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), 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 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
1	colCounts	0.002328	0.0026845	0.0034006	0.0031785	0.0036755	0.020089
2	colSums	0.003611	0.0042750	0.0050358	0.0048795	0.0054905	0.018323
3	apply+sum	0.023981	0.0250105	0.0268431	0.0256210	0.0263100	0.115911

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.00000	1.0000000
2	colSums	1.551117	1.592475	1.480874	1.535158	1.49381	0.9120912
3	apply+sum	10.301117	9.316633	7.893680	8.060721	7.15821	5.7698741

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
1	rowCounts	0.002142	0.0025260	0.0031219	0.0030625	0.0033300	0.019303
2	rowSums	0.004027	0.0047025	0.0052483	0.0051865	0.0055520	0.013782
3	apply+sum	0.023156	0.0238310	0.0255243	0.0243485	0.0249995	0.113094

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowSums	1.880019	1.861639	1.681142	1.693551	1.667267	0.7139823
3	apply+sum	10.810458	9.434283	8.175985	7.950531	7.507357	5.8588820

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
2	rowCounts	2.142	2.5260	3.12186	3.0625	3.3300	19.303
1	colCounts	2.328	2.6845	3.40058	3.1785	3.6755	20.089

	expr	min	lq	mean	median	uq	max
2	rowCounts	1.000000	1.000000	1.00000	1.000000	1.000000	1.000000
1	colCounts	1.086835	1.062747	1.08928	1.037878	1.103754	1.040719

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
2	colSums	0.028677	0.0319080	0.0339170	0.033191	0.0347105	0.083212
1	colCounts	0.036823	0.0406280	0.0427344	0.043041	0.0445850	0.055337
3	apply+sum	0.249355	0.2666405	0.2764543	0.276950	0.2815840	0.369515

	expr	min	lq	mean	median	uq	max
2	colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colCounts	1.284060	1.273286	1.259971	1.296767	1.284482	0.6650123
3	apply+sum	8.695296	8.356541	8.150916	8.344129	8.112358	4.4406456

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
1	rowCounts	0.039630	0.0419400	0.0438029	0.0435405	0.0449350	0.060316
2	rowSums	0.052529	0.0549360	0.0574219	0.0569215	0.0584745	0.095847
3	apply+sum	0.201360	0.2126885	0.2198666	0.2181190	0.2247705	0.316050

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums	1.325486	1.309871	1.310915	1.307323	1.301313	1.589081
3	apply+sum	5.080999	5.071257	5.019455	5.009566	5.002125	5.239903

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
1	colCounts	36.823	40.628	42.73438	43.0410	44.585	55.337
2	rowCounts	39.630	41.940	43.80289	43.5405	44.935	60.316

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.00000	1.000000
2	rowCounts	1.076229	1.032293	1.025003	1.011605	1.00785	1.089976

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
2	colSums	0.028710	0.0302335	0.0321662	0.0309535	0.0324555	0.079907
1	colCounts	0.035595	0.0385470	0.0397576	0.0393385	0.0409075	0.056743
3	apply+sum	0.137407	0.1414435	0.1456927	0.1446505	0.1482805	0.192041

	expr	min	lq	mean	median	uq	max
2	colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCounts	1.239812	1.274976	1.236005	1.270890	1.260418	0.710113
3	apply+sum	4.786033	4.678370	4.529368	4.673155	4.568733	2.403306

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
1	rowCounts	0.042749	0.0442715	0.0460280	0.0453575	0.046274	0.086405
3	apply+sum	0.090446	0.0961030	0.0993340	0.0975900	0.100060	0.165663
2	rowSums	0.150703	0.1518345	0.1538293	0.1525445	0.154458	0.169657

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	apply+sum	2.115745	2.170765	2.158122	2.151574	2.162337	1.917285
2	rowSums	3.525299	3.429622	3.342081	3.363159	3.337900	1.963509

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
1	colCounts	35.595	38.5470	39.75761	39.3385	40.9075	56.743
2	rowCounts	42.749	44.2715	46.02800	45.3575	46.2740	86.405

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCounts	1.200983	1.148507	1.157715	1.153005	1.131186	1.522743

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
2	colSums	0.030470	0.0347585	0.0369110	0.035816	0.0379420	0.094213
1	colCounts	0.048898	0.0507590	0.0533154	0.052875	0.0544615	0.071200
3	apply+sum	1.152888	1.3644255	1.4148292	1.416287	1.4737975	1.647667

	expr	min	lq	mean	median	uq	max
2	colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colCounts	1.604792	1.460333	1.444431	1.476295	1.435388	0.7557343
3	apply+sum	37.836823	39.254441	38.330819	39.543416	38.843432	17.4887436

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
1	rowCounts	0.040192	0.046462	0.0480386	0.0479915	0.0494305	0.057295
2	rowSums	0.045889	0.050253	0.0518240	0.0512735	0.0527490	0.083858
3	apply+sum	1.144308	1.334182	1.3896115	1.3898330	1.4537275	1.626345

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums	1.141745	1.081594	1.078799	1.068387	1.067135	1.463618
3	apply+sum	28.471039	28.715553	28.926964	28.959982	29.409524	28.385461

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
2	rowCounts	40.192	46.462	48.03862	47.9915	49.4305	57.295
1	colCounts	48.898	50.759	53.31541	52.8750	54.4615	71.200

	expr	min	lq	mean	median	uq	max
2	rowCounts	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCounts	1.21661	1.092484	1.109845	1.101758	1.101779	1.242691

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
2	colSums	0.251793	0.278771	0.2837448	0.282931	0.2886400	0.373711
1	colCounts	0.367332	0.411473	0.4158703	0.414132	0.4252835	0.453008
3	apply+sum	2.358500	2.597695	2.7831135	2.643990	2.6813615	17.475270

	expr	min	lq	mean	median	uq	max
2	colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCounts	1.458865	1.476025	1.465649	1.463721	1.473405	1.212188
3	apply+sum	9.366821	9.318383	9.808508	9.344999	9.289639	46.761455

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
1	rowCounts	0.381769	0.423853	0.4341455	0.434970	0.4390095	0.594224
2	rowSums	0.393760	0.443222	0.4479370	0.446323	0.4558705	0.584000
3	apply+sum	1.911745	2.104206	2.2907912	2.135264	2.1851250	17.030164

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowSums	1.031409	1.045697	1.031767	1.026101	1.038407	0.9827944
3	apply+sum	5.007596	4.964471	5.276551	4.908991	4.977398	28.6595021

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
1	colCounts	367.332	411.473	415.8703	414.132	425.2835	453.008
2	rowCounts	381.769	423.853	434.1455	434.970	439.0095	594.224

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
2	rowCounts	1.039302	1.030087	1.043945	1.050317	1.032275	1.31173

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
2	colSums	0.246997	0.2684990	0.3471783	0.2780810	0.2879720	7.077920
1	colCounts	0.350538	0.3812510	0.3936804	0.3929965	0.4051735	0.465459
3	apply+sum	0.795796	0.8459175	0.8681194	0.8652085	0.8879385	1.033500

	expr	min	lq	mean	median	uq	max
2	colSums	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colCounts	1.419199	1.419935	1.133943	1.413245	1.406989	0.0657621
3	apply+sum	3.221885	3.150542	2.500500	3.111354	3.083420	0.1460175

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
1	rowCounts	0.383917	0.4188915	0.4288461	0.4291130	0.4403380	0.497454
2	rowSums	0.490445	0.5427390	0.5491929	0.5447770	0.5591100	0.625087
3	apply+sum	0.867725	0.8985720	0.9925708	0.9146845	0.9323115	7.680937

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums	1.277477	1.295655	1.280629	1.269542	1.269729	1.256573
3	apply+sum	2.260189	2.145119	2.314515	2.131570	2.117263	15.440497

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
1	colCounts	350.538	381.2510	393.6804	392.9965	405.1735	465.459
2	rowCounts	383.917	418.8915	428.8461	429.1130	440.3380	497.454

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.0000	1.000000	1.000000
2	rowCounts	1.095222	1.098729	1.089325	1.0919	1.086789	1.068739

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), 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 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
1	colCounts	0.002294	0.0027695	0.0033275	0.0031695	0.0035730	0.018437
2	colSums	0.003521	0.0042510	0.0050100	0.0047865	0.0053525	0.018981
3	apply+sum	0.023711	0.0244435	0.0260772	0.0249880	0.0254990	0.110113

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colSums	1.534874	1.534934	1.505642	1.510175	1.498041	1.029506
3	apply+sum	10.336094	8.825961	7.836960	7.883893	7.136580	5.972392

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
1	rowCounts	0.002224	0.0025580	0.0031205	0.003065	0.003297	0.019767
2	rowSums	0.004278	0.0046755	0.0052453	0.005149	0.005538	0.013490
3	apply+sum	0.023192	0.0241550	0.0258686	0.024641	0.025128	0.115457

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowSums	1.923561	1.827795	1.680907	1.679935	1.679709	0.6824505
3	apply+sum	10.428058	9.442924	8.289889	8.039478	7.621474	5.8408964

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
2	rowCounts	2.224	2.5580	3.12050	3.0650	3.297	19.767
1	colCounts	2.294	2.7695	3.32747	3.1695	3.573	18.437

	expr	min	lq	mean	median	uq	max
2	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colCounts	1.031475	1.082682	1.066326	1.034095	1.083713	0.9327161

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
1	colCounts	0.009680	0.0109585	0.0118371	0.011584	0.0122145	0.026120
2	colSums	0.028641	0.0305650	0.0325819	0.032243	0.0333430	0.068816
3	apply+sum	0.202369	0.2136885	0.2224504	0.220758	0.2277980	0.314136

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colSums	2.958781	2.789159	2.752523	2.783408	2.729788	2.634609
3	apply+sum	20.905888	19.499795	18.792643	19.057148	18.649801	12.026646

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
1	rowCounts	0.009782	0.0107015	0.0116533	0.0115075	0.012229	0.025980
2	rowSums	0.052254	0.0543355	0.0570388	0.0568945	0.058840	0.082481
3	apply+sum	0.198398	0.2091705	0.2170734	0.2168195	0.220956	0.291754

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums	5.341852	5.077372	4.894644	4.944123	4.811514	3.174788
3	apply+sum	20.281946	19.545905	18.627605	18.841582	18.068199	11.229946

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
2	rowCounts	9.782	10.7015	11.65332	11.5075	12.2290	25.98
1	colCounts	9.680	10.9585	11.83710	11.5840	12.2145	26.12

	expr	min	lq	mean	median	uq	max
2	rowCounts	1.0000000	1.000000	1.000000	1.000000	1.0000000	1.000000
1	colCounts	0.9895727	1.024015	1.015771	1.006648	0.9988143	1.005389

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
1	colCounts	0.008485	0.0091530	0.0100014	0.0096835	0.0102135	0.026648
2	colSums	0.028789	0.0297235	0.0312879	0.0304580	0.0314395	0.078758
3	apply+sum	0.094330	0.0970910	0.1004045	0.0986365	0.1004485	0.154553

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000
2	colSums	3.392929	3.247405	3.128366	3.14535	3.078230	2.955494
3	apply+sum	11.117266	10.607560	10.039082	10.18604	9.834875	5.799797

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
1	rowCounts	0.009922	0.010539	0.0120016	0.0113540	0.0118370	0.048509
3	apply+sum	0.092502	0.095798	0.1002758	0.0967795	0.0998375	0.183271
2	rowSums	0.150489	0.151916	0.1576946	0.1550520	0.1577745	0.249131

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	apply+sum	9.322919	9.089857	8.355213	8.523824	8.434358	3.778082
2	rowSums	15.167204	14.414650	13.139489	13.656156	13.328926	5.135769

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
1	colCounts	8.485	9.153	10.00136	9.6835	10.2135	26.648
2	rowCounts	9.922	10.539	12.00158	11.3540	11.8370	48.509

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000
2	rowCounts	1.169358	1.151426	1.199995	1.17251	1.158956	1.820362

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
1	colCounts	0.013271	0.0144885	0.0166902	0.0155940	0.0182175	0.036133
2	colSums	0.030669	0.0344525	0.0374436	0.0356975	0.0384700	0.083503
3	apply+sum	1.143084	1.3664885	1.4343462	1.4163240	1.4798550	1.910101

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.00000	1.000000	1.000000	1.000000	1.00000
2	colSums	2.310979	2.37792	2.243454	2.289182	2.111706	2.31099
3	apply+sum	86.133976	94.31539	85.939578	90.824933	81.232606	52.86306

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
1	rowCounts	0.009825	0.011479	0.0127799	0.0126325	0.0135450	0.025046
2	rowSums	0.047333	0.050072	0.0513539	0.0510790	0.0521535	0.064745
3	apply+sum	1.121708	1.333550	1.3801187	1.3870145	1.4488205	1.596805

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums	4.817608	4.362052	4.018324	4.043459	3.850388	2.585043
3	apply+sum	114.168753	116.173012	107.991099	109.797308	106.963492	63.754891

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
2	rowCounts	9.825	11.4790	12.77993	12.6325	13.5450	25.046
1	colCounts	13.271	14.4885	16.69017	15.5940	18.2175	36.133

	expr	min	lq	mean	median	uq	max
2	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCounts	1.350738	1.262174	1.305967	1.234435	1.344961	1.442665

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
1	colCounts	0.077557	0.0839765	0.0875862	0.0866995	0.0897325	0.131382
2	colSums	0.244317	0.2777990	0.2820857	0.2842420	0.2903470	0.311446
3	apply+sum	1.911778	2.0975750	2.2806285	2.1376915	2.1812515	16.733862

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colSums	3.150161	3.308056	3.220665	3.278473	3.235695	2.370538
3	apply+sum	24.649974	24.978119	26.038688	24.656330	24.308378	127.367996

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
1	rowCounts	0.066644	0.074960	0.0777305	0.0771330	0.0793215	0.097746
2	rowSums	0.395390	0.448138	0.4542746	0.4581175	0.4610180	0.498521
3	apply+sum	1.930299	2.111248	2.2927815	2.1475750	2.1861240	16.694900

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums	5.932867	5.978362	5.844227	5.939319	5.812018	5.100168
3	apply+sum	28.964333	28.164995	29.496549	27.842493	27.560296	170.798805

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
2	rowCounts	66.644	74.9600	77.73050	77.1330	79.3215	97.746
1	colCounts	77.557	83.9765	87.58615	86.6995	89.7325	131.382

	expr	min	lq	mean	median	uq	max
2	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCounts	1.163751	1.120284	1.126793	1.124026	1.131251	1.344116

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
1	colCounts	0.062391	0.0709410	0.0731314	0.0727470	0.0747695	0.097505
2	colSums	0.250656	0.2761615	0.3521502	0.2848915	0.2939970	7.008896
3	apply+sum	0.820201	0.8754505	0.8908282	0.8886385	0.9077550	1.048810

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.00000	1.000000	1.000000	1.00000
2	colSums	4.017502	3.892834	4.81531	3.916196	3.932044	71.88243
3	apply+sum	13.146143	12.340544	12.18120	12.215466	12.140712	10.75647

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
1	rowCounts	0.077440	0.0868970	0.0913577	0.0894245	0.0918630	0.259978
2	rowSums	0.489164	0.5421850	0.5558072	0.5589465	0.5721430	0.675897
3	apply+sum	0.828744	0.8977505	0.9865928	0.9072540	0.9307245	7.791024

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.00000	1.000000	1.00000	1.000000
2	rowSums	6.316684	6.239398	6.08386	6.250485	6.22822	2.599824
3	apply+sum	10.701756	10.331202	10.79923	10.145475	10.13166	29.968013

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
1	colCounts	62.391	70.941	73.13138	72.7470	74.7695	97.505
2	rowCounts	77.440	86.897	91.35766	89.4245	91.8630	259.978

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCounts	1.241205	1.224919	1.249227	1.229253	1.228616	2.666304

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), 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 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
1	colCounts	0.002400	0.002824	0.0033545	0.0032355	0.0036150	0.015873
2	colSums	0.003528	0.004263	0.0049980	0.0048640	0.0054525	0.017406
3	apply+sum	0.024579	0.025424	0.0273306	0.0261875	0.0268355	0.098683

	expr	min	lq	mean	median	uq	max
1	colCounts	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colSums	1.47000	1.509561	1.489945	1.503322	1.508299	1.096579
3	apply+sum	10.24125	9.002833	8.147501	8.093803	7.423375	6.217035

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
1	rowCounts	0.002119	0.0024580	0.0030468	0.0030460	0.0033125	0.016031
2	rowSums	0.004028	0.0045980	0.0051230	0.0049595	0.0054595	0.012398
3	apply+sum	0.022943	0.0238265	0.0250884	0.0241670	0.0247565	0.082336

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowSums	1.900897	1.870626	1.681399	1.628201	1.648151	0.7733766
3	apply+sum	10.827277	9.693450	8.234203	7.934012	7.473660	5.1360489

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
2	rowCounts	2.119	2.458	3.04685	3.0460	3.3125	16.031
1	colCounts	2.400	2.824	3.35448	3.2355	3.6150	15.873

	expr	min	lq	mean	median	uq	max
2	rowCounts	1.00000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colCounts	1.13261	1.148902	1.100967	1.062213	1.091321	0.9901441

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
1	colCounts	0.013077	0.0143520	0.0150193	0.0148825	0.0156385	0.027298
2	colSums	0.024183	0.0256325	0.0271621	0.0268515	0.0278580	0.057314
3	apply+sum	0.241223	0.2513305	0.2632340	0.2645980	0.2703540	0.332537

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colSums	1.849277	1.785988	1.808484	1.804233	1.781373	2.099568
3	apply+sum	18.446356	17.511880	17.526415	17.779137	17.287719	12.181735

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
1	rowCounts	0.013647	0.0148050	0.0159620	0.0156780	0.0165370	0.032710
2	rowSums	0.048004	0.0501075	0.0526763	0.0529405	0.0542335	0.091946
3	apply+sum	0.197127	0.2090720	0.2178631	0.2183785	0.2224440	0.312596

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums	3.51755	3.384498	3.300104	3.376738	3.279525	2.810945
3	apply+sum	14.44471	14.121716	13.648853	13.928977	13.451291	9.556588

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
1	colCounts	13.077	14.352	15.01927	14.8825	15.6385	27.298
2	rowCounts	13.647	14.805	15.96201	15.6780	16.5370	32.710

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCounts	1.043588	1.031564	1.062769	1.053452	1.057454	1.198256

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
1	colCounts	0.012190	0.0131750	0.0140944	0.0137900	0.0143500	0.028055
2	colSums	0.023972	0.0253625	0.0268999	0.0263130	0.0275100	0.051356
3	apply+sum	0.133603	0.1370685	0.1438736	0.1440655	0.1486545	0.184863

	expr	min	lq	mean	median	uq	max
1	colCounts	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colSums	1.96653	1.925047	1.908552	1.908122	1.917073	1.830547
3	apply+sum	10.96005	10.403681	10.207848	10.447099	10.359199	6.589307

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
1	rowCounts	0.013342	0.014339	0.0152904	0.0147195	0.0152260	0.046777
3	apply+sum	0.091434	0.094447	0.0971692	0.0967245	0.0978970	0.166123
2	rowSums	0.144738	0.147814	0.1532645	0.1543355	0.1552815	0.182930

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	apply+sum	6.853096	6.586722	6.354906	6.571181	6.429594	3.551382
2	rowSums	10.848299	10.308529	10.023564	10.485105	10.198443	3.910683

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
1	colCounts	12.190	13.175	14.09441	13.7900	14.350	28.055
2	rowCounts	13.342	14.339	15.29042	14.7195	15.226	46.777

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCounts	1.094504	1.088349	1.084857	1.067404	1.061045	1.667332

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
1	colCounts	0.015487	0.0167375	0.0179957	0.0177835	0.0185205	0.029386
2	colSums	0.025288	0.0288670	0.0304565	0.0298915	0.0315220	0.052157
3	apply+sum	1.128363	1.3575160	1.3910470	1.3982795	1.4497070	1.595142

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
2	colSums	1.632853	1.72469	1.692434	1.680856	1.702006	1.774893
3	apply+sum	72.858720	81.10626	77.298898	78.627914	78.275802	54.282379

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
1	rowCounts	0.014309	0.0162555	0.0176940	0.0173895	0.018379	0.030320
2	rowSums	0.038846	0.0451725	0.0468241	0.0465485	0.047673	0.063300
3	apply+sum	1.138963	1.3534980	1.3938169	1.4042220	1.454298	1.630223

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums	2.714795	2.778906	2.646327	2.676817	2.593884	2.087731
3	apply+sum	79.597666	83.264003	78.773378	80.751143	79.128271	53.767249

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
2	rowCounts	14.309	16.2555	17.69401	17.3895	18.3790	30.320
1	colCounts	15.487	16.7375	17.99569	17.7835	18.5205	29.386

	expr	min	lq	mean	median	uq	max
2	rowCounts	1.000000	1.000000	1.00000	1.000000	1.000000	1.0000000
1	colCounts	1.082326	1.029651	1.01705	1.022657	1.007699	0.9691953

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
1	colCounts	0.106030	0.1215205	0.1270202	0.1230815	0.130991	0.194324
2	colSums	0.205654	0.2300105	0.3694346	0.2353750	0.244164	13.281108
3	apply+sum	2.331110	2.5964585	2.7965672	2.6262150	2.672417	16.346528

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
2	colSums	1.939583	1.892771	2.908472	1.912351	1.863975	68.34518
3	apply+sum	21.985381	21.366424	22.016716	21.337203	20.401531	84.11996

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
1	rowCounts	0.110568	0.1244755	0.1309001	0.129068	0.1337675	0.175034
2	rowSums	0.346945	0.3912795	0.4061822	0.402256	0.4198065	0.477486
3	apply+sum	1.912307	2.1099115	2.4360798	2.150076	2.1997965	15.678354

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.00000	1.000000
2	rowSums	3.137843	3.143426	3.102993	3.116621	3.13833	2.727961
3	apply+sum	17.295302	16.950416	18.610215	16.658479	16.44493	89.573191

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
1	colCounts	106.030	121.5205	127.0202	123.0815	130.9910	194.324
2	rowCounts	110.568	124.4755	130.9001	129.0680	133.7675	175.034

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowCounts	1.042799	1.024317	1.030546	1.048639	1.021196	0.9007328

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
1	colCounts	0.099038	0.1137430	0.1152469	0.1154190	0.1176060	0.129126
2	colSums	0.200269	0.2308030	0.2927586	0.2363550	0.2418085	5.926186
3	apply+sum	0.798039	0.8596005	0.9269429	0.8695245	0.8835600	6.472874

	expr	min	lq	mean	median	uq	max
1	colCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
2	colSums	2.022143	2.029162	2.540272	2.047800	2.056090	45.89460
3	apply+sum	8.057907	7.557392	8.043102	7.533634	7.512882	50.12836

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
1	rowCounts	0.115612	0.1307065	0.1348954	0.1355175	0.137661	0.186606
2	rowSums	0.442426	0.5029895	0.5149004	0.5174200	0.521725	0.911519
3	apply+sum	0.850686	0.8945170	1.0342045	0.9017175	0.923892	6.506540

	expr	min	lq	mean	median	uq	max
1	rowCounts	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowSums	3.826817	3.848236	3.817036	3.818105	3.789926	4.884725
3	apply+sum	7.358112	6.843707	7.666716	6.653882	6.711356	34.867796

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
1	colCounts	99.038	113.7430	115.2469	115.4190	117.606	129.126
2	rowCounts	115.612	130.7065	134.8954	135.5175	137.661	186.606

	expr	min	lq	mean	median	uq	max
1	colCounts	1.00000	1.000000	1.00000	1.000000	1.000000	1.000000
2	rowCounts	1.16735	1.149139	1.17049	1.174135	1.170527	1.445147

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 3.6.1 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 37.43 secs.

Reproducibility

To reproduce this report, do:

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

---

Copyright Henrik Bengtsson. Last updated on 2019-09-10 20:37:25 (-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>