matrixStats.benchmarks

---

colDiffs() and rowDiffs() benchmarks on subsetted computation

This report benchmark the performance of colDiffs() and rowDiffs() 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 5229403 279.3    7916910 422.9  7916910 422.9
Vcells 9937010  75.9   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5216771 278.7    7916910 422.9  7916910 422.9
Vcells 9895059  75.5   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs_X_S	0.002420	0.0026685	0.0048089	0.0027600	0.0028790	0.204439
2	colDiffs(X, rows, cols)	0.002779	0.0029955	0.0031543	0.0031085	0.0032035	0.005495
3	colDiffs(X[rows, cols])	0.003340	0.0037500	0.0040192	0.0039085	0.0040985	0.007430

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
2	colDiffs(X, rows, cols)	1.148347	1.122541	0.6559254	1.126268	1.112713	0.0268784
3	colDiffs(X[rows, cols])	1.380165	1.405284	0.8357889	1.416123	1.423585	0.0363434

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.002488	0.0027240	0.0028994	0.0027925	0.0029155	0.004651
2	rowDiffs(X, cols, rows)	0.002885	0.0030920	0.0052886	0.0031660	0.0032910	0.210292
3	rowDiffs(X[cols, rows])	0.003421	0.0037915	0.0039987	0.0039240	0.0040790	0.008344

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	1.159566	1.135095	1.824040	1.133751	1.128794	45.214363
3	rowDiffs(X[cols, rows])	1.375000	1.391887	1.379136	1.405193	1.399074	1.794023

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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	colDiffs_X_S	2.420	2.6685	4.80887	2.7600	2.8790	204.439
2	rowDiffs_X_S	2.488	2.7240	2.89941	2.7925	2.9155	4.651

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
2	rowDiffs_X_S	1.028099	1.020798	0.6029296	1.011775	1.012678	0.0227501

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 5215597 278.6    7916910 422.9  7916910 422.9
Vcells 9564472  73.0   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5215591 278.6    7916910 422.9  7916910 422.9
Vcells 9569555  73.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs_X_S	0.008978	0.0095325	0.0099565	0.0096740	0.0099190	0.015421
2	colDiffs(X, rows, cols)	0.014254	0.0148565	0.0153480	0.0151305	0.0153765	0.024972
3	colDiffs(X[rows, cols])	0.020055	0.0208915	0.0221599	0.0213495	0.0215975	0.074278

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colDiffs(X, rows, cols)	1.587659	1.558510	1.541497	1.564038	1.550207	1.619350
3	colDiffs(X[rows, cols])	2.233794	2.191608	2.225669	2.206895	2.177387	4.816679

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.008817	0.0092940	0.0098796	0.0094825	0.0097075	0.033440
2	rowDiffs(X, cols, rows)	0.013550	0.0141555	0.0149148	0.0144580	0.0147035	0.051330
3	rowDiffs(X[cols, rows])	0.019358	0.0203160	0.0213836	0.0208975	0.0213250	0.038412

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	1.536804	1.523079	1.509646	1.524703	1.514654	1.534988
3	rowDiffs(X[cols, rows])	2.195531	2.185926	2.164415	2.203797	2.196755	1.148684

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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	rowDiffs_X_S	8.817	9.2940	9.87964	9.4825	9.7075	33.440
1	colDiffs_X_S	8.978	9.5325	9.95653	9.6740	9.9190	15.421

	expr	min	lq	mean	median	uq	max
2	rowDiffs_X_S	1.00000	1.000000	1.000000	1.000000	1.000000	1.0000000
1	colDiffs_X_S	1.01826	1.025662	1.007783	1.020195	1.021787	0.4611543

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 5216340 278.6    7916910 422.9  7916910 422.9
Vcells 9568547  73.1   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5216334 278.6    7916910 422.9  7916910 422.9
Vcells 9573630  73.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs_X_S	0.008689	0.0093540	0.0095785	0.0094925	0.009697	0.012686
2	colDiffs(X, rows, cols)	0.014674	0.0155950	0.0161768	0.0158745	0.016380	0.025598
3	colDiffs(X[rows, cols])	0.019645	0.0208555	0.0219744	0.0212350	0.021991	0.064434

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colDiffs(X, rows, cols)	1.688802	1.667201	1.688858	1.672320	1.689182	2.017815
3	colDiffs(X[rows, cols])	2.260905	2.229581	2.294131	2.237029	2.267815	5.079142

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.009367	0.0097975	0.0107860	0.0100585	0.0104255	0.026405
2	rowDiffs(X, cols, rows)	0.016303	0.0170330	0.0181807	0.0175700	0.0179835	0.035707
3	rowDiffs(X[cols, rows])	0.022716	0.0234795	0.0258039	0.0242395	0.0267305	0.058727

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	1.740472	1.738505	1.685578	1.746781	1.724953	1.352282
3	rowDiffs(X[cols, rows])	2.425109	2.396479	2.392348	2.409852	2.563954	2.224086

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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	colDiffs_X_S	8.689	9.3540	9.57852	9.4925	9.6970	12.686
2	rowDiffs_X_S	9.367	9.7975	10.78601	10.0585	10.4255	26.405

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs_X_S	1.07803	1.047413	1.126062	1.059626	1.075126	2.081428

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 5216545 278.6    7916910 422.9  7916910 422.9
Vcells 9569363  73.1   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5216539 278.6    7916910 422.9  7916910 422.9
Vcells 9574446  73.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs_X_S	0.008169	0.0087070	0.0093797	0.0088880	0.0090550	0.036902
2	colDiffs(X, rows, cols)	0.014261	0.0148310	0.0155227	0.0152705	0.0156240	0.029455
3	colDiffs(X[rows, cols])	0.020580	0.0217685	0.0227336	0.0224755	0.0228255	0.035877

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colDiffs(X, rows, cols)	1.745746	1.703342	1.654926	1.718103	1.725456	0.7981952
3	colDiffs(X[rows, cols])	2.519280	2.500115	2.423710	2.528747	2.520762	0.9722237

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.007989	0.0083965	0.0088602	0.0085775	0.0088745	0.015705
2	rowDiffs(X, cols, rows)	0.012948	0.0135115	0.0143816	0.0138020	0.0142520	0.057249
3	rowDiffs(X[cols, rows])	0.019107	0.0201940	0.0212245	0.0203955	0.0212220	0.039549

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	1.620729	1.609182	1.623172	1.609094	1.605950	3.645272
3	rowDiffs(X[cols, rows])	2.391663	2.405050	2.395485	2.377791	2.391346	2.518243

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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	rowDiffs_X_S	7.989	8.3965	8.86020	8.5775	8.8745	15.705
1	colDiffs_X_S	8.169	8.7070	9.37968	8.8880	9.0550	36.902

	expr	min	lq	mean	median	uq	max
2	rowDiffs_X_S	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
1	colDiffs_X_S	1.022531	1.03698	1.058631	1.036199	1.020339	2.349697

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 5216756 278.7    7916910 422.9  7916910 422.9
Vcells 9592026  73.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5216750 278.7    7916910 422.9  7916910 422.9
Vcells 9642109  73.6   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs_X_S	0.049948	0.0534205	0.0582087	0.0552685	0.0605225	0.109774
2	colDiffs(X, rows, cols)	0.092337	0.0988005	0.1076287	0.1020370	0.1123740	0.192729
3	colDiffs(X[rows, cols])	0.130857	0.1395655	0.1544117	0.1474420	0.1620730	0.318177

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colDiffs(X, rows, cols)	1.848663	1.849487	1.849014	1.846205	1.856731	1.755689
3	colDiffs(X[rows, cols])	2.619865	2.612583	2.652727	2.667740	2.677897	2.898473

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.051704	0.0542250	0.0618808	0.0595485	0.0661530	0.110843
2	rowDiffs(X, cols, rows)	0.088411	0.0916985	0.1004919	0.0952035	0.1068870	0.181371
3	rowDiffs(X[cols, rows])	0.131275	0.1363815	0.1510804	0.1457250	0.1605755	0.234687

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	1.709945	1.691074	1.623960	1.598756	1.615754	1.636287
3	rowDiffs(X[cols, rows])	2.538972	2.515104	2.441475	2.447165	2.427335	2.117292

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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	colDiffs_X_S	49.948	53.4205	58.20868	55.2685	60.5225	109.774
2	rowDiffs_X_S	51.704	54.2250	61.88078	59.5485	66.1530	110.843

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.00000	1.000000	1.00000	1.000000	1.000000
2	rowDiffs_X_S	1.035157	1.01506	1.063085	1.07744	1.093031	1.009738

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 5216966 278.7    7916910 422.9  7916910 422.9
Vcells 9592729  73.2   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5216960 278.7    7916910 422.9  7916910 422.9
Vcells 9642812  73.6   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs_X_S	0.050507	0.0525055	0.0575358	0.0542325	0.0601415	0.105880
2	colDiffs(X, rows, cols)	0.090679	0.0941415	0.1047782	0.1004060	0.1112930	0.223794
3	colDiffs(X[rows, cols])	0.127945	0.1332485	0.1471991	0.1424475	0.1563025	0.242569

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colDiffs(X, rows, cols)	1.795375	1.792984	1.821096	1.851399	1.850519	2.113657
3	colDiffs(X[rows, cols])	2.533213	2.537801	2.558393	2.626608	2.598913	2.290980

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.051975	0.0542915	0.0612367	0.0579800	0.0658050	0.131392
2	rowDiffs(X, cols, rows)	0.093554	0.0983985	0.1099080	0.1046005	0.1147120	0.192893
3	rowDiffs(X[cols, rows])	0.136226	0.1429505	0.1575831	0.1526775	0.1679315	0.230444

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	1.799981	1.812411	1.794806	1.804079	1.743211	1.468073
3	rowDiffs(X[cols, rows])	2.620991	2.633018	2.573345	2.633279	2.551957	1.753866

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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	colDiffs_X_S	50.507	52.5055	57.53578	54.2325	60.1415	105.880
2	rowDiffs_X_S	51.975	54.2915	61.23669	57.9800	65.8050	131.392

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.00000	1.000000
2	rowDiffs_X_S	1.029065	1.034015	1.064324	1.069101	1.09417	1.240952

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 5217184 278.7    7916910 422.9  7916910 422.9
Vcells 9683814  73.9   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5217169 278.7    7916910 422.9  7916910 422.9
Vcells 9683982  73.9   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs_X_S	0.002598	0.0028820	0.0033337	0.0029690	0.0030740	0.036233
2	colDiffs(X, rows, cols)	0.002965	0.0032415	0.0034179	0.0033460	0.0034925	0.005915
3	colDiffs(X[rows, cols])	0.003580	0.0039650	0.0041717	0.0040425	0.0042085	0.010714

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colDiffs(X, rows, cols)	1.141263	1.124740	1.025266	1.126979	1.136142	0.1632490
3	colDiffs(X[rows, cols])	1.377983	1.375781	1.251378	1.361570	1.369063	0.2956973

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.002587	0.002927	0.0031885	0.0030325	0.0031845	0.005777
2	rowDiffs(X, cols, rows)	0.002990	0.003338	0.0039580	0.0034180	0.0036160	0.044428
3	rowDiffs(X[cols, rows])	0.003640	0.004029	0.0042729	0.0041380	0.0042915	0.009662

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	1.155779	1.140417	1.241316	1.127123	1.135500	7.690497
3	rowDiffs(X[cols, rows])	1.407035	1.376495	1.340096	1.364551	1.347621	1.672494

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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	colDiffs_X_S	2.598	2.882	3.33370	2.9690	3.0740	36.233
2	rowDiffs_X_S	2.587	2.927	3.18851	3.0325	3.1845	5.777

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
2	rowDiffs_X_S	0.995766	1.015614	0.9564478	1.021388	1.035947	0.1594403

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 5217383 278.7    7916910 422.9  7916910 422.9
Vcells 9689746  74.0   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5217377 278.7    7916910 422.9  7916910 422.9
Vcells 9699829  74.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs_X_S	0.007164	0.0083080	0.0098073	0.0089865	0.011152	0.015938
2	colDiffs(X, rows, cols)	0.013762	0.0146380	0.0175394	0.0151860	0.019411	0.052173
3	colDiffs(X[rows, cols])	0.023134	0.0245945	0.0279257	0.0253645	0.028766	0.076470

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colDiffs(X, rows, cols)	1.920994	1.761916	1.788405	1.689868	1.740585	3.273497
3	colDiffs(X[rows, cols])	3.229202	2.960339	2.847443	2.822512	2.579448	4.797967

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.006891	0.0080755	0.0085462	0.0085930	0.0089015	0.014724
2	rowDiffs(X, cols, rows)	0.020214	0.0206815	0.0217297	0.0210985	0.0215935	0.050693
3	rowDiffs(X[cols, rows])	0.022603	0.0241970	0.0253184	0.0248330	0.0254855	0.055220

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	2.933391	2.561018	2.542609	2.455312	2.425827	3.442882
3	rowDiffs(X[cols, rows])	3.280076	2.996347	2.962531	2.889910	2.863057	3.750340

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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	rowDiffs_X_S	6.891	8.0755	8.54621	8.5930	8.9015	14.724
1	colDiffs_X_S	7.164	8.3080	9.80730	8.9865	11.1520	15.938

	expr	min	lq	mean	median	uq	max
2	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
1	colDiffs_X_S	1.039617	1.028791	1.147561	1.045793	1.252823	1.08245

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 5217582 278.7    7916910 422.9  7916910 422.9
Vcells 9691124  74.0   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5217576 278.7    7916910 422.9  7916910 422.9
Vcells 9701207  74.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs_X_S	0.007125	0.007709	0.0079392	0.0079070	0.0081005	0.011660
2	colDiffs(X, rows, cols)	0.014359	0.015187	0.0157568	0.0155245	0.0157385	0.029336
3	colDiffs(X[rows, cols])	0.022770	0.023781	0.0251061	0.0243505	0.0251880	0.068829

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colDiffs(X, rows, cols)	2.015298	1.970035	1.984667	1.963387	1.942905	2.515952
3	colDiffs(X[rows, cols])	3.195790	3.084836	3.162278	3.079613	3.109438	5.903002

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.007531	0.0087035	0.0099673	0.0090945	0.010117	0.032648
2	rowDiffs(X, cols, rows)	0.024179	0.0245890	0.0254423	0.0248090	0.025324	0.038812
3	rowDiffs(X[cols, rows])	0.027785	0.0288615	0.0323594	0.0295170	0.031131	0.097305

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	3.210596	2.825185	2.552587	2.727912	2.503114	1.188802
3	rowDiffs(X[cols, rows])	3.689417	3.316080	3.246565	3.245588	3.077098	2.980428

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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
1	colDiffs_X_S	7.125	7.7090	7.93924	7.9070	8.1005	11.660
2	rowDiffs_X_S	7.531	8.7035	9.96727	9.0945	10.1170	32.648

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0
2	rowDiffs_X_S	1.056982	1.129005	1.255444	1.150183	1.248935	2.8

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 5217787 278.7    7916910 422.9  7916910 422.9
Vcells 9691260  74.0   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5217781 278.7    7916910 422.9  7916910 422.9
Vcells 9701343  74.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs_X_S	0.006839	0.0076695	0.0087254	0.0079775	0.0081955	0.044311
2	colDiffs(X, rows, cols)	0.013908	0.0147035	0.0154618	0.0151300	0.0154845	0.034809
3	colDiffs(X[rows, cols])	0.024994	0.0261475	0.0275320	0.0265645	0.0272785	0.051190

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colDiffs(X, rows, cols)	2.033631	1.917139	1.772046	1.896584	1.889390	0.7855611
3	colDiffs(X[rows, cols])	3.654628	3.409283	3.155388	3.329928	3.328473	1.1552436

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.006065	0.0070640	0.0078138	0.0078115	0.0082765	0.012731
2	rowDiffs(X, cols, rows)	0.017590	0.0185175	0.0198295	0.0191790	0.0198720	0.050255
3	rowDiffs(X[cols, rows])	0.021654	0.0231510	0.0249392	0.0247880	0.0254340	0.040921

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	2.900247	2.621390	2.537756	2.455226	2.401015	3.947451
3	rowDiffs(X[cols, rows])	3.570321	3.277322	3.191686	3.173270	3.073038	3.214280

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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	rowDiffs_X_S	6.065	7.0640	7.81381	7.8115	8.2765	12.731
1	colDiffs_X_S	6.839	7.6695	8.72539	7.9775	8.1955	44.311

	expr	min	lq	mean	median	uq	max
2	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.0000000	1.000000
1	colDiffs_X_S	1.127617	1.085716	1.116663	1.021251	0.9902133	3.480559

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 5217998 278.7    7916910 422.9  7916910 422.9
Vcells 9736691  74.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5217992 278.7    7916910 422.9  7916910 422.9
Vcells 9836774  75.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs_X_S	0.041705	0.0445245	0.0648304	0.0473525	0.0515275	0.182889
2	colDiffs(X, rows, cols)	0.097882	0.0993725	0.1230300	0.1098810	0.1188305	0.253727
3	colDiffs(X[rows, cols])	0.171747	0.1806615	0.2187814	0.1943055	0.2081430	0.451833

	expr	min	lq	mean	median	uq	max
1	colDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colDiffs(X, rows, cols)	2.347009	2.231861	1.897720	2.320490	2.306157	1.387328
3	colDiffs(X[rows, cols])	4.118139	4.057575	3.374672	4.103384	4.039455	2.470531

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.037139	0.0391735	0.0438025	0.0411120	0.0444100	0.114574
2	rowDiffs(X, cols, rows)	0.123215	0.1275580	0.1386520	0.1318980	0.1460495	0.259629
3	rowDiffs(X[cols, rows])	0.152225	0.1627130	0.1749235	0.1689935	0.1834580	0.322587

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	3.317671	3.256232	3.165392	3.208260	3.288663	2.266038
3	rowDiffs(X[cols, rows])	4.098791	4.153650	3.993463	4.110564	4.131006	2.815534

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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	rowDiffs_X_S	37.139	39.1735	43.80246	41.1120	44.4100	114.574
1	colDiffs_X_S	41.705	44.5245	64.83041	47.3525	51.5275	182.889

	expr	min	lq	mean	median	uq	max
2	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colDiffs_X_S	1.122944	1.136597	1.480063	1.151793	1.160268	1.596252

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 5218208 278.7    7916910 422.9  7916910 422.9
Vcells 9736832  74.3   33191153 253.3 53339345 407.0
> colStats <- microbenchmark(colDiffs_X_S = colDiffs(X_S), `colDiffs(X, rows, cols)` = colDiffs(X, 
+     rows = rows, cols = cols), `colDiffs(X[rows, cols])` = colDiffs(X[rows, cols]), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5218202 278.7    7916910 422.9  7916910 422.9
Vcells 9836915  75.1   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowDiffs_X_S = rowDiffs(X_S), `rowDiffs(X, cols, rows)` = rowDiffs(X, 
+     rows = cols, cols = rows), `rowDiffs(X[cols, rows])` = rowDiffs(X[cols, rows]), unit = "ms")

Table: Benchmarking of colDiffs_X_S(), colDiffs(X, rows, cols)() and colDiffs(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	colDiffs(X, rows, cols)	0.102611	0.1117310	0.1735545	0.1233620	0.2333815	0.472639
1	colDiffs_X_S	0.045599	0.0576005	0.1264480	0.1666655	0.1816485	0.285409
3	colDiffs(X[rows, cols])	0.184446	0.2047845	0.3120772	0.2313235	0.4371380	0.671058

	expr	min	lq	mean	median	uq	max
2	colDiffs(X, rows, cols)	1.0000000	1.0000000	1.000000	1.000000	1.0000000	1.0000000
1	colDiffs_X_S	0.4443871	0.5155284	0.728578	1.351028	0.7783329	0.6038626
3	colDiffs(X[rows, cols])	1.7975266	1.8328351	1.798152	1.875160	1.8730619	1.4198109

Table: Benchmarking of rowDiffs_X_S(), rowDiffs(X, cols, rows)() and rowDiffs(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	rowDiffs_X_S	0.039184	0.040845	0.0467808	0.0440665	0.049739	0.108445
2	rowDiffs(X, cols, rows)	0.139614	0.144773	0.1610847	0.1543450	0.167373	0.289987
3	rowDiffs(X[cols, rows])	0.163911	0.170386	0.1886539	0.1813995	0.203806	0.279092

	expr	min	lq	mean	median	uq	max
1	rowDiffs_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowDiffs(X, cols, rows)	3.563036	3.544449	3.443395	3.502547	3.365025	2.674047
3	rowDiffs(X[cols, rows])	4.183110	4.171526	4.032722	4.116494	4.097509	2.573581

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

Table: Benchmarking of colDiffs_X_S() and rowDiffs_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	rowDiffs_X_S	39.184	40.8450	46.78078	44.0665	49.7390	108.445
1	colDiffs_X_S	45.599	57.6005	126.44796	166.6655	181.6485	285.409

	expr	min	lq	mean	median	uq	max
2	rowDiffs_X_S	1.000000	1.000000	1.00000	1.000000	1.000000	1.000000
1	colDiffs_X_S	1.163715	1.410222	2.70299	3.782136	3.652034	2.631832

Figure: Benchmarking of colDiffs_X_S() and rowDiffs_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 23.15 secs.

Reproducibility

To reproduce this report, do:

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

---

Copyright Dongcan Jiang. Last updated on 2021-08-25 22:15:36 (+0200 UTC). Powered by RSP.