsS3_1.7.0   

loaded via a namespace (and not attached):
 [1] base64enc_0.1-2  colorspace_1.2-6 digest_0.6.8     evaluate_0.5.5  
 [5] formatR_1.1      grid_3.0.2       gtable_0.1.2     labeling_0.3    
 [9] MASS_7.3-29      mime_0.2         munsell_0.4.2    plyr_1.8.1      
[13] proto_0.3-10     R.cache_0.10.0   Rcpp_0.11.5      reshape2_1.4.1  
[17] R.rsp_0.20.0     scales_0.2.4     stringr_0.6.2    tools_3.0.2     

Total processing time was 23.32 secs.

Reproducibility

To reproduce this report, do:

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

---

Copyright Henrik Bengtsson. Last updated on 2015-04-18 19:54:09 (+0000 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>