matrixStats.benchmarks

colQuantiles() and rowQuantiles() benchmarks on subsetted computation

This report benchmark the performance of colQuantiles() and rowQuantiles() on subsetted computation.

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 = "double")

Results

10x10 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  5291916 282.7    7916910 422.9  7916910 422.9
Vcells 10501223  80.2   33191153 253.3 53339345 407.0
> probs <- seq(from = 0, to = 1, by = 0.25)
> colStats <- microbenchmark(colQuantiles_X_S = colQuantiles(X_S, probs = probs, na.rm = FALSE), `colQuantiles(X, rows, cols)` = colQuantiles(X, 
+     rows = rows, cols = cols, probs = probs, na.rm = FALSE), `colQuantiles(X[rows, cols])` = colQuantiles(X[rows, 
+     cols], probs = probs, na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5280058 282.0    7916910 422.9  7916910 422.9
Vcells 10461111  79.9   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowQuantiles_X_S = rowQuantiles(X_S, probs = probs, na.rm = FALSE), `rowQuantiles(X, cols, rows)` = rowQuantiles(X, 
+     rows = cols, cols = rows, probs = probs, na.rm = FALSE), `rowQuantiles(X[cols, rows])` = rowQuantiles(X[cols, 
+     rows], probs = probs, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 10x10 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
3	colQuantiles(X[rows, cols])	0.193864	0.2191920	0.2395020	0.2338630	0.250885	0.325129
1	colQuantiles_X_S	0.191803	0.2195265	0.2481782	0.2345925	0.259328	1.006207
2	colQuantiles(X, rows, cols)	0.198249	0.2212015	0.2476609	0.2365175	0.268017	0.373556

	expr	min	lq	mean	median	uq	max
3	colQuantiles(X[rows, cols])	1.0000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	colQuantiles_X_S	0.9893688	1.001526	1.036226	1.003119	1.033653	3.094793
2	colQuantiles(X, rows, cols)	1.0226189	1.009168	1.034066	1.011351	1.068286	1.148947

Table: Benchmarking of rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on 10x10 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
3	rowQuantiles(X[cols, rows])	0.190332	0.2057045	0.2267560	0.2148715	0.2395465	0.401357
1	rowQuantiles_X_S	0.187668	0.2024100	0.2304964	0.2232495	0.2462815	0.376155
2	rowQuantiles(X, cols, rows)	0.190403	0.1996565	0.2419617	0.2276835	0.2496985	0.889379

	expr	min	lq	mean	median	uq	max
3	rowQuantiles(X[cols, rows])	1.0000000	1.0000000	1.000000	1.000000	1.000000	1.000000
1	rowQuantiles_X_S	0.9860034	0.9839843	1.016495	1.038991	1.028116	0.937208
2	rowQuantiles(X, cols, rows)	1.0003730	0.9705986	1.067057	1.059626	1.042380	2.215930

Figure: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 10x10 data as well as rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

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

	expr	min	lq	mean	median	uq	max
2	rowQuantiles_X_S	187.668	202.4100	230.4964	223.2495	246.2815	376.155
1	colQuantiles_X_S	191.803	219.5265	248.1782	234.5925	259.3280	1006.207

	expr	min	lq	mean	median	uq	max
2	rowQuantiles_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.00000
1	colQuantiles_X_S	1.022034	1.084564	1.076712	1.050809	1.052974	2.67498

Figure: Benchmarking of colQuantiles_X_S() and rowQuantiles_X_S() on 10x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x100 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  5279660 282.0    7916910 422.9  7916910 422.9
Vcells 10134066  77.4   33191153 253.3 53339345 407.0
> probs <- seq(from = 0, to = 1, by = 0.25)
> colStats <- microbenchmark(colQuantiles_X_S = colQuantiles(X_S, probs = probs, na.rm = FALSE), `colQuantiles(X, rows, cols)` = colQuantiles(X, 
+     rows = rows, cols = cols, probs = probs, na.rm = FALSE), `colQuantiles(X[rows, cols])` = colQuantiles(X[rows, 
+     cols], probs = probs, na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5279648 282.0    7916910 422.9  7916910 422.9
Vcells 10144139  77.4   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowQuantiles_X_S = rowQuantiles(X_S, probs = probs, na.rm = FALSE), `rowQuantiles(X, cols, rows)` = rowQuantiles(X, 
+     rows = cols, cols = rows, probs = probs, na.rm = FALSE), `rowQuantiles(X[cols, rows])` = rowQuantiles(X[cols, 
+     rows], probs = probs, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 100x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	1.225778	1.252869	1.354780	1.286105	1.353386	2.311531
2	colQuantiles(X, rows, cols)	1.245639	1.269805	1.376652	1.304566	1.387719	2.182813
3	colQuantiles(X[rows, cols])	1.244523	1.264433	1.471976	1.311001	1.435142	10.505936

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colQuantiles(X, rows, cols)	1.016203	1.013517	1.016144	1.014354	1.025369	0.9443148
3	colQuantiles(X[rows, cols])	1.015292	1.009230	1.086506	1.019358	1.060408	4.5450119

Table: Benchmarking of rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on 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	rowQuantiles_X_S	1.255388	1.272221	1.379842	1.289094	1.372680	2.192171
3	rowQuantiles(X[cols, rows])	1.268772	1.281665	1.369053	1.296639	1.364815	2.194366
2	rowQuantiles(X, cols, rows)	1.265596	1.282952	1.477202	1.300990	1.368350	10.359479

	expr	min	lq	mean	median	uq	max
1	rowQuantiles_X_S	1.000000	1.000000	1.0000000	1.000000	1.0000000	1.000000
3	rowQuantiles(X[cols, rows])	1.010661	1.007423	0.9921812	1.005853	0.9942700	1.001001
2	rowQuantiles(X, cols, rows)	1.008131	1.008434	1.0705588	1.009229	0.9968452	4.725671

Figure: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 100x100 data as well as rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

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

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	1.225778	1.252869	1.354780	1.286105	1.353386	2.311531
2	rowQuantiles_X_S	1.255388	1.272221	1.379842	1.289094	1.372680	2.192171

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowQuantiles_X_S	1.024156	1.015446	1.018499	1.002324	1.014256	0.9483632

Figure: Benchmarking of colQuantiles_X_S() and rowQuantiles_X_S() on 100x100 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x10 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  5280409 282.1    7916910 422.9  7916910 422.9
Vcells 10138126  77.4   33191153 253.3 53339345 407.0
> probs <- seq(from = 0, to = 1, by = 0.25)
> colStats <- microbenchmark(colQuantiles_X_S = colQuantiles(X_S, probs = probs, na.rm = FALSE), `colQuantiles(X, rows, cols)` = colQuantiles(X, 
+     rows = rows, cols = cols, probs = probs, na.rm = FALSE), `colQuantiles(X[rows, cols])` = colQuantiles(X[rows, 
+     cols], probs = probs, na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5280391 282.1    7916910 422.9  7916910 422.9
Vcells 10148189  77.5   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowQuantiles_X_S = rowQuantiles(X_S, probs = probs, na.rm = FALSE), `rowQuantiles(X, cols, rows)` = rowQuantiles(X, 
+     rows = cols, cols = rows, probs = probs, na.rm = FALSE), `rowQuantiles(X[cols, rows])` = rowQuantiles(X[cols, 
+     rows], probs = probs, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 1000x10 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	0.370728	0.3727030	0.4200224	0.381484	0.4541030	0.617224
3	colQuantiles(X[rows, cols])	0.382117	0.3861615	0.4301175	0.390001	0.4528365	0.825006
2	colQuantiles(X, rows, cols)	0.381225	0.3864250	0.4304690	0.390269	0.4602770	0.692628

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	colQuantiles(X[rows, cols])	1.030721	1.036110	1.024035	1.022326	0.997211	1.336639
2	colQuantiles(X, rows, cols)	1.028315	1.036817	1.024872	1.023028	1.013596	1.122166

Table: Benchmarking of rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on 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	rowQuantiles_X_S	0.385827	0.388216	0.4227291	0.3912555	0.4255555	0.616402
3	rowQuantiles(X[cols, rows])	0.399713	0.404013	0.4412166	0.4061035	0.4605050	0.817110
2	rowQuantiles(X, cols, rows)	0.402190	0.404894	0.4423714	0.4122980	0.4760700	0.621445

	expr	min	lq	mean	median	uq	max
1	rowQuantiles_X_S	1.00000	1.000000	1.000000	1.000000	1.000000	1.000000
3	rowQuantiles(X[cols, rows])	1.03599	1.040691	1.043734	1.037950	1.082127	1.325612
2	rowQuantiles(X, cols, rows)	1.04241	1.042961	1.046465	1.053782	1.118702	1.008181

Figure: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 1000x10 data as well as rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colQuantiles_X_S() and rowQuantiles_X_S() on 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	colQuantiles_X_S	370.728	372.703	420.0224	381.4840	454.1030	617.224
2	rowQuantiles_X_S	385.827	388.216	422.7291	391.2555	425.5555	616.402

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	1.000000	1.000000	1.000000	1.000000	1.0000000	1.0000000
2	rowQuantiles_X_S	1.040728	1.041623	1.006444	1.025614	0.9371343	0.9986682

Figure: Benchmarking of colQuantiles_X_S() and rowQuantiles_X_S() on 1000x10 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

10x1000 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  5280608 282.1    7916910 422.9  7916910 422.9
Vcells 10139102  77.4   33191153 253.3 53339345 407.0
> probs <- seq(from = 0, to = 1, by = 0.25)
> colStats <- microbenchmark(colQuantiles_X_S = colQuantiles(X_S, probs = probs, na.rm = FALSE), `colQuantiles(X, rows, cols)` = colQuantiles(X, 
+     rows = rows, cols = cols, probs = probs, na.rm = FALSE), `colQuantiles(X[rows, cols])` = colQuantiles(X[rows, 
+     cols], probs = probs, na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5280596 282.1    7916910 422.9  7916910 422.9
Vcells 10149175  77.5   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowQuantiles_X_S = rowQuantiles(X_S, probs = probs, na.rm = FALSE), `rowQuantiles(X, cols, rows)` = rowQuantiles(X, 
+     rows = cols, cols = rows, probs = probs, na.rm = FALSE), `rowQuantiles(X[cols, rows])` = rowQuantiles(X[cols, 
+     rows], probs = probs, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 10x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	colQuantiles(X, rows, cols)	9.311982	9.526577	10.76101	9.882511	10.57142	18.86934
3	colQuantiles(X[rows, cols])	9.336797	9.664976	10.59511	9.965283	10.35032	17.60338
1	colQuantiles_X_S	9.322680	9.623390	10.29404	10.035808	10.43551	16.82942

	expr	min	lq	mean	median	uq	max
2	colQuantiles(X, rows, cols)	1.000000	1.000000	1.0000000	1.000000	1.0000000	1.0000000
3	colQuantiles(X[rows, cols])	1.002665	1.014528	0.9845829	1.008376	0.9790849	0.9329090
1	colQuantiles_X_S	1.001149	1.010162	0.9566059	1.015512	0.9871438	0.8918923

Table: Benchmarking of rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on 10x1000 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
3	rowQuantiles(X[cols, rows])	9.318816	9.671371	10.38362	9.922601	10.47725	18.05593
2	rowQuantiles(X, cols, rows)	9.418175	9.712003	11.03026	10.020728	10.98179	24.69222
1	rowQuantiles_X_S	9.355191	9.644142	10.40783	10.024606	10.32139	16.76068

	expr	min	lq	mean	median	uq	max
3	rowQuantiles(X[cols, rows])	1.000000	1.0000000	1.000000	1.000000	1.0000000	1.0000000
2	rowQuantiles(X, cols, rows)	1.010662	1.0042013	1.062275	1.009889	1.0481553	1.3675403
1	rowQuantiles_X_S	1.003903	0.9971846	1.002331	1.010280	0.9851235	0.9282643

Figure: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 10x1000 data as well as rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colQuantiles_X_S() and rowQuantiles_X_S() on 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	rowQuantiles_X_S	9.355191	9.644142	10.40783	10.02461	10.32139	16.76068
1	colQuantiles_X_S	9.322680	9.623390	10.29404	10.03581	10.43551	16.82942

	expr	min	lq	mean	median	uq	max
2	rowQuantiles_X_S	1.0000000	1.0000000	1.0000000	1.000000	1.000000	1.000000
1	colQuantiles_X_S	0.9965248	0.9978482	0.9890675	1.001118	1.011057	1.004102

Figure: Benchmarking of colQuantiles_X_S() and rowQuantiles_X_S() on 10x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

100x1000 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  5280818 282.1    7916910 422.9  7916910 422.9
Vcells 10183847  77.7   33191153 253.3 53339345 407.0
> probs <- seq(from = 0, to = 1, by = 0.25)
> colStats <- microbenchmark(colQuantiles_X_S = colQuantiles(X_S, probs = probs, na.rm = FALSE), `colQuantiles(X, rows, cols)` = colQuantiles(X, 
+     rows = rows, cols = cols, probs = probs, na.rm = FALSE), `colQuantiles(X[rows, cols])` = colQuantiles(X[rows, 
+     cols], probs = probs, na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5280806 282.1    7916910 422.9  7916910 422.9
Vcells 10283920  78.5   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowQuantiles_X_S = rowQuantiles(X_S, probs = probs, na.rm = FALSE), `rowQuantiles(X, cols, rows)` = rowQuantiles(X, 
+     rows = cols, cols = rows, probs = probs, na.rm = FALSE), `rowQuantiles(X[cols, rows])` = rowQuantiles(X[cols, 
+     rows], probs = probs, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 100x1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	11.52600	12.38966	14.61118	13.30258	14.56454	38.26557
3	colQuantiles(X[rows, cols])	11.69649	12.47478	14.41368	13.47034	14.54926	26.50288
2	colQuantiles(X, rows, cols)	11.78100	12.46030	15.42942	13.92336	15.77830	28.62961

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	1.000000	1.000000	1.0000000	1.000000	1.0000000	1.0000000
3	colQuantiles(X[rows, cols])	1.014791	1.006870	0.9864831	1.012611	0.9989507	0.6926037
2	colQuantiles(X, rows, cols)	1.022124	1.005702	1.0560009	1.046666	1.0833364	0.7481819

Table: Benchmarking of rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on 100x1000 data (transposed). The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	rowQuantiles(X, cols, rows)	12.36452	13.40608	14.51832	13.71598	14.47456	30.05745
1	rowQuantiles_X_S	12.16054	13.33660	14.59470	13.80858	15.02680	24.57460
3	rowQuantiles(X[cols, rows])	12.19137	13.44041	14.51761	13.81128	14.67734	26.34513

	expr	min	lq	mean	median	uq	max
2	rowQuantiles(X, cols, rows)	1.0000000	1.0000000	1.0000000	1.000000	1.000000	1.0000000
1	rowQuantiles_X_S	0.9835027	0.9948169	1.0052609	1.006751	1.038153	0.8175878
3	rowQuantiles(X[cols, rows])	0.9859963	1.0025601	0.9999514	1.006948	1.014010	0.8764925

Figure: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 100x1000 data as well as rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colQuantiles_X_S() and rowQuantiles_X_S() on 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	colQuantiles_X_S	11.52600	12.38966	14.61118	13.30258	14.56454	38.26557
2	rowQuantiles_X_S	12.16054	13.33660	14.59470	13.80858	15.02680	24.57460

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	1.000000	1.00000	1.0000000	1.000000	1.000000	1.0000000
2	rowQuantiles_X_S	1.055052	1.07643	0.9988718	1.038038	1.031739	0.6422119

Figure: Benchmarking of colQuantiles_X_S() and rowQuantiles_X_S() on 100x1000 data (original and transposed). Outliers are displayed as crosses. Times are in milliseconds.

1000x100 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  5281031 282.1    7916910 422.9  7916910 422.9
Vcells 10184677  77.8   33191153 253.3 53339345 407.0
> probs <- seq(from = 0, to = 1, by = 0.25)
> colStats <- microbenchmark(colQuantiles_X_S = colQuantiles(X_S, probs = probs, na.rm = FALSE), `colQuantiles(X, rows, cols)` = colQuantiles(X, 
+     rows = rows, cols = cols, probs = probs, na.rm = FALSE), `colQuantiles(X[rows, cols])` = colQuantiles(X[rows, 
+     cols], probs = probs, na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5281019 282.1    7916910 422.9  7916910 422.9
Vcells 10284750  78.5   33191153 253.3 53339345 407.0
> rowStats <- microbenchmark(rowQuantiles_X_S = rowQuantiles(X_S, probs = probs, na.rm = FALSE), `rowQuantiles(X, cols, rows)` = rowQuantiles(X, 
+     rows = cols, cols = rows, probs = probs, na.rm = FALSE), `rowQuantiles(X[cols, rows])` = rowQuantiles(X[cols, 
+     rows], probs = probs, na.rm = FALSE), unit = "ms")

Table: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 1000x100 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	2.880728	3.120188	3.413697	3.210253	3.330167	13.560352
3	colQuantiles(X[rows, cols])	3.061107	3.220135	3.391236	3.334631	3.465556	4.785324
2	colQuantiles(X, rows, cols)	3.096268	3.252244	3.522540	3.346129	3.467803	13.800664

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
3	colQuantiles(X[rows, cols])	1.062616	1.032032	0.9934202	1.038744	1.040655	0.3528908
2	colQuantiles(X, rows, cols)	1.074821	1.042323	1.0318841	1.042326	1.041330	1.0177217

Table: Benchmarking of rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on 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	rowQuantiles_X_S	3.090674	3.472739	3.580201	3.557223	3.642758	4.549436
3	rowQuantiles(X[cols, rows])	3.337567	3.637268	3.807218	3.754910	3.893598	6.240238
2	rowQuantiles(X, cols, rows)	3.339166	3.652345	4.036156	3.765739	3.932219	15.623647

	expr	min	lq	mean	median	uq	max
1	rowQuantiles_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
3	rowQuantiles(X[cols, rows])	1.079883	1.047378	1.063409	1.055573	1.068860	1.371651
2	rowQuantiles(X, cols, rows)	1.080401	1.051719	1.127355	1.058618	1.079462	3.434194

Figure: Benchmarking of colQuantiles_X_S(), colQuantiles(X, rows, cols)() and colQuantiles(X[rows, cols])() on 1000x100 data as well as rowQuantiles_X_S(), rowQuantiles(X, cols, rows)() and rowQuantiles(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colQuantiles_X_S() and rowQuantiles_X_S() on 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	colQuantiles_X_S	2.880728	3.120188	3.413697	3.210253	3.330167	13.560352
2	rowQuantiles_X_S	3.090674	3.472739	3.580201	3.557223	3.642758	4.549436

	expr	min	lq	mean	median	uq	max
1	colQuantiles_X_S	1.000000	1.00000	1.000000	1.000000	1.000000	1.0000000
2	rowQuantiles_X_S	1.072879	1.11299	1.048775	1.108082	1.093867	0.3354954

Figure: Benchmarking of colQuantiles_X_S() and rowQuantiles_X_S() on 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 30.22 secs.

Reproducibility

To reproduce this report, do:

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