matrixStats.benchmarks

---

colCumsums() and rowCumsums() benchmarks on subsetted computation

This report benchmark the performance of colCumsums() and rowCumsums() on subsetted computation.

Data type “integer”

Data

> rmatrix <- function(nrow, ncol, mode = c("logical", "double", "integer", "index"), range = c(-100, 
+     +100), na_prob = 0) {
+     mode <- match.arg(mode)
+     n <- nrow * ncol
+     if (mode == "logical") {
+         x <- sample(c(FALSE, TRUE), size = n, replace = TRUE)
+     }     else if (mode == "index") {
+         x <- seq_len(n)
+         mode <- "integer"
+     }     else {
+         x <- runif(n, min = range[1], max = range[2])
+     }
+     storage.mode(x) <- mode
+     if (na_prob > 0) 
+         x[sample(n, size = na_prob * n)] <- NA
+     dim(x) <- c(nrow, ncol)
+     x
+ }
> rmatrices <- function(scale = 10, seed = 1, ...) {
+     set.seed(seed)
+     data <- list()
+     data[[1]] <- rmatrix(nrow = scale * 1, ncol = scale * 1, ...)
+     data[[2]] <- rmatrix(nrow = scale * 10, ncol = scale * 10, ...)
+     data[[3]] <- rmatrix(nrow = scale * 100, ncol = scale * 1, ...)
+     data[[4]] <- t(data[[3]])
+     data[[5]] <- rmatrix(nrow = scale * 10, ncol = scale * 100, ...)
+     data[[6]] <- t(data[[5]])
+     names(data) <- sapply(data, FUN = function(x) paste(dim(x), collapse = "x"))
+     data
+ }
> data <- rmatrices(mode = mode)

Results

10x10 integer matrix

> X <- data[["10x10"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5221525 278.9    7916910 422.9  7916910 422.9
Vcells 9882645  75.4   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5208790 278.2    7916910 422.9  7916910 422.9
Vcells 9840435  75.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.002187	0.0023425	0.0038885	0.0024185	0.0025225	0.146407
2	colCumsums(X, rows, cols)	0.002436	0.0026880	0.0028176	0.0027590	0.0028480	0.005069
3	colCummins(X[rows, cols])	0.002996	0.0034180	0.0035988	0.0035370	0.0036645	0.007258

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
2	colCumsums(X, rows, cols)	1.113855	1.147492	0.7246133	1.140790	1.129039	0.0346227
3	colCummins(X[rows, cols])	1.369913	1.459125	0.9255104	1.462477	1.452725	0.0495741

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.002192	0.0024270	0.0025563	0.0024875	0.0026365	0.004748
2	rowCumsums(X, cols, rows)	0.002492	0.0027175	0.0043897	0.0027865	0.0028960	0.158584
3	rowCumsums(X[cols, rows])	0.003003	0.0034560	0.0036023	0.0035515	0.0036320	0.007585

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCumsums(X, cols, rows)	1.136861	1.119695	1.717188	1.120201	1.098426	33.400168
3	rowCumsums(X[cols, rows])	1.369982	1.423980	1.409172	1.427739	1.377584	1.597515

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on integer+10x10 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

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

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	2.187	2.3425	3.88846	2.4185	2.5225	146.407
2	rowCumsums_X_S	2.192	2.4270	2.55633	2.4875	2.6365	4.748

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.0000000	1.00000	1.000000	1.0000000
2	rowCumsums_X_S	1.002286	1.036073	0.6574145	1.02853	1.045193	0.0324301

Figure: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on integer+10x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x100 integer matrix

> X <- data[["100x100"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5207495 278.2    7916910 422.9  7916910 422.9
Vcells 9509455  72.6   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5207489 278.2    7916910 422.9  7916910 422.9
Vcells 9514538  72.6   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.014966	0.015374	0.0158363	0.0154960	0.0156730	0.027424
2	colCumsums(X, rows, cols)	0.023868	0.024176	0.0246965	0.0243465	0.0244700	0.041711
3	colCummins(X[rows, cols])	0.026372	0.026909	0.0281198	0.0271350	0.0273285	0.059165

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCumsums(X, rows, cols)	1.594815	1.572525	1.559489	1.571147	1.561284	1.520967
3	colCummins(X[rows, cols])	1.762128	1.750293	1.775658	1.751097	1.743667	2.157417

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.020501	0.0216520	0.0222227	0.0218275	0.0222305	0.031656
2	rowCumsums(X, cols, rows)	0.023009	0.0242820	0.0253671	0.0244575	0.0248665	0.070186
3	rowCumsums(X[cols, rows])	0.031315	0.0329925	0.0346110	0.0332590	0.0340310	0.067481

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000
2	rowCumsums(X, cols, rows)	1.122335	1.121467	1.141493	1.12049	1.118576	2.217147
3	rowCumsums(X[cols, rows])	1.527486	1.523762	1.557458	1.52372	1.530825	2.131697

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on integer+100x100 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on integer+100x100 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	14.966	15.374	15.83628	15.4960	15.6730	27.424
2	rowCumsums_X_S	20.501	21.652	22.22272	21.8275	22.2305	31.656

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCumsums_X_S	1.369838	1.408352	1.403279	1.408589	1.418395	1.154317

Figure: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on integer+100x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x10 integer matrix

> X <- data[["1000x10"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5208237 278.2    7916910 422.9  7916910 422.9
Vcells 9513506  72.6   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5208231 278.2    7916910 422.9  7916910 422.9
Vcells 9518589  72.7   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.014539	0.014901	0.0153663	0.0152690	0.0154305	0.030117
2	colCumsums(X, rows, cols)	0.023026	0.024256	0.0251102	0.0246265	0.0253655	0.038718
3	colCummins(X[rows, cols])	0.024820	0.026322	0.0273808	0.0270345	0.0276205	0.052923

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCumsums(X, rows, cols)	1.583740	1.627810	1.634107	1.612843	1.643855	1.285586
3	colCummins(X[rows, cols])	1.707132	1.766459	1.781873	1.770548	1.789994	1.757247

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.020600	0.021898	0.0229122	0.0226535	0.023488	0.038163
2	rowCumsums(X, cols, rows)	0.024954	0.026149	0.0274845	0.0276460	0.028356	0.039416
3	rowCumsums(X[cols, rows])	0.033005	0.034743	0.0365598	0.0361615	0.037538	0.065182

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCumsums(X, cols, rows)	1.211359	1.194127	1.199558	1.220385	1.207255	1.032833
3	rowCumsums(X[cols, rows])	1.602184	1.586583	1.595650	1.596288	1.598178	1.707989

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on integer+1000x10 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() 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	colCumsums_X_S	14.539	14.901	15.36630	15.2690	15.4305	30.117
2	rowCumsums_X_S	20.600	21.898	22.91216	22.6535	23.4880	38.163

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.00000	1.000000
2	rowCumsums_X_S	1.416879	1.469566	1.491065	1.483627	1.52218	1.267158

Figure: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on integer+1000x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

10x1000 integer matrix

> X <- data[["10x1000"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5208442 278.2    7916910 422.9  7916910 422.9
Vcells 9514334  72.6   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5208436 278.2    7916910 422.9  7916910 422.9
Vcells 9519417  72.7   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.015101	0.0159295	0.0167881	0.016561	0.0167885	0.035414
2	colCumsums(X, rows, cols)	0.025431	0.0266525	0.0277261	0.027781	0.0280280	0.042531
3	colCummins(X[rows, cols])	0.028260	0.0299825	0.0307777	0.030895	0.0311495	0.043917

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCumsums(X, rows, cols)	1.684061	1.673154	1.651534	1.677495	1.669476	1.200966
3	colCummins(X[rows, cols])	1.871399	1.882200	1.833308	1.865527	1.855407	1.240103

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.018594	0.0201560	0.0205578	0.0203280	0.0206045	0.033723
2	rowCumsums(X, cols, rows)	0.021955	0.0235780	0.0246662	0.0238220	0.0241395	0.063707
3	rowCumsums(X[cols, rows])	0.029738	0.0320155	0.0324293	0.0322925	0.0326635	0.047299

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCumsums(X, cols, rows)	1.180757	1.169776	1.199849	1.171881	1.171564	1.889126
3	rowCumsums(X[cols, rows])	1.599333	1.588386	1.577474	1.588572	1.585260	1.402574

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on integer+10x1000 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on integer+10x1000 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	15.101	15.9295	16.78808	16.561	16.7885	35.414
2	rowCumsums_X_S	18.594	20.1560	20.55776	20.328	20.6045	33.723

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowCumsums_X_S	1.231309	1.265325	1.224545	1.227462	1.227298	0.9522505

Figure: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on integer+10x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x1000 integer matrix

> X <- data[["100x1000"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5208653 278.2    7916910 422.9  7916910 422.9
Vcells 9537005  72.8   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5208647 278.2    7916910 422.9  7916910 422.9
Vcells 9587088  73.2   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.085659	0.0935475	0.1058033	0.1009105	0.1145250	0.152515
2	colCumsums(X, rows, cols)	0.143120	0.1570405	0.1777168	0.1722325	0.1949645	0.284269
3	colCummins(X[rows, cols])	0.157417	0.1737415	0.2003399	0.1900500	0.2192475	0.320763

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCumsums(X, rows, cols)	1.670811	1.678725	1.679691	1.706785	1.702375	1.863876
3	colCummins(X[rows, cols])	1.837717	1.857254	1.893514	1.883352	1.914407	2.103157

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.116090	0.1361245	0.1502013	0.1459120	0.1627330	0.222236
2	rowCumsums(X, cols, rows)	0.131060	0.1424760	0.1632429	0.1601415	0.1759785	0.274895
3	rowCumsums(X[cols, rows])	0.179466	0.2028175	0.2281330	0.2247085	0.2418230	0.314098

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCumsums(X, cols, rows)	1.128952	1.046660	1.086827	1.097521	1.081394	1.236951
3	rowCumsums(X[cols, rows])	1.545921	1.489941	1.518849	1.540028	1.486011	1.413353

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on integer+100x1000 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on integer+100x1000 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	85.659	93.5475	105.8033	100.9105	114.525	152.515
2	rowCumsums_X_S	116.090	136.1245	150.2013	145.9120	162.733	222.236

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCumsums_X_S	1.355258	1.455138	1.419628	1.445955	1.420939	1.457142

Figure: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on integer+100x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x100 integer matrix

> X <- data[["1000x100"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5208863 278.2    7916910 422.9  7916910 422.9
Vcells 9537781  72.8   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5208857 278.2    7916910 422.9  7916910 422.9
Vcells 9587864  73.2   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.085421	0.0960880	0.1051075	0.1023735	0.1125505	0.158771
2	colCumsums(X, rows, cols)	0.142818	0.1601600	0.1767880	0.1709135	0.1898255	0.232814
3	colCummins(X[rows, cols])	0.154752	0.1737175	0.1935330	0.1871860	0.2144070	0.295733

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCumsums(X, rows, cols)	1.671931	1.666805	1.681973	1.669509	1.686581	1.466351
3	colCummins(X[rows, cols])	1.811639	1.807900	1.841286	1.828461	1.904985	1.862639

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.120129	0.1305330	0.1520348	0.1486140	0.1692605	0.234467
2	rowCumsums(X, cols, rows)	0.134202	0.1560545	0.1767003	0.1693115	0.1873500	0.342235
3	rowCumsums(X[cols, rows])	0.184329	0.2055035	0.2388848	0.2318885	0.2660125	0.392404

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
2	rowCumsums(X, cols, rows)	1.117149	1.195518	1.162236	1.139270	1.106874	1.45963
3	rowCumsums(X[cols, rows])	1.534425	1.574341	1.571250	1.560341	1.571616	1.67360

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on integer+1000x100 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() 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	colCumsums_X_S	85.421	96.088	105.1075	102.3735	112.5505	158.771
2	rowCumsums_X_S	120.129	130.533	152.0349	148.6140	169.2605	234.467

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.00000	1.000000	1.000000	1.000000
2	rowCumsums_X_S	1.406317	1.358473	1.44647	1.451684	1.503863	1.476762

Figure: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on integer+1000x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

Data type “double”

Data

> rmatrix <- function(nrow, ncol, mode = c("logical", "double", "integer", "index"), range = c(-100, 
+     +100), na_prob = 0) {
+     mode <- match.arg(mode)
+     n <- nrow * ncol
+     if (mode == "logical") {
+         x <- sample(c(FALSE, TRUE), size = n, replace = TRUE)
+     }     else if (mode == "index") {
+         x <- seq_len(n)
+         mode <- "integer"
+     }     else {
+         x <- runif(n, min = range[1], max = range[2])
+     }
+     storage.mode(x) <- mode
+     if (na_prob > 0) 
+         x[sample(n, size = na_prob * n)] <- NA
+     dim(x) <- c(nrow, ncol)
+     x
+ }
> rmatrices <- function(scale = 10, seed = 1, ...) {
+     set.seed(seed)
+     data <- list()
+     data[[1]] <- rmatrix(nrow = scale * 1, ncol = scale * 1, ...)
+     data[[2]] <- rmatrix(nrow = scale * 10, ncol = scale * 10, ...)
+     data[[3]] <- rmatrix(nrow = scale * 100, ncol = scale * 1, ...)
+     data[[4]] <- t(data[[3]])
+     data[[5]] <- rmatrix(nrow = scale * 10, ncol = scale * 100, ...)
+     data[[6]] <- t(data[[5]])
+     names(data) <- sapply(data, FUN = function(x) paste(dim(x), collapse = "x"))
+     data
+ }
> data <- rmatrices(mode = mode)

Results

10x10 double matrix

> X <- data[["10x10"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5209081 278.2    7916910 422.9  7916910 422.9
Vcells 9628888  73.5   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5209066 278.2    7916910 422.9  7916910 422.9
Vcells 9629056  73.5   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.002109	0.0022375	0.0025103	0.0023225	0.0024230	0.016412
2	colCumsums(X, rows, cols)	0.002292	0.0025845	0.0027037	0.0026600	0.0027540	0.005152
3	colCummins(X[rows, cols])	0.003126	0.0033755	0.0035408	0.0034990	0.0036175	0.007606

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colCumsums(X, rows, cols)	1.086771	1.155084	1.077076	1.145317	1.136607	0.3139166
3	colCummins(X[rows, cols])	1.482219	1.508603	1.410539	1.506566	1.492984	0.4634414

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.002113	0.0023265	0.0027674	0.0024285	0.0025725	0.010611
2	rowCumsums(X, cols, rows)	0.002334	0.0026040	0.0031750	0.0027805	0.0029730	0.019302
3	rowCumsums(X[cols, rows])	0.003246	0.0035160	0.0039242	0.0036210	0.0037795	0.010398

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowCumsums(X, cols, rows)	1.104591	1.119278	1.147272	1.144945	1.155685	1.8190557
3	rowCumsums(X[cols, rows])	1.536204	1.511283	1.418008	1.491044	1.469193	0.9799265

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on double+10x10 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

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

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	2.109	2.2375	2.51026	2.3225	2.4230	16.412
2	rowCumsums_X_S	2.113	2.3265	2.76739	2.4285	2.5725	10.611

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.00000	1.0000	1.0000000
2	rowCumsums_X_S	1.001897	1.039777	1.102432	1.04564	1.0617	0.6465391

Figure: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on double+10x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x100 double matrix

> X <- data[["100x100"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5209280 278.3    7916910 422.9  7916910 422.9
Vcells 9634838  73.6   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5209274 278.3    7916910 422.9  7916910 422.9
Vcells 9644921  73.6   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.008385	0.008832	0.0091381	0.0090265	0.0092485	0.012737
2	colCumsums(X, rows, cols)	0.013312	0.013789	0.0140690	0.0140650	0.0142015	0.017707
3	colCummins(X[rows, cols])	0.024722	0.025222	0.0262576	0.0253960	0.0256840	0.057829

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCumsums(X, rows, cols)	1.587597	1.561254	1.539601	1.558190	1.535546	1.390202
3	colCummins(X[rows, cols])	2.948360	2.855752	2.873430	2.813494	2.777099	4.540237

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.006892	0.0076505	0.0082450	0.008164	0.0085210	0.013263
2	rowCumsums(X, cols, rows)	0.013694	0.0140090	0.0149928	0.014199	0.0144140	0.045128
3	rowCumsums(X[cols, rows])	0.023596	0.0240160	0.0249024	0.024710	0.0252885	0.035033

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCumsums(X, cols, rows)	1.986941	1.831122	1.818416	1.739221	1.691585	3.402548
3	rowCumsums(X[cols, rows])	3.423680	3.139141	3.020310	3.026703	2.967786	2.641408

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on double+100x100 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on double+100x100 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	rowCumsums_X_S	6.892	7.6505	8.24499	8.1640	8.5210	13.263
1	colCumsums_X_S	8.385	8.8320	9.13808	9.0265	9.2485	12.737

	expr	min	lq	mean	median	uq	max
2	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colCumsums_X_S	1.216628	1.154434	1.108319	1.105647	1.085377	0.9603408

Figure: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on double+100x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x10 double matrix

> X <- data[["1000x10"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5209479 278.3    7916910 422.9  7916910 422.9
Vcells 9636239  73.6   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5209473 278.3    7916910 422.9  7916910 422.9
Vcells 9646322  73.6   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.010257	0.0109105	0.0114655	0.0110805	0.0112095	0.042591
2	colCumsums(X, rows, cols)	0.014083	0.0150475	0.0155989	0.0152585	0.0155235	0.030346
3	colCummins(X[rows, cols])	0.026451	0.0278425	0.0288871	0.0280870	0.0283535	0.057928

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colCumsums(X, rows, cols)	1.373014	1.379176	1.360508	1.377059	1.384852	0.7124979
3	colCummins(X[rows, cols])	2.578824	2.551900	2.519485	2.534813	2.529417	1.3600996

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.009068	0.0097245	0.0109601	0.0102480	0.0110915	0.018957
2	rowCumsums(X, cols, rows)	0.017052	0.0177405	0.0197559	0.0180285	0.0186185	0.049472
3	rowCumsums(X[cols, rows])	0.029108	0.0298330	0.0330052	0.0307890	0.0314115	0.100503

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCumsums(X, cols, rows)	1.880459	1.824310	1.802529	1.759221	1.678628	2.609696
3	rowCumsums(X[cols, rows])	3.209969	3.067818	3.011400	3.004391	2.832034	5.301630

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on double+1000x10 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

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

	expr	min	lq	mean	median	uq	max
2	rowCumsums_X_S	9.068	9.7245	10.96010	10.2480	11.0915	18.957
1	colCumsums_X_S	10.257	10.9105	11.46547	11.0805	11.2095	42.591

	expr	min	lq	mean	median	uq	max
2	rowCumsums_X_S	1.00000	1.00000	1.00000	1.000000	1.000000	1.000000
1	colCumsums_X_S	1.13112	1.12196	1.04611	1.081235	1.010639	2.246716

Figure: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on double+1000x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

10x1000 double matrix

> X <- data[["10x1000"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5209684 278.3    7916910 422.9  7916910 422.9
Vcells 9636375  73.6   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5209678 278.3    7916910 422.9  7916910 422.9
Vcells 9646458  73.6   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.008074	0.0088665	0.0102586	0.0091860	0.0097660	0.048346
2	colCumsums(X, rows, cols)	0.015422	0.0162015	0.0178825	0.0166055	0.0175145	0.036718
3	colCummins(X[rows, cols])	0.027596	0.0290160	0.0315073	0.0292980	0.0300845	0.069444

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colCumsums(X, rows, cols)	1.910082	1.827271	1.743170	1.807697	1.793416	0.7594837
3	colCummins(X[rows, cols])	3.417885	3.272543	3.071309	3.189419	3.080535	1.4363960

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.006717	0.0075325	0.0081717	0.0080655	0.0083825	0.017332
2	rowCumsums(X, cols, rows)	0.014666	0.0150885	0.0156827	0.0152660	0.0154865	0.038996
3	rowCumsums(X[cols, rows])	0.023749	0.0244450	0.0255443	0.0250915	0.0256115	0.056241

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCumsums(X, cols, rows)	2.183415	2.003120	1.919141	1.892753	1.847480	2.249942
3	rowCumsums(X[cols, rows])	3.535656	3.245271	3.125940	3.110966	3.055353	3.244923

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on double+10x1000 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() 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	rowCumsums_X_S	6.717	7.5325	8.17173	8.0655	8.3825	17.332
1	colCumsums_X_S	8.074	8.8665	10.25858	9.1860	9.7660	48.346

	expr	min	lq	mean	median	uq	max
2	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCumsums_X_S	1.202025	1.177099	1.255374	1.138925	1.165046	2.789407

Figure: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on double+10x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x1000 double matrix

> X <- data[["100x1000"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5209895 278.3    7916910 422.9  7916910 422.9
Vcells 9681827  73.9   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5209889 278.3    7916910 422.9  7916910 422.9
Vcells 9781910  74.7   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.052834	0.0604730	0.0887941	0.0680615	0.0867560	0.271492
2	colCumsums(X, rows, cols)	0.094961	0.0992340	0.1383374	0.1064625	0.1384710	0.371713
3	colCummins(X[rows, cols])	0.178834	0.1890475	0.2551509	0.2082200	0.2843435	0.672803

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCumsums(X, rows, cols)	1.797346	1.640964	1.557956	1.564210	1.596097	1.369149
3	colCummins(X[rows, cols])	3.384828	3.126147	2.873511	3.059292	3.277508	2.478169

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.040027	0.0418715	0.0494413	0.0440965	0.0505560	0.107859
2	rowCumsums(X, cols, rows)	0.083877	0.0872310	0.0955318	0.0892380	0.0931000	0.164449
3	rowCumsums(X[cols, rows])	0.157927	0.1653100	0.1802106	0.1691215	0.1768535	0.303600

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCumsums(X, cols, rows)	2.095511	2.083302	1.932228	2.023698	1.841522	1.524666
3	rowCumsums(X[cols, rows])	3.945512	3.948031	3.644943	3.835259	3.498170	2.814786

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on double+100x1000 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on double+100x1000 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	rowCumsums_X_S	40.027	41.8715	49.44127	44.0965	50.556	107.859
1	colCumsums_X_S	52.834	60.4730	88.79412	68.0615	86.756	271.492

	expr	min	lq	mean	median	uq	max
2	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCumsums_X_S	1.319959	1.444252	1.795951	1.543467	1.716038	2.517101

Figure: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on double+100x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x100 double matrix

> X <- data[["1000x100"]]
> rows <- sample.int(nrow(X), size = nrow(X) * 0.7)
> cols <- sample.int(ncol(X), size = ncol(X) * 0.7)
> X_S <- X[rows, cols]
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5210105 278.3    7916910 422.9  7916910 422.9
Vcells 9681968  73.9   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCumsums_X_S = colCumsums(X_S), `colCumsums(X, rows, cols)` = colCumsums(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCumsums(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5210099 278.3    7916910 422.9  7916910 422.9
Vcells 9782051  74.7   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCumsums_X_S = rowCumsums(X_S), `rowCumsums(X, cols, rows)` = rowCumsums(X, 
+     rows = cols, cols = rows), `rowCumsums(X[cols, rows])` = rowCumsums(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() 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	colCumsums_X_S	0.067361	0.0726580	0.1369943	0.109547	0.2004660	0.217652
2	colCumsums(X, rows, cols)	0.090478	0.1009495	0.1576254	0.126204	0.2241475	0.265737
3	colCummins(X[rows, cols])	0.189788	0.2197365	0.3164680	0.261764	0.4417790	0.487193

	expr	min	lq	mean	median	uq	max
1	colCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCumsums(X, rows, cols)	1.343181	1.389379	1.150598	1.152054	1.118132	1.220926
3	colCummins(X[rows, cols])	2.817476	3.024258	2.310082	2.389513	2.203760	2.238403

Table: Benchmarking of rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() 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	rowCumsums_X_S	0.041043	0.0432690	0.0494292	0.0463665	0.0512750	0.095122
2	rowCumsums(X, cols, rows)	0.091367	0.0955605	0.1073948	0.1015550	0.1141825	0.237039
3	rowCumsums(X[cols, rows])	0.163954	0.1725130	0.1903939	0.1836175	0.2059775	0.275066

	expr	min	lq	mean	median	uq	max
1	rowCumsums_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCumsums(X, cols, rows)	2.226129	2.208521	2.172701	2.190267	2.226865	2.491947
3	rowCumsums(X[cols, rows])	3.994689	3.986988	3.851852	3.960133	4.017114	2.891718

Figure: Benchmarking of colCumsums_X_S(), colCumsums(X, rows, cols)() and colCummins(X[rows, cols])() on double+1000x100 data as well as rowCumsums_X_S(), rowCumsums(X, cols, rows)() and rowCumsums(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colCumsums_X_S() and rowCumsums_X_S() on double+1000x100 data (original and transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	rowCumsums_X_S	41.043	43.269	49.42919	46.3665	51.275	95.122
1	colCumsums_X_S	67.361	72.658	136.99427	109.5470	200.466	217.652

	expr	min	lq	mean	median	uq	max
2	rowCumsums_X_S	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCumsums_X_S	1.64123	1.679216	2.771526	2.362633	3.909625	2.288135

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

Appendix

Session information

R version 4.1.1 Patched (2021-08-10 r80727)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.5 LTS

Matrix products: default
BLAS:   /home/hb/software/R-devel/R-4-1-branch/lib/R/lib/libRblas.so
LAPACK: /home/hb/software/R-devel/R-4-1-branch/lib/R/lib/libRlapack.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] microbenchmark_1.4-7   matrixStats_0.60.0     ggplot2_3.3.5         
[4] knitr_1.33             R.devices_2.17.0       R.utils_2.10.1        
[7] R.oo_1.24.0            R.methodsS3_1.8.1-9001 history_0.0.1-9000    

loaded via a namespace (and not attached):
 [1] Biobase_2.52.0          httr_1.4.2              splines_4.1.1          
 [4] bit64_4.0.5             network_1.17.1          assertthat_0.2.1       
 [7] highr_0.9               stats4_4.1.1            blob_1.2.2             
[10] GenomeInfoDbData_1.2.6  robustbase_0.93-8       pillar_1.6.2           
[13] RSQLite_2.2.8           lattice_0.20-44         glue_1.4.2             
[16] digest_0.6.27           XVector_0.32.0          colorspace_2.0-2       
[19] Matrix_1.3-4            XML_3.99-0.7            pkgconfig_2.0.3        
[22] zlibbioc_1.38.0         genefilter_1.74.0       purrr_0.3.4            
[25] ergm_4.1.2              xtable_1.8-4            scales_1.1.1           
[28] tibble_3.1.4            annotate_1.70.0         KEGGREST_1.32.0        
[31] farver_2.1.0            generics_0.1.0          IRanges_2.26.0         
[34] ellipsis_0.3.2          cachem_1.0.6            withr_2.4.2            
[37] BiocGenerics_0.38.0     mime_0.11               survival_3.2-13        
[40] magrittr_2.0.1          crayon_1.4.1            statnet.common_4.5.0   
[43] memoise_2.0.0           laeken_0.5.1            fansi_0.5.0            
[46] R.cache_0.15.0          MASS_7.3-54             R.rsp_0.44.0           
[49] progressr_0.8.0         tools_4.1.1             lifecycle_1.0.0        
[52] S4Vectors_0.30.0        trust_0.1-8             munsell_0.5.0          
[55] tabby_0.0.1-9001        AnnotationDbi_1.54.1    Biostrings_2.60.2      
[58] compiler_4.1.1          GenomeInfoDb_1.28.1     rlang_0.4.11           
[61] grid_4.1.1              RCurl_1.98-1.4          cwhmisc_6.6            
[64] rstudioapi_0.13         rappdirs_0.3.3          startup_0.15.0         
[67] labeling_0.4.2          bitops_1.0-7            base64enc_0.1-3        
[70] boot_1.3-28             gtable_0.3.0            DBI_1.1.1              
[73] markdown_1.1            R6_2.5.1                lpSolveAPI_5.5.2.0-17.7
[76] rle_0.9.2               dplyr_1.0.7             fastmap_1.1.0          
[79] bit_4.0.4               utf8_1.2.2              parallel_4.1.1         
[82] Rcpp_1.0.7              vctrs_0.3.8             png_0.1-7              
[85] DEoptimR_1.0-9          tidyselect_1.1.1        xfun_0.25              
[88] coda_0.19-4            

Total processing time was 23.3 secs.

Reproducibility

To reproduce this report, do:

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

---

Copyright Dongcan Jiang. Last updated on 2021-08-25 22:14:45 (+0200 UTC). Powered by RSP.