mean2_subset

matrixStats: Benchmark report

---

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  3237341 172.9    5709258  305  5709258 305.0
Vcells 12940607  98.8   28038728  214 87357391 666.5
> 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
1	mean2_x_S	0.001391	0.0014275	0.0014746	0.0014445	0.0014750	0.001891
2	mean2_x_S_no_refine	0.001392	0.0014325	0.0014915	0.0014470	0.0015025	0.002195
4	mean2_no_refine(x, idxs)	0.002036	0.0020680	0.0021376	0.0020810	0.0021125	0.004911
3	mean2(x, idxs)	0.002022	0.0020710	0.0021228	0.0020830	0.0021230	0.002821
5	mean2(x[idxs])	0.002792	0.0029155	0.0030071	0.0029735	0.0030495	0.003842
6	mean2_no_refine(x[idxs])	0.002820	0.0029085	0.0039671	0.0029995	0.0031200	0.096489

	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	1.000719	1.003503	1.011509	1.001731	1.018644	1.160762
4	mean2_no_refine(x, idxs)	1.463695	1.448686	1.449683	1.440637	1.432203	2.597039
3	mean2(x, idxs)	1.453631	1.450788	1.439660	1.442022	1.439322	1.491803
5	mean2(x[idxs])	2.007189	2.042382	2.039327	2.058498	2.067458	2.031729
6	mean2_no_refine(x[idxs])	2.027318	2.037478	2.690380	2.076497	2.115254	51.025383

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  3233777 172.8    5709258  305  5709258 305.0
Vcells 11808939  90.1   28038728  214 87357391 666.5
> 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	0.007993	0.0080635	0.0082027	0.0081170	0.0081970	0.011061
2	mean2_x_S_no_refine	0.007994	0.0080785	0.0083935	0.0081425	0.0082375	0.031185
4	mean2_no_refine(x, idxs)	0.013556	0.0136325	0.0137183	0.0136720	0.0137450	0.014743
3	mean2(x, idxs)	0.013537	0.0136295	0.0137370	0.0136985	0.0137950	0.014099
5	mean2(x[idxs])	0.019365	0.0197740	0.0202089	0.0199010	0.0201385	0.033319
6	mean2_no_refine(x[idxs])	0.019468	0.0198135	0.0203902	0.0199815	0.0201645	0.032497

	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	1.000125	1.001860	1.023257	1.003142	1.004941	2.819365
4	mean2_no_refine(x, idxs)	1.695984	1.690643	1.672411	1.684366	1.676833	1.332881
3	mean2(x, idxs)	1.693607	1.690271	1.674683	1.687631	1.682933	1.274659
5	mean2(x[idxs])	2.422745	2.452285	2.463677	2.451768	2.456813	3.012296
6	mean2_no_refine(x[idxs])	2.435631	2.457184	2.485788	2.461685	2.459985	2.937980

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  3233876 172.8    5709258  305  5709258 305.0
Vcells 11872517  90.6   28038728  214 87357391 666.5
> 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
1	mean2_x_S	0.073192	0.0733580	0.0735297	0.0734010	0.0735615	0.076538
2	mean2_x_S_no_refine	0.073240	0.0733670	0.0739529	0.0734355	0.0735930	0.100184
4	mean2_no_refine(x, idxs)	0.139412	0.1395635	0.1418755	0.1396525	0.1398425	0.216691
3	mean2(x, idxs)	0.139339	0.1395675	0.1402904	0.1396880	0.1398400	0.163529
5	mean2(x[idxs])	0.222900	0.2238220	0.2259060	0.2243030	0.2255700	0.293821
6	mean2_no_refine(x[idxs])	0.223118	0.2237985	0.2253905	0.2243285	0.2250305	0.259972

	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	1.000656	1.000123	1.005756	1.000470	1.000428	1.308945
4	mean2_no_refine(x, idxs)	1.904744	1.902499	1.929499	1.902597	1.901028	2.831156
3	mean2(x, idxs)	1.903746	1.902553	1.907943	1.903080	1.900994	2.136573
5	mean2(x[idxs])	3.045415	3.051092	3.072310	3.055858	3.066414	3.838891
6	mean2_no_refine(x[idxs])	3.048393	3.050772	3.065300	3.056205	3.059080	3.396640

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  3233975 172.8    5709258  305  5709258 305.0
Vcells 12502795  95.4   28038728  214 87357391 666.5
> 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
2	mean2_x_S_no_refine	0.819987	0.864055	0.8904314	0.872912	0.9101030	1.059168
1	mean2_x_S	0.746681	0.864348	0.8836420	0.874443	0.9106205	1.020441
3	mean2(x, idxs)	2.070680	2.352730	2.4511763	2.445139	2.5226930	2.946063
4	mean2_no_refine(x, idxs)	2.045235	2.341083	2.4554177	2.446915	2.5772595	2.792526
5	mean2(x[idxs])	3.233744	4.660319	4.9976479	4.812597	5.0010920	15.858845
6	mean2_no_refine(x[idxs])	3.281051	4.671983	4.9904019	4.823245	4.9698770	16.243384

	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	0.910601	1.000339	0.9923752	1.001754	1.000569	0.9634364
3	mean2(x, idxs)	2.525260	2.722893	2.7527964	2.801129	2.771876	2.7814879
4	mean2_no_refine(x, idxs)	2.494229	2.709415	2.7575596	2.803164	2.831833	2.6365279
5	mean2(x[idxs])	3.943653	5.393544	5.6126142	5.513267	5.495083	14.9729269
6	mean2_no_refine(x[idxs])	4.001345	5.407044	5.6044766	5.525465	5.460785	15.3359845

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  3234071 172.8    5709258 305.0  5709258 305.0
Vcells 18803453 143.5   33726473 257.4 87357391 666.5
> 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
2	mean2_x_S_no_refine	9.053797	12.12507	13.27875	12.24600	16.13622	18.12965
1	mean2_x_S	9.069925	11.65445	13.40851	12.38970	16.42170	18.62974
4	mean2_no_refine(x, idxs)	80.103785	95.49647	99.35035	97.53584	104.74655	117.62011
3	mean2(x, idxs)	79.685783	96.42031	100.16765	99.56728	105.24282	111.31561
5	mean2(x[idxs])	128.063078	135.78485	146.60189	141.03012	145.42353	418.95981
6	mean2_no_refine(x[idxs])	114.357685	136.14016	140.93776	141.52260	145.64412	165.26707

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	1.000000	1.0000000	1.000000	1.000000	1.000000	1.000000
1	mean2_x_S	1.001781	0.9611864	1.009772	1.011735	1.017692	1.027584
4	mean2_no_refine(x, idxs)	8.847535	7.8759513	7.481908	7.964710	6.491391	6.487721
3	mean2(x, idxs)	8.801366	7.9521435	7.543457	8.130596	6.522146	6.139976
5	mean2(x[idxs])	14.144682	11.1986848	11.040342	11.516423	9.012240	23.109094
6	mean2_no_refine(x[idxs])	12.630909	11.2279881	10.613786	11.556639	9.025911	9.115844

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  3234179 172.8    5709258 305.0  5709258 305.0
Vcells 17360460 132.5   40551767 309.4 87357391 666.5
> 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	0.001412	0.0014460	0.0014958	0.0014610	0.0014805	0.002581
4	mean2_no_refine(x, idxs)	0.002062	0.0020965	0.0021521	0.0021165	0.0021555	0.002529
1	mean2_x_S	0.002132	0.0021630	0.0022130	0.0021830	0.0022170	0.002508
3	mean2(x, idxs)	0.002782	0.0028230	0.0028867	0.0028415	0.0029060	0.004155
6	mean2_no_refine(x[idxs])	0.002972	0.0031115	0.0034228	0.0031920	0.0033410	0.017197
5	mean2(x[idxs])	0.003691	0.0038675	0.0040252	0.0039535	0.0041580	0.005126

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
4	mean2_no_refine(x, idxs)	1.460340	1.449862	1.438747	1.448665	1.455927	0.9798528
1	mean2_x_S	1.509915	1.495851	1.479460	1.494182	1.497467	0.9717164
3	mean2(x, idxs)	1.970255	1.952282	1.929779	1.944901	1.962850	1.6098411
6	mean2_no_refine(x[idxs])	2.104816	2.151798	2.288177	2.184805	2.256670	6.6629213
5	mean2(x[idxs])	2.614023	2.674620	2.690898	2.706023	2.808511	1.9860519

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  3234272 172.8    5709258 305.0  5709258 305.0
Vcells 17370413 132.6   40551767 309.4 87357391 666.5
> 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	0.008046	0.0082290	0.0084524	0.0082940	0.0083945	0.021646
4	mean2_no_refine(x, idxs)	0.013675	0.0138290	0.0139197	0.0139055	0.0139785	0.014878
1	mean2_x_S	0.015307	0.0154260	0.0155253	0.0155170	0.0156065	0.015898
3	mean2(x, idxs)	0.020930	0.0210695	0.0212282	0.0211535	0.0212655	0.024405
6	mean2_no_refine(x[idxs])	0.021730	0.0222470	0.0230560	0.0224820	0.0227340	0.046212
5	mean2(x[idxs])	0.029109	0.0295505	0.0299383	0.0297495	0.0299955	0.035199

	expr	min	lq	mean	median	uq	max
2	mean2_x_S_no_refine	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
4	mean2_no_refine(x, idxs)	1.699602	1.680520	1.646840	1.676573	1.665198	0.6873325
1	mean2_x_S	1.902436	1.874590	1.836793	1.870871	1.859134	0.7344544
3	mean2(x, idxs)	2.601293	2.560396	2.511502	2.550458	2.533266	1.1274600
6	mean2_no_refine(x[idxs])	2.700721	2.703488	2.727750	2.710634	2.708202	2.1348979
5	mean2(x[idxs])	3.617822	3.591020	3.541990	3.586870	3.573233	1.6261203

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  3234371 172.8    5709258 305.0  5709258 305.0
Vcells 17465522 133.3   40551767 309.4 87357391 666.5
> 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	0.073383	0.0735780	0.0748593	0.0739205	0.0747670	0.112598
1	mean2_x_S	0.145355	0.1455485	0.1467679	0.1458990	0.1471985	0.159998
4	mean2_no_refine(x, idxs)	0.151656	0.1520600	0.1542243	0.1522575	0.1524755	0.310611
3	mean2(x, idxs)	0.251454	0.2519835	0.2533889	0.2521780	0.2524295	0.288298
6	mean2_no_refine(x[idxs])	0.257278	0.2661055	0.3567621	0.3921090	0.3957230	0.410325
5	mean2(x[idxs])	0.331040	0.4081495	0.4351360	0.4647425	0.4678355	0.514656

	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	1.980772	1.978152	1.960583	1.973728	1.968763	1.420967
4	mean2_no_refine(x, idxs)	2.066637	2.066650	2.060188	2.059747	2.039342	2.758584
3	mean2(x, idxs)	3.426597	3.424713	3.384868	3.411476	3.376215	2.560418
6	mean2_no_refine(x[idxs])	3.505962	3.616645	4.765767	5.304469	5.292750	3.644159
5	mean2(x[idxs])	4.511127	5.547168	5.812717	6.287058	6.257246	4.570738

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  3234470 172.8    5709258 305.0  5709258 305.0
Vcells 18410589 140.5   40551767 309.4 87357391 666.5
> 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	1.109013	1.205344	1.248485	1.247311	1.268617	1.450124
1	mean2_x_S	2.270884	2.405847	2.495035	2.488111	2.528594	2.798773
4	mean2_no_refine(x, idxs)	5.103129	5.342005	5.462168	5.455947	5.544610	6.269207
6	mean2_no_refine(x[idxs])	6.055889	7.278324	8.882099	9.561957	9.753134	18.643234
3	mean2(x, idxs)	8.936119	9.448269	9.695674	9.636117	9.842493	11.401684
5	mean2(x[idxs])	7.863381	8.747281	10.445387	10.761168	11.028589	20.290996

	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	2.047662	1.995984	1.998450	1.994780	1.993190	1.930023
4	mean2_no_refine(x, idxs)	4.601505	4.431934	4.375036	4.374168	4.370596	4.323221
6	mean2_no_refine(x[idxs])	5.460611	6.038379	7.114300	7.666057	7.688008	12.856303
3	mean2(x, idxs)	8.057723	7.838649	7.765950	7.725513	7.758447	7.862558
5	mean2(x[idxs])	7.090432	7.257083	8.366448	8.627494	8.693399	13.992594

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  3234569 172.8    5709258 305.0  5709258 305.0
Vcells 27861305 212.6   48742120 371.9 87357391 666.5
> 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	7.416486	10.97654	13.64744	12.08673	18.71371	21.16901
1	mean2_x_S	15.439379	20.59652	25.01414	23.11340	32.18062	36.78187
4	mean2_no_refine(x, idxs)	95.059996	132.90154	143.80882	140.24072	160.59192	189.81906
6	mean2_no_refine(x[idxs])	159.109667	167.80084	180.08796	173.82238	186.44614	452.33853
5	mean2(x[idxs])	169.306052	179.77445	199.31009	193.45681	198.79157	471.73459
3	mean2(x, idxs)	230.929040	297.02068	310.27159	309.52590	325.38551	386.11923

	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	2.081765	1.876413	1.832882	1.912295	1.719628	1.737534
4	mean2_no_refine(x, idxs)	12.817390	12.107783	10.537422	11.602868	8.581514	8.966837
6	mean2_no_refine(x[idxs])	21.453511	15.287228	13.195733	14.381259	9.963080	21.367959
5	mean2(x[idxs])	22.828338	16.378064	14.604212	16.005720	10.622780	22.284206
3	mean2(x, idxs)	31.137258	27.059596	22.734786	25.608739	17.387551	18.239834

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 3.6.1 Patched (2019-08-27 r77078)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.3 LTS

Matrix products: default
BLAS:   /home/hb/software/R-devel/R-3-6-branch/lib/R/lib/libRblas.so
LAPACK: /home/hb/software/R-devel/R-3-6-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_1.4-6    matrixStats_0.55.0-9000 ggplot2_3.2.1          
[4] knitr_1.24              R.devices_2.16.0        R.utils_2.9.0          
[7] R.oo_1.22.0             R.methodsS3_1.7.1       history_0.0.0-9002     

loaded via a namespace (and not attached):
 [1] Biobase_2.45.0       bit64_0.9-7          splines_3.6.1       
 [4] network_1.15         assertthat_0.2.1     highr_0.8           
 [7] stats4_3.6.1         blob_1.2.0           robustbase_0.93-5   
[10] pillar_1.4.2         RSQLite_2.1.2        backports_1.1.4     
[13] lattice_0.20-38      glue_1.3.1           digest_0.6.20       
[16] colorspace_1.4-1     sandwich_2.5-1       Matrix_1.2-17       
[19] XML_3.98-1.20        lpSolve_5.6.13.3     pkgconfig_2.0.2     
[22] genefilter_1.66.0    purrr_0.3.2          ergm_3.10.4         
[25] xtable_1.8-4         mvtnorm_1.0-11       scales_1.0.0        
[28] tibble_2.1.3         annotate_1.62.0      IRanges_2.18.2      
[31] TH.data_1.0-10       withr_2.1.2          BiocGenerics_0.30.0 
[34] lazyeval_0.2.2       mime_0.7             survival_2.44-1.1   
[37] magrittr_1.5         crayon_1.3.4         statnet.common_4.3.0
[40] memoise_1.1.0        laeken_0.5.0         R.cache_0.13.0      
[43] MASS_7.3-51.4        R.rsp_0.43.1         tools_3.6.1         
[46] multcomp_1.4-10      S4Vectors_0.22.1     trust_0.1-7         
[49] munsell_0.5.0        AnnotationDbi_1.46.1 compiler_3.6.1      
[52] rlang_0.4.0          grid_3.6.1           RCurl_1.95-4.12     
[55] cwhmisc_6.6          rappdirs_0.3.1       labeling_0.3        
[58] bitops_1.0-6         base64enc_0.1-3      boot_1.3-23         
[61] gtable_0.3.0         codetools_0.2-16     DBI_1.0.0           
[64] markdown_1.1         R6_2.4.0             zoo_1.8-6           
[67] dplyr_0.8.3          bit_1.1-14           zeallot_0.1.0       
[70] parallel_3.6.1       Rcpp_1.0.2           vctrs_0.2.0         
[73] DEoptimR_1.0-8       tidyselect_0.2.5     xfun_0.9            
[76] coda_0.19-3         

Total processing time was 2.85 mins.

Reproducibility

To reproduce this report, do:

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

---

Copyright Dongcan Jiang. Last updated on 2019-09-10 21:06:50 (-0700 UTC). Powered by RSP.

<script> var link = document.createElement('link'); link.rel = 'icon'; link.href = "data:image/png;base64,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" document.getElementsByTagName('head')[0].appendChild(link); </script>