matrixStats.benchmarks

---

colCummins() and rowCummins() benchmarks on subsetted computation

This report benchmark the performance of colCummins() and rowCummins() 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 5205545 278.1    7916910 422.9  7916910 422.9
Vcells 9772989  74.6   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5193013 277.4    7916910 422.9  7916910 422.9
Vcells 9731461  74.3   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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	colCummins_X_S	0.002330	0.0024935	0.0041899	0.002534	0.0026285	0.163776
2	colCummins(X, rows, cols)	0.002512	0.0028250	0.0029663	0.002905	0.0030145	0.005294
3	colCummins(X[rows, cols])	0.003221	0.0035685	0.0037370	0.003672	0.0037845	0.007849

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
2	colCummins(X, rows, cols)	1.078112	1.132946	0.7079807	1.146409	1.146852	0.0323246
3	colCummins(X[rows, cols])	1.382403	1.431121	0.8919152	1.449092	1.439795	0.0479252

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.002180	0.0025405	0.0028294	0.0026840	0.002929	0.005550
2	rowCummins(X, cols, rows)	0.002590	0.0029520	0.0049077	0.0031060	0.003385	0.172772
3	rowCummins(X[cols, rows])	0.003161	0.0037645	0.0042124	0.0040355	0.004395	0.008983

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCummins(X, cols, rows)	1.188073	1.161976	1.734530	1.157228	1.155685	31.130090
3	rowCummins(X[cols, rows])	1.450000	1.481795	1.488793	1.503540	1.500512	1.618559

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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	colCummins_X_S	2.33	2.4935	4.18986	2.534	2.6285	163.776
2	rowCummins_X_S	2.18	2.5405	2.82940	2.684	2.9290	5.550

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.0000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowCummins_X_S	0.9356223	1.018849	0.675297	1.059195	1.114324	0.0338877

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 5191738 277.3    7916910 422.9  7916910 422.9
Vcells 9400288  71.8   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5191732 277.3    7916910 422.9  7916910 422.9
Vcells 9405371  71.8   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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	colCummins_X_S	0.014722	0.0150745	0.0154053	0.0152420	0.0156375	0.020775
2	colCummins(X, rows, cols)	0.019863	0.0200705	0.0204905	0.0202300	0.0209000	0.024746
3	colCummins(X[rows, cols])	0.025919	0.0264095	0.0274265	0.0266065	0.0275770	0.051017

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCummins(X, rows, cols)	1.349205	1.331421	1.330091	1.327254	1.336531	1.191143
3	colCummins(X[rows, cols])	1.760562	1.751932	1.780325	1.745604	1.763517	2.455692

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.011269	0.0116405	0.0120914	0.0117745	0.011904	0.021832
2	rowCummins(X, cols, rows)	0.020076	0.0203055	0.0210412	0.0204785	0.020640	0.046656
3	rowCummins(X[cols, rows])	0.022696	0.0230005	0.0232872	0.0231585	0.023364	0.028107

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCummins(X, cols, rows)	1.781524	1.744384	1.740175	1.739225	1.733871	2.137046
3	rowCummins(X[cols, rows])	2.014021	1.975903	1.925929	1.966835	1.962702	1.287422

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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
2	rowCummins_X_S	11.269	11.6405	12.09144	11.7745	11.9040	21.832
1	colCummins_X_S	14.722	15.0745	15.40532	15.2420	15.6375	20.775

	expr	min	lq	mean	median	uq	max
2	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colCummins_X_S	1.306416	1.295004	1.274068	1.294492	1.313634	0.9515848

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 5192480 277.4    7916910 422.9  7916910 422.9
Vcells 9404338  71.8   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5192474 277.4    7916910 422.9  7916910 422.9
Vcells 9409421  71.8   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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	colCummins_X_S	0.015670	0.0158360	0.0161365	0.0160075	0.0161700	0.022060
2	colCummins(X, rows, cols)	0.020882	0.0211305	0.0213286	0.0212610	0.0213995	0.024265
3	colCummins(X[rows, cols])	0.027668	0.0281160	0.0294205	0.0282215	0.0284410	0.062390

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCummins(X, rows, cols)	1.332610	1.334333	1.321760	1.328190	1.323408	1.099955
3	colCummins(X[rows, cols])	1.765667	1.775448	1.823221	1.763017	1.758874	2.828196

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.014166	0.0144155	0.0148361	0.0145505	0.0147625	0.029874
2	rowCummins(X, cols, rows)	0.023063	0.0233835	0.0240219	0.0236370	0.0238950	0.041511
3	rowCummins(X[cols, rows])	0.027575	0.0281980	0.0289943	0.0283870	0.0287650	0.057518

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCummins(X, cols, rows)	1.628053	1.622108	1.619154	1.624480	1.618628	1.389536
3	rowCummins(X[cols, rows])	1.946562	1.956089	1.954311	1.950929	1.948518	1.925353

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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
2	rowCummins_X_S	14.166	14.4155	14.83608	14.5505	14.7625	29.874
1	colCummins_X_S	15.670	15.8360	16.13655	16.0075	16.1700	22.060

	expr	min	lq	mean	median	uq	max
2	rowCummins_X_S	1.00000	1.00000	1.000000	1.000000	1.000000	1.0000000
1	colCummins_X_S	1.10617	1.09854	1.087656	1.100134	1.095343	0.7384348

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 5192685 277.4    7916910 422.9  7916910 422.9
Vcells 9405170  71.8   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5192679 277.4    7916910 422.9  7916910 422.9
Vcells 9410253  71.8   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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	colCummins_X_S	0.012417	0.0129775	0.0133815	0.0130920	0.0132620	0.027088
2	colCummins(X, rows, cols)	0.021947	0.0228485	0.0232299	0.0229705	0.0231175	0.037744
3	colCummins(X[rows, cols])	0.026242	0.0268645	0.0271684	0.0270525	0.0272390	0.031215

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCummins(X, rows, cols)	1.767496	1.760624	1.735977	1.754545	1.743138	1.393385
3	colCummins(X[rows, cols])	2.113393	2.070083	2.030304	2.066338	2.053913	1.152355

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.010842	0.0112060	0.0114377	0.0113875	0.0115715	0.014782
2	rowCummins(X, cols, rows)	0.019605	0.0200830	0.0205675	0.0202240	0.0204085	0.044026
3	rowCummins(X[cols, rows])	0.022824	0.0232155	0.0235122	0.0233875	0.0236185	0.028475

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCummins(X, cols, rows)	1.808246	1.792165	1.798223	1.775982	1.763687	2.978352
3	rowCummins(X[cols, rows])	2.105147	2.071703	2.055684	2.053787	2.041092	1.926329

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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
2	rowCummins_X_S	10.842	11.2060	11.43768	11.3875	11.5715	14.782
1	colCummins_X_S	12.417	12.9775	13.38146	13.0920	13.2620	27.088

	expr	min	lq	mean	median	uq	max
2	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCummins_X_S	1.145268	1.158085	1.169945	1.149682	1.146092	1.832499

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 5192896 277.4    7916910 422.9  7916910 422.9
Vcells 9427841  72.0   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5192890 277.4    7916910 422.9  7916910 422.9
Vcells 9477924  72.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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	colCummins_X_S	0.091849	0.097813	0.1084712	0.1021645	0.1173995	0.165053
2	colCummins(X, rows, cols)	0.124766	0.133437	0.1502803	0.1379690	0.1668310	0.283260
3	colCummins(X[rows, cols])	0.169853	0.181177	0.2021601	0.1929750	0.2217820	0.275184

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCummins(X, rows, cols)	1.358382	1.364205	1.385439	1.350459	1.421054	1.716176
3	colCummins(X[rows, cols])	1.849263	1.852279	1.863721	1.888866	1.889122	1.667246

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.065341	0.0697605	0.0782191	0.0747945	0.0866420	0.124541
2	rowCummins(X, cols, rows)	0.127408	0.1325905	0.1477554	0.1377975	0.1600035	0.301935
3	rowCummins(X[cols, rows])	0.141966	0.1514080	0.1663348	0.1574730	0.1824585	0.265356

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.00000	1.000000
2	rowCummins(X, cols, rows)	1.949894	1.900653	1.888994	1.842348	1.84672	2.424382
3	rowCummins(X[cols, rows])	2.172694	2.170397	2.126526	2.105409	2.10589	2.130672

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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
2	rowCummins_X_S	65.341	69.7605	78.21906	74.7945	86.6420	124.541
1	colCummins_X_S	91.849	97.8130	108.47125	102.1645	117.3995	165.053

	expr	min	lq	mean	median	uq	max
2	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
1	colCummins_X_S	1.405687	1.402126	1.386762	1.365936	1.354995	1.32529

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 5193106 277.4    7916910 422.9  7916910 422.9
Vcells 9428618  72.0   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5193100 277.4    7916910 422.9  7916910 422.9
Vcells 9478701  72.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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	colCummins_X_S	0.097288	0.1043695	0.1141792	0.1094445	0.1224050	0.167498
2	colCummins(X, rows, cols)	0.124448	0.1325890	0.1513855	0.1430235	0.1623875	0.308314
3	colCummins(X[rows, cols])	0.171952	0.1848175	0.2125704	0.2042670	0.2365570	0.341080

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCummins(X, rows, cols)	1.279171	1.270381	1.325859	1.306813	1.326641	1.840703
3	colCummins(X[rows, cols])	1.767453	1.770800	1.861725	1.866398	1.932576	2.036323

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.064526	0.0690160	0.0752052	0.0727410	0.0794415	0.114851
2	rowCummins(X, cols, rows)	0.125718	0.1342185	0.1461374	0.1403835	0.1541915	0.264038
3	rowCummins(X[cols, rows])	0.142518	0.1527550	0.1666361	0.1579085	0.1786145	0.240792

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCummins(X, cols, rows)	1.948331	1.944745	1.943183	1.929909	1.940944	2.298961
3	rowCummins(X[cols, rows])	2.208691	2.213327	2.215754	2.170832	2.248378	2.096560

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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
2	rowCummins_X_S	64.526	69.0160	75.20515	72.7410	79.4415	114.851
1	colCummins_X_S	97.288	104.3695	114.17923	109.4445	122.4050	167.498

	expr	min	lq	mean	median	uq	max
2	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCummins_X_S	1.507733	1.512251	1.518237	1.504578	1.540819	1.458394

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 5193324 277.4    7916910 422.9  7916910 422.9
Vcells 9519723  72.7   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5193309 277.4    7916910 422.9  7916910 422.9
Vcells 9519891  72.7   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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	colCummins_X_S	0.002216	0.0023310	0.0027661	0.002414	0.0024975	0.032504
2	colCummins(X, rows, cols)	0.002441	0.0027025	0.0028328	0.002778	0.0028675	0.005693
3	colCummins(X[rows, cols])	0.003184	0.0033950	0.0035779	0.003496	0.0036075	0.009843

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colCummins(X, rows, cols)	1.101534	1.159374	1.024117	1.150787	1.148148	0.1751477
3	colCummins(X[rows, cols])	1.436823	1.456457	1.293463	1.448219	1.444444	0.3028243

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.002192	0.0024240	0.0026295	0.0025030	0.0026140	0.009779
2	rowCummins(X, cols, rows)	0.002463	0.0026520	0.0030682	0.0027145	0.0028265	0.033493
3	rowCummins(X[cols, rows])	0.003037	0.0034525	0.0035934	0.0035190	0.0036480	0.008236

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowCummins(X, cols, rows)	1.123631	1.094059	1.166833	1.084499	1.081293	3.4249923
3	rowCummins(X[cols, rows])	1.385493	1.424299	1.366561	1.405913	1.395562	0.8422129

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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	colCummins_X_S	2.216	2.331	2.76614	2.414	2.4975	32.504
2	rowCummins_X_S	2.192	2.424	2.62952	2.503	2.6140	9.779

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.0000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
2	rowCummins_X_S	0.9891697	1.039897	0.9506099	1.036868	1.046647	0.3008553

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 5193524 277.4    7916910 422.9  7916910 422.9
Vcells 9525683  72.7   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5193518 277.4    7916910 422.9  7916910 422.9
Vcells 9535766  72.8   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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	colCummins_X_S	0.018548	0.019370	0.0214916	0.0196115	0.0203645	0.081237
2	colCummins(X, rows, cols)	0.027487	0.028266	0.0297218	0.0285560	0.0297905	0.044972
3	colCummins(X[rows, cols])	0.035384	0.035747	0.0406520	0.0361620	0.0378185	0.145265

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colCummins(X, rows, cols)	1.481939	1.459267	1.382948	1.456084	1.462864	0.5535901
3	colCummins(X[rows, cols])	1.907699	1.845483	1.891528	1.843918	1.857080	1.7881630

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.008610	0.009274	0.0098582	0.0096510	0.0099870	0.024965
2	rowCummins(X, cols, rows)	0.016725	0.017617	0.0187327	0.0179720	0.0185155	0.049009
3	rowCummins(X[cols, rows])	0.023891	0.024936	0.0259356	0.0256355	0.0264860	0.038386

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCummins(X, cols, rows)	1.942509	1.899612	1.900225	1.862190	1.853960	1.963108
3	rowCummins(X[cols, rows])	2.774797	2.688807	2.630879	2.656253	2.652048	1.537593

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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	rowCummins_X_S	8.610	9.274	9.85815	9.6510	9.9870	24.965
1	colCummins_X_S	18.548	19.370	21.49163	19.6115	20.3645	81.237

	expr	min	lq	mean	median	uq	max
2	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCummins_X_S	2.154239	2.088635	2.180087	2.032069	2.039101	3.254036

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 5193723 277.4    7916910 422.9  7916910 422.9
Vcells 9527087  72.7   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5193717 277.4    7916910 422.9  7916910 422.9
Vcells 9537170  72.8   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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	colCummins_X_S	0.020001	0.0211825	0.0216423	0.0213465	0.021506	0.051758
2	colCummins(X, rows, cols)	0.024540	0.0256575	0.0261151	0.0259225	0.026106	0.040537
3	colCummins(X[rows, cols])	0.036050	0.0369165	0.0388757	0.0383145	0.038539	0.077128

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colCummins(X, rows, cols)	1.226939	1.211259	1.206670	1.214368	1.213894	0.7832026
3	colCummins(X[rows, cols])	1.802410	1.742783	1.796283	1.794884	1.792012	1.4901658

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.010577	0.0114020	0.0118177	0.0118060	0.0120895	0.014964
2	rowCummins(X, cols, rows)	0.019738	0.0208580	0.0214364	0.0217220	0.0220050	0.024602
3	rowCummins(X[cols, rows])	0.028957	0.0303755	0.0324387	0.0316365	0.0320795	0.067983

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCummins(X, cols, rows)	1.866125	1.829328	1.813913	1.839912	1.820175	1.644079
3	rowCummins(X[cols, rows])	2.737733	2.664050	2.744916	2.679697	2.653501	4.543103

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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	rowCummins_X_S	10.577	11.4020	11.81774	11.8060	12.0895	14.964
1	colCummins_X_S	20.001	21.1825	21.64233	21.3465	21.5060	51.758

	expr	min	lq	mean	median	uq	max
2	rowCummins_X_S	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCummins_X_S	1.89099	1.857788	1.831343	1.808106	1.778899	3.458835

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 5193928 277.4    7916910 422.9  7916910 422.9
Vcells 9527223  72.7   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5193922 277.4    7916910 422.9  7916910 422.9
Vcells 9537306  72.8   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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	colCummins_X_S	0.011280	0.0118475	0.0123158	0.0120555	0.0123515	0.031398
3	colCummins(X[rows, cols])	0.030099	0.0313330	0.0320822	0.0319560	0.0322320	0.058855
2	colCummins(X, rows, cols)	0.038250	0.0401205	0.0420835	0.0413480	0.0430510	0.068609

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colCummins(X[rows, cols])	2.668351	2.644693	2.604954	2.650740	2.609562	1.874483
2	colCummins(X, rows, cols)	3.390957	3.386411	3.417025	3.429804	3.485488	2.185139

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.009841	0.0103465	0.0155926	0.0179385	0.0190945	0.022135
2	rowCummins(X, cols, rows)	0.019463	0.0268685	0.0312358	0.0334295	0.0348230	0.057469
3	rowCummins(X[cols, rows])	0.027592	0.0281150	0.0398897	0.0444620	0.0459760	0.096757

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCummins(X, cols, rows)	1.977746	2.596868	2.003247	1.863562	1.823719	2.596296
3	rowCummins(X[cols, rows])	2.803780	2.717344	2.558243	2.478580	2.407814	4.371222

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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
1	colCummins_X_S	11.280	11.8475	12.31584	12.0555	12.3515	31.398
2	rowCummins_X_S	9.841	10.3465	15.59261	17.9385	19.0945	22.135

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.0000000	1.0000000	1.000000	1.000000	1.000000	1.0000000
2	rowCummins_X_S	0.8724291	0.8733066	1.266061	1.487993	1.545926	0.7049812

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 5194139 277.4    7916910 422.9  7916910 422.9
Vcells 9572677  73.1   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5194133 277.4    7916910 422.9  7916910 422.9
Vcells 9672760  73.8   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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	colCummins_X_S	0.123995	0.1489515	0.1902407	0.1716795	0.2377835	0.274697
2	colCummins(X, rows, cols)	0.206843	0.2215395	0.2682690	0.2676780	0.3009670	0.383653
3	colCummins(X[rows, cols])	0.249138	0.2781345	0.3626980	0.3239760	0.4529690	0.547241

	expr	min	lq	mean	median	uq	max
1	colCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colCummins(X, rows, cols)	1.668156	1.487326	1.410155	1.559173	1.265719	1.396641
3	colCummins(X[rows, cols])	2.009258	1.867282	1.906522	1.887098	1.904964	1.992162

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.053265	0.0568025	0.0655034	0.0628090	0.0690690	0.116550
2	rowCummins(X, cols, rows)	0.117202	0.1187275	0.1311534	0.1257885	0.1389795	0.255642
3	rowCummins(X[cols, rows])	0.173150	0.1790470	0.2006694	0.1961750	0.2154230	0.351481

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCummins(X, cols, rows)	2.200357	2.090181	2.002237	2.002715	2.012183	2.193411
3	rowCummins(X[cols, rows])	3.250727	3.152097	3.063494	3.123358	3.118953	3.015710

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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	rowCummins_X_S	53.265	56.8025	65.50343	62.8090	69.0690	116.550
1	colCummins_X_S	123.995	148.9515	190.24072	171.6795	237.7835	274.697

	expr	min	lq	mean	median	uq	max
2	rowCummins_X_S	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
1	colCummins_X_S	2.327889	2.62227	2.904286	2.733358	3.442695	2.356903

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 5194349 277.5    7916910 422.9  7916910 422.9
Vcells 9572818  73.1   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colCummins_X_S = colCummins(X_S), `colCummins(X, rows, cols)` = colCummins(X, 
+     rows = rows, cols = cols), `colCummins(X[rows, cols])` = colCummins(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5194343 277.5    7916910 422.9  7916910 422.9
Vcells 9672901  73.8   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowCummins_X_S = rowCummins(X_S), `rowCummins(X, cols, rows)` = rowCummins(X, 
+     rows = cols, cols = rows), `rowCummins(X[cols, rows])` = rowCummins(X[cols, rows]), unit = "ms")

Table: Benchmarking of colCummins_X_S(), colCummins(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
2	colCummins(X, rows, cols)	0.170143	0.1879635	0.2392096	0.2400995	0.2819365	0.323949
1	colCummins_X_S	0.145330	0.1700135	0.2202261	0.2402445	0.2609510	0.313648
3	colCummins(X[rows, cols])	0.261099	0.3078045	0.3916748	0.3450270	0.4830445	0.541359

	expr	min	lq	mean	median	uq	max
2	colCummins(X, rows, cols)	1.0000000	1.0000000	1.0000000	1.000000	1.0000000	1.0000000
1	colCummins_X_S	0.8541639	0.9045027	0.9206408	1.000604	0.9255666	0.9682018
3	colCummins(X[rows, cols])	1.5345856	1.6375759	1.6373704	1.437017	1.7133096	1.6711242

Table: Benchmarking of rowCummins_X_S(), rowCummins(X, cols, rows)() and rowCummins(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	rowCummins_X_S	0.053698	0.0563250	0.0643461	0.0603185	0.0690515	0.106838
2	rowCummins(X, cols, rows)	0.118892	0.1234485	0.1388333	0.1360840	0.1488855	0.308372
3	rowCummins(X[cols, rows])	0.174977	0.1816975	0.2041870	0.1964685	0.2244635	0.354954

	expr	min	lq	mean	median	uq	max
1	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowCummins(X, cols, rows)	2.214086	2.191718	2.157603	2.256091	2.156152	2.886351
3	rowCummins(X[cols, rows])	3.258538	3.225877	3.173263	3.257185	3.250668	3.322357

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

Table: Benchmarking of colCummins_X_S() and rowCummins_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	rowCummins_X_S	53.698	56.3250	64.34607	60.3185	69.0515	106.838
1	colCummins_X_S	145.330	170.0135	220.22613	240.2445	260.9510	313.648

	expr	min	lq	mean	median	uq	max
2	rowCummins_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colCummins_X_S	2.706432	3.018438	3.422527	3.982932	3.779078	2.935735

Figure: Benchmarking of colCummins_X_S() and rowCummins_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 23.5 secs.

Reproducibility

To reproduce this report, do:

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

---

Copyright Dongcan Jiang. Last updated on 2021-08-25 22:13:04 (+0200 UTC). Powered by RSP.