matrixStats.benchmarks

---

mean2() benchmarks on subsetted computation

This report benchmark the performance of mean2() on subsetted computation.

Data type “integer”

Data

> rvector <- function(n, mode = c("logical", "double", "integer"), range = c(-100, +100), na_prob = 0) {
+     mode <- match.arg(mode)
+     if (mode == "logical") {
+         x <- sample(c(FALSE, TRUE), size = n, replace = TRUE)
+     }     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
+     x
+ }
> rvectors <- function(scale = 10, seed = 1, ...) {
+     set.seed(seed)
+     data <- list()
+     data[[1]] <- rvector(n = scale * 100, ...)
+     data[[2]] <- rvector(n = scale * 1000, ...)
+     data[[3]] <- rvector(n = scale * 10000, ...)
+     data[[4]] <- rvector(n = scale * 1e+05, ...)
+     data[[5]] <- rvector(n = scale * 1e+06, ...)
+     names(data) <- sprintf("n = %d", sapply(data, FUN = length))
+     data
+ }
> data <- rvectors(mode = mode)

Results

n = 1000 vector

> x <- data[["n = 1000"]]
> idxs <- sample.int(length(x), size = length(x) * 0.7)
> x_S <- x[idxs]
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5346229 285.6    7916910 422.9  7916910 422.9
Vcells 17014748 129.9   39539356 301.7 94934136 724.3
> stats <- microbenchmark(mean2_x_S = mean2(x_S, refine = TRUE), mean2_x_S_no_refine = mean2(x_S, refine = FALSE), 
+     `mean2(x, idxs)` = mean2(x, idxs = idxs, refine = TRUE), `mean2_no_refine(x, idxs)` = mean2(x, 
+         idxs = idxs, refine = FALSE), `mean2(x[idxs])` = mean2(x[idxs], refine = TRUE), `mean2_no_refine(x[idxs])` = mean2(x[idxs], 
+         refine = FALSE), unit = "ms")

Table: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on integer+n = 1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	0.002679	0.0027150	0.0028934	0.0027715	0.0028810	0.009689
1	mean2_x_S	0.002662	0.0027160	0.0028218	0.0027740	0.0028430	0.003517
4	mean2_no_refine(x, idxs)	0.004343	0.0044340	0.0045670	0.0044755	0.0045800	0.008436
3	mean2(x, idxs)	0.004358	0.0044315	0.0045226	0.0044845	0.0045545	0.005723
5	mean2(x[idxs])	0.005448	0.0056455	0.0058082	0.0057385	0.0059025	0.007304
6	mean2_no_refine(x[idxs])	0.005491	0.0056645	0.0075553	0.0057805	0.0059085	0.177777

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	1.0000000	1.000000	1.0000000	1.000000	1.0000000	1.0000000
1	mean2_x_S	0.9936543	1.000368	0.9752813	1.000902	0.9868101	0.3629890
4	mean2_no_refine(x, idxs)	1.6211273	1.633149	1.5784486	1.614829	1.5897258	0.8706781
3	mean2(x, idxs)	1.6267264	1.632228	1.5630996	1.618077	1.5808747	0.5906698
5	mean2(x[idxs])	2.0335946	2.079374	2.0074274	2.070539	2.0487678	0.7538446
6	mean2_no_refine(x[idxs])	2.0496454	2.086372	2.6112513	2.085694	2.0508504	18.3483332

Figure: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on integer+n = 1000 data. Outliers are displayed as crosses. Times are in milliseconds.

n = 10000 vector

> x <- data[["n = 10000"]]
> idxs <- sample.int(length(x), size = length(x) * 0.7)
> x_S <- x[idxs]
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5343562 285.4    7916910 422.9  7916910 422.9
Vcells 15885751 121.2   39539356 301.7 94934136 724.3
> stats <- microbenchmark(mean2_x_S = mean2(x_S, refine = TRUE), mean2_x_S_no_refine = mean2(x_S, refine = FALSE), 
+     `mean2(x, idxs)` = mean2(x, idxs = idxs, refine = TRUE), `mean2_no_refine(x, idxs)` = mean2(x, 
+         idxs = idxs, refine = FALSE), `mean2(x[idxs])` = mean2(x[idxs], refine = TRUE), `mean2_no_refine(x[idxs])` = mean2(x[idxs], 
+         refine = FALSE), unit = "ms")

Table: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on integer+n = 10000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	mean2_x_S	0.011806	0.0128000	0.0133537	0.0133305	0.0135930	0.024226
2	mean2_x_S_no_refine	0.011800	0.0128075	0.0135270	0.0134160	0.0136760	0.031628
4	mean2_no_refine(x, idxs)	0.025117	0.0271390	0.0287798	0.0283000	0.0285155	0.079392
3	mean2(x, idxs)	0.024984	0.0267295	0.0279374	0.0283230	0.0286770	0.039879
6	mean2_no_refine(x[idxs])	0.031601	0.0341050	0.0358402	0.0356755	0.0362440	0.048649
5	mean2(x[idxs])	0.032194	0.0345280	0.0359147	0.0358030	0.0364095	0.054310

	expr	min	lq	mean	median	uq	max
1	mean2_x_S	1.0000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	mean2_x_S_no_refine	0.9994918	1.000586	1.012974	1.006414	1.006106	1.305540
4	mean2_no_refine(x, idxs)	2.1274776	2.120234	2.155192	2.122951	2.097808	3.277140
3	mean2(x, idxs)	2.1162121	2.088242	2.092106	2.124677	2.109689	1.646124
6	mean2_no_refine(x[idxs])	2.6766898	2.664453	2.683917	2.676231	2.666372	2.008132
5	mean2(x[idxs])	2.7269185	2.697500	2.689494	2.685796	2.678548	2.241806

Figure: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on integer+n = 10000 data. Outliers are displayed as crosses. Times are in milliseconds.

n = 100000 vector

> x <- data[["n = 100000"]]
> idxs <- sample.int(length(x), size = length(x) * 0.7)
> x_S <- x[idxs]
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5343661 285.4    7916910 422.9  7916910 422.9
Vcells 15949329 121.7   39539356 301.7 94934136 724.3
> stats <- microbenchmark(mean2_x_S = mean2(x_S, refine = TRUE), mean2_x_S_no_refine = mean2(x_S, refine = FALSE), 
+     `mean2(x, idxs)` = mean2(x, idxs = idxs, refine = TRUE), `mean2_no_refine(x, idxs)` = mean2(x, 
+         idxs = idxs, refine = FALSE), `mean2(x[idxs])` = mean2(x[idxs], refine = TRUE), `mean2_no_refine(x[idxs])` = mean2(x[idxs], 
+         refine = FALSE), unit = "ms")

Table: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on integer+n = 100000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	0.073474	0.0735800	0.0841975	0.0737455	0.0915115	0.141297
1	mean2_x_S	0.073487	0.0735915	0.0813215	0.0737755	0.0917265	0.115134
3	mean2(x, idxs)	0.175472	0.1757280	0.1931086	0.1758890	0.2018180	0.288645
4	mean2_no_refine(x, idxs)	0.175481	0.1757460	0.1985619	0.1760840	0.2210415	0.302460
5	mean2(x[idxs])	0.223684	0.2252090	0.2606633	0.2270950	0.2903330	0.479048
6	mean2_no_refine(x[idxs])	0.223673	0.2251325	0.2644887	0.2372160	0.2899310	0.435036

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
1	mean2_x_S	1.000177	1.000156	0.9658427	1.000407	1.002349	0.8148368
3	mean2(x, idxs)	2.388219	2.388258	2.2935192	2.385081	2.205384	2.0428247
4	mean2_no_refine(x, idxs)	2.388342	2.388502	2.3582876	2.387725	2.415450	2.1405975
5	mean2(x[idxs])	3.044397	3.060737	3.0958559	3.079442	3.172639	3.3903621
6	mean2_no_refine(x[idxs])	3.044247	3.059697	3.1412896	3.216684	3.168247	3.0788764

Figure: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on integer+n = 100000 data. Outliers are displayed as crosses. Times are in milliseconds.

n = 1000000 vector

> x <- data[["n = 1000000"]]
> idxs <- sample.int(length(x), size = length(x) * 0.7)
> x_S <- x[idxs]
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5343760 285.4    7916910 422.9  7916910 422.9
Vcells 16579615 126.5   39539356 301.7 94934136 724.3
> stats <- microbenchmark(mean2_x_S = mean2(x_S, refine = TRUE), mean2_x_S_no_refine = mean2(x_S, refine = FALSE), 
+     `mean2(x, idxs)` = mean2(x, idxs = idxs, refine = TRUE), `mean2_no_refine(x, idxs)` = mean2(x, 
+         idxs = idxs, refine = FALSE), `mean2(x[idxs])` = mean2(x[idxs], refine = TRUE), `mean2_no_refine(x[idxs])` = mean2(x[idxs], 
+         refine = FALSE), unit = "ms")

Table: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on integer+n = 1000000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	mean2_x_S	0.797124	0.8867120	0.9276662	0.921480	0.964449	1.208392
2	mean2_x_S_no_refine	0.792487	0.8909325	0.9344205	0.930534	0.974582	1.108519
3	mean2(x, idxs)	2.304069	2.5442640	2.8272858	2.739886	2.932145	6.009656
4	mean2_no_refine(x, idxs)	2.330342	2.5830600	2.9430296	2.748366	3.006038	6.111460
5	mean2(x[idxs])	3.028176	3.5891290	4.5712891	4.148489	5.073321	15.984561
6	mean2_no_refine(x[idxs])	3.049728	3.5750345	4.3986760	4.232615	4.926449	14.930561

	expr	min	lq	mean	median	uq	max
1	mean2_x_S	1.0000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	mean2_x_S_no_refine	0.9941828	1.004760	1.007281	1.009826	1.010506	0.9173505
3	mean2(x, idxs)	2.8904775	2.869324	3.047740	2.973354	3.040228	4.9732670
4	mean2_no_refine(x, idxs)	2.9234373	2.913077	3.172509	2.982557	3.116844	5.0575144
5	mean2(x[idxs])	3.7988770	4.047683	4.927731	4.501985	5.260331	13.2279600
6	mean2_no_refine(x[idxs])	3.8259142	4.031788	4.741658	4.593280	5.108045	12.3557265

Figure: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on integer+n = 1000000 data. Outliers are displayed as crosses. Times are in milliseconds.

n = 10000000 vector

> x <- data[["n = 10000000"]]
> idxs <- sample.int(length(x), size = length(x) * 0.7)
> x_S <- x[idxs]
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5343859 285.4    7916910 422.9  7916910 422.9
Vcells 22880280 174.6   39539356 301.7 94934136 724.3
> stats <- microbenchmark(mean2_x_S = mean2(x_S, refine = TRUE), mean2_x_S_no_refine = mean2(x_S, refine = FALSE), 
+     `mean2(x, idxs)` = mean2(x, idxs = idxs, refine = TRUE), `mean2_no_refine(x, idxs)` = mean2(x, 
+         idxs = idxs, refine = FALSE), `mean2(x[idxs])` = mean2(x[idxs], refine = TRUE), `mean2_no_refine(x[idxs])` = mean2(x[idxs], 
+         refine = FALSE), unit = "ms")

Table: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on integer+n = 10000000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
1	mean2_x_S	9.392315	12.13718	13.8573	13.22215	14.70686	20.28899
2	mean2_x_S_no_refine	9.473921	12.62147	14.8040	13.91564	17.32832	22.79262
3	mean2(x, idxs)	83.504952	106.03663	112.2134	113.38030	118.42927	135.76966
4	mean2_no_refine(x, idxs)	86.889893	104.18003	112.2513	113.50355	121.66386	142.11418
6	mean2_no_refine(x[idxs])	114.336552	144.33245	151.0674	152.17082	161.03930	186.60792
5	mean2(x[idxs])	104.054642	141.70683	152.4366	155.60037	165.71799	189.46797

	expr	min	lq	mean	median	uq	max
1	mean2_x_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	mean2_x_S_no_refine	1.008689	1.039901	1.068318	1.052449	1.178247	1.123398
3	mean2(x, idxs)	8.890774	8.736510	8.097789	8.575025	8.052657	6.691789
4	mean2_no_refine(x, idxs)	9.251169	8.583542	8.100524	8.584346	8.272594	7.004497
6	mean2_no_refine(x[idxs])	12.173415	11.891758	10.901648	11.508776	10.949947	9.197496
5	mean2(x[idxs])	11.078700	11.675429	11.000457	11.768156	11.268077	9.338462

Figure: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on integer+n = 10000000 data. Outliers are displayed as crosses. Times are in milliseconds.

Data type “double”

Data

> rvector <- function(n, mode = c("logical", "double", "integer"), range = c(-100, +100), na_prob = 0) {
+     mode <- match.arg(mode)
+     if (mode == "logical") {
+         x <- sample(c(FALSE, TRUE), size = n, replace = TRUE)
+     }     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
+     x
+ }
> rvectors <- function(scale = 10, seed = 1, ...) {
+     set.seed(seed)
+     data <- list()
+     data[[1]] <- rvector(n = scale * 100, ...)
+     data[[2]] <- rvector(n = scale * 1000, ...)
+     data[[3]] <- rvector(n = scale * 10000, ...)
+     data[[4]] <- rvector(n = scale * 1e+05, ...)
+     data[[5]] <- rvector(n = scale * 1e+06, ...)
+     names(data) <- sprintf("n = %d", sapply(data, FUN = length))
+     data
+ }
> data <- rvectors(mode = mode)

Results

n = 1000 vector

> x <- data[["n = 1000"]]
> idxs <- sample.int(length(x), size = length(x) * 0.7)
> x_S <- x[idxs]
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5343961 285.4    7916910 422.9  7916910 422.9
Vcells 21437275 163.6   39539356 301.7 94934136 724.3
> stats <- microbenchmark(mean2_x_S = mean2(x_S, refine = TRUE), mean2_x_S_no_refine = mean2(x_S, refine = FALSE), 
+     `mean2(x, idxs)` = mean2(x, idxs = idxs, refine = TRUE), `mean2_no_refine(x, idxs)` = mean2(x, 
+         idxs = idxs, refine = FALSE), `mean2(x[idxs])` = mean2(x[idxs], refine = TRUE), `mean2_no_refine(x[idxs])` = mean2(x[idxs], 
+         refine = FALSE), unit = "ms")

Table: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on double+n = 1000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	0.002568	0.0026230	0.0027211	0.0026865	0.0027715	0.004095
1	mean2_x_S	0.003718	0.0037865	0.0038905	0.0038435	0.0039230	0.004834
4	mean2_no_refine(x, idxs)	0.004207	0.0042835	0.0043910	0.0043495	0.0044375	0.005166
6	mean2_no_refine(x[idxs])	0.004932	0.0051495	0.0056211	0.0052980	0.0055275	0.029869
3	mean2(x, idxs)	0.005676	0.0057890	0.0060261	0.0058575	0.0059915	0.013501
5	mean2(x[idxs])	0.006182	0.0063395	0.0065188	0.0064440	0.0066095	0.008110

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	mean2_x_S	1.447819	1.443576	1.429779	1.430672	1.415479	1.180464
4	mean2_no_refine(x, idxs)	1.638240	1.633054	1.613703	1.619021	1.601118	1.261538
6	mean2_no_refine(x[idxs])	1.920561	1.963210	2.065761	1.972083	1.994407	7.294017
3	mean2(x, idxs)	2.210280	2.207015	2.214621	2.180346	2.161826	3.296947
5	mean2(x[idxs])	2.407321	2.416889	2.395683	2.398660	2.384810	1.980464

Figure: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on double+n = 1000 data. Outliers are displayed as crosses. Times are in milliseconds.

n = 10000 vector

> x <- data[["n = 10000"]]
> idxs <- sample.int(length(x), size = length(x) * 0.7)
> x_S <- x[idxs]
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5344057 285.5    7916910 422.9  7916910 422.9
Vcells 21447233 163.7   39539356 301.7 94934136 724.3
> stats <- microbenchmark(mean2_x_S = mean2(x_S, refine = TRUE), mean2_x_S_no_refine = mean2(x_S, refine = FALSE), 
+     `mean2(x, idxs)` = mean2(x, idxs = idxs, refine = TRUE), `mean2_no_refine(x, idxs)` = mean2(x, 
+         idxs = idxs, refine = FALSE), `mean2(x[idxs])` = mean2(x[idxs], refine = TRUE), `mean2_no_refine(x[idxs])` = mean2(x[idxs], 
+         refine = FALSE), unit = "ms")

Table: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on double+n = 10000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	0.011529	0.0124345	0.0131386	0.0134575	0.0137425	0.016491
1	mean2_x_S	0.021683	0.0232830	0.0247345	0.0246010	0.0255325	0.039620
4	mean2_no_refine(x, idxs)	0.024507	0.0265645	0.0285094	0.0278905	0.0290310	0.063720
6	mean2_no_refine(x[idxs])	0.030129	0.0328290	0.0350074	0.0347835	0.0355420	0.048748
3	mean2(x, idxs)	0.037147	0.0400865	0.0426263	0.0434240	0.0437385	0.067930
5	mean2(x[idxs])	0.040544	0.0441535	0.0467372	0.0469495	0.0478245	0.077817

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	mean2_x_S	1.880735	1.872452	1.882578	1.828051	1.857922	2.402523
4	mean2_no_refine(x, idxs)	2.125683	2.136354	2.169896	2.072487	2.112498	3.863926
6	mean2_no_refine(x[idxs])	2.613323	2.640154	2.664471	2.584693	2.586283	2.956037
3	mean2(x, idxs)	3.222049	3.223813	3.244352	3.226751	3.182718	4.119217
5	mean2(x[idxs])	3.516697	3.550887	3.557239	3.488724	3.480044	4.718756

Figure: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on double+n = 10000 data. Outliers are displayed as crosses. Times are in milliseconds.

n = 100000 vector

> x <- data[["n = 100000"]]
> idxs <- sample.int(length(x), size = length(x) * 0.7)
> x_S <- x[idxs]
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5344156 285.5    7916910 422.9  7916910 422.9
Vcells 21542339 164.4   39539356 301.7 94934136 724.3
> stats <- microbenchmark(mean2_x_S = mean2(x_S, refine = TRUE), mean2_x_S_no_refine = mean2(x_S, refine = FALSE), 
+     `mean2(x, idxs)` = mean2(x, idxs = idxs, refine = TRUE), `mean2_no_refine(x, idxs)` = mean2(x, 
+         idxs = idxs, refine = FALSE), `mean2(x[idxs])` = mean2(x[idxs], refine = TRUE), `mean2_no_refine(x[idxs])` = mean2(x[idxs], 
+         refine = FALSE), unit = "ms")

Table: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on double+n = 100000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	0.073473	0.076329	0.0895030	0.0841105	0.0935915	0.147389
1	mean2_x_S	0.145473	0.149827	0.1702611	0.1551085	0.1838625	0.520770
4	mean2_no_refine(x, idxs)	0.188426	0.193833	0.2180048	0.2058290	0.2375080	0.315463
6	mean2_no_refine(x[idxs])	0.252566	0.266935	0.3164315	0.2964040	0.3314100	0.582847
3	mean2(x, idxs)	0.288981	0.296944	0.3267770	0.3055625	0.3333595	0.540610
5	mean2(x[idxs])	0.322798	0.337867	0.3950561	0.3620625	0.4264285	0.829984

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	mean2_x_S	1.979952	1.962911	1.902294	1.844104	1.964521	3.533303
4	mean2_no_refine(x, idxs)	2.564561	2.539441	2.435725	2.447126	2.537709	2.140343
6	mean2_no_refine(x[idxs])	3.437535	3.497164	3.535428	3.523983	3.541027	3.954481
3	mean2(x, idxs)	3.933159	3.890317	3.651016	3.632870	3.561857	3.667913
5	mean2(x[idxs])	4.393423	4.426456	4.413886	4.304605	4.556274	5.631248

Figure: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on double+n = 100000 data. Outliers are displayed as crosses. Times are in milliseconds.

n = 1000000 vector

> x <- data[["n = 1000000"]]
> idxs <- sample.int(length(x), size = length(x) * 0.7)
> x_S <- x[idxs]
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5344255 285.5    7916910 422.9  7916910 422.9
Vcells 22487406 171.6   39539356 301.7 94934136 724.3
> stats <- microbenchmark(mean2_x_S = mean2(x_S, refine = TRUE), mean2_x_S_no_refine = mean2(x_S, refine = FALSE), 
+     `mean2(x, idxs)` = mean2(x, idxs = idxs, refine = TRUE), `mean2_no_refine(x, idxs)` = mean2(x, 
+         idxs = idxs, refine = FALSE), `mean2(x[idxs])` = mean2(x[idxs], refine = TRUE), `mean2_no_refine(x[idxs])` = mean2(x[idxs], 
+         refine = FALSE), unit = "ms")

Table: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on double+n = 1000000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	1.230885	1.343892	1.439630	1.437428	1.522612	1.797705
1	mean2_x_S	2.489166	2.652147	2.842999	2.835109	2.960653	3.524161
4	mean2_no_refine(x, idxs)	5.964255	6.614252	7.457118	7.203841	7.824255	12.777483
6	mean2_no_refine(x[idxs])	7.221800	8.150077	9.065001	8.576329	9.560261	16.890504
5	mean2(x[idxs])	8.529981	9.520125	10.625877	9.909907	10.849048	28.367464
3	mean2(x, idxs)	10.193729	11.444287	13.008162	12.409895	14.001493	21.960129

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	mean2_x_S	2.022257	1.973483	1.974812	1.972349	1.944457	1.960367
4	mean2_no_refine(x, idxs)	4.845501	4.921716	5.179884	5.011618	5.138705	7.107664
6	mean2_no_refine(x[idxs])	5.867161	6.064535	6.296756	5.966440	6.278856	9.395593
5	mean2(x[idxs])	6.929958	7.083998	7.380976	6.894194	7.125287	15.779822
3	mean2(x, idxs)	8.281626	8.515782	9.035765	8.633403	9.195707	12.215647

Figure: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on double+n = 1000000 data. Outliers are displayed as crosses. Times are in milliseconds.

n = 10000000 vector

> x <- data[["n = 10000000"]]
> idxs <- sample.int(length(x), size = length(x) * 0.7)
> x_S <- x[idxs]
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5344354 285.5    7916910 422.9  7916910 422.9
Vcells 31938122 243.7   47527227 362.7 94934136 724.3
> stats <- microbenchmark(mean2_x_S = mean2(x_S, refine = TRUE), mean2_x_S_no_refine = mean2(x_S, refine = FALSE), 
+     `mean2(x, idxs)` = mean2(x, idxs = idxs, refine = TRUE), `mean2_no_refine(x, idxs)` = mean2(x, 
+         idxs = idxs, refine = FALSE), `mean2(x[idxs])` = mean2(x[idxs], refine = TRUE), `mean2_no_refine(x[idxs])` = mean2(x[idxs], 
+         refine = FALSE), unit = "ms")

Table: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on double+n = 10000000 data. The top panel shows times in milliseconds and the bottom panel shows relative times.

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	9.945101	13.01269	15.49916	14.25683	18.21093	24.03539
1	mean2_x_S	18.255518	24.04461	27.35695	25.87535	29.05412	40.64428
4	mean2_no_refine(x, idxs)	113.465042	149.09081	163.52154	167.07174	177.39111	211.85577
6	mean2_no_refine(x[idxs])	151.800584	181.34268	194.35322	192.54680	205.41341	254.45399
5	mean2(x[idxs])	146.893860	196.24733	214.24513	211.35711	223.84974	576.89962
3	mean2(x, idxs)	203.682130	300.38533	323.19607	328.22621	350.22217	395.24234

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	1.000000	1.000000	1.00000	1.000000	1.000000	1.000000
1	mean2_x_S	1.835629	1.847782	1.76506	1.814945	1.595423	1.691018
4	mean2_no_refine(x, idxs)	11.409139	11.457341	10.55035	11.718719	9.740917	8.814325
6	mean2_no_refine(x[idxs])	15.263855	13.935835	12.53960	13.505586	11.279680	10.586638
5	mean2(x[idxs])	14.770474	15.081229	13.82302	14.824976	12.292057	24.002088
3	mean2(x, idxs)	20.480650	23.084033	20.85249	23.022390	19.231431	16.444181

Figure: Benchmarking of mean2_x_S(), mean2_x_S_no_refine(), mean2(x, idxs)(), mean2_no_refine(x, idxs)(), mean2(x[idxs])() and mean2_no_refine(x[idxs])() on double+n = 10000000 data. 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_0.15.0-9000    
[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 2.82 mins.

Reproducibility

To reproduce this report, do:

html <- matrixStats:::benchmark('mean2_subset')

---

Copyright Dongcan Jiang. Last updated on 2021-08-25 22:43:50 (+0200 UTC). Powered by RSP.