matrixStats.benchmarks

---

anyMissing() benchmarks on subsetted computation

This report benchmark the performance of anyMissing() 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  5162916 275.8    7916910 422.9  7916910 422.9
Vcells 25710602 196.2   53955392 411.7 50098692 382.3
> stats <- microbenchmark(anyMissing_x_S = anyMissing(x_S), `anyMissing(x, idxs)` = anyMissing(x, idxs = idxs), 
+     `anyMissing(x[idxs])` = anyMissing(x[idxs]), unit = "ms")

Table: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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	anyMissing_x_S	0.001299	0.0013310	0.0014001	0.0013850	0.0014175	0.002203
2	anyMissing(x, idxs)	0.002861	0.0029085	0.0029626	0.0029380	0.0029755	0.004156
3	anyMissing(x[idxs])	0.003853	0.0039675	0.0052384	0.0040435	0.0041210	0.121189

	expr	min	lq	mean	median	uq	max
1	anyMissing_x_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	anyMissing(x, idxs)	2.202463	2.185199	2.116021	2.121300	2.099118	1.886518
3	anyMissing(x[idxs])	2.966128	2.980841	3.741459	2.919495	2.907231	55.010894

Figure: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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  5148112  275    7916910 422.9  7916910 422.9
Vcells 14536728  111   53955392 411.7 50098692 382.3
> stats <- microbenchmark(anyMissing_x_S = anyMissing(x_S), `anyMissing(x, idxs)` = anyMissing(x, idxs = idxs), 
+     `anyMissing(x[idxs])` = anyMissing(x[idxs]), unit = "ms")

Table: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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	anyMissing_x_S	0.006821	0.006938	0.0071182	0.0070145	0.0071065	0.012310
2	anyMissing(x, idxs)	0.021872	0.021980	0.0226106	0.0220715	0.0222115	0.037837
3	anyMissing(x[idxs])	0.028426	0.028889	0.0302136	0.0292280	0.0297295	0.067615

	expr	min	lq	mean	median	uq	max
1	anyMissing_x_S	1.000000	1.00000	1.000000	1.000000	1.000000	1.000000
2	anyMissing(x, idxs)	3.206568	3.16806	3.176452	3.146554	3.125519	3.073680
3	anyMissing(x[idxs])	4.167424	4.16388	4.244565	4.166797	4.183424	5.492689

Figure: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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  5148184 275.0    7916910 422.9  7916910 422.9
Vcells 14600288 111.4   53955392 411.7 50098692 382.3
> stats <- microbenchmark(anyMissing_x_S = anyMissing(x_S), `anyMissing(x, idxs)` = anyMissing(x, idxs = idxs), 
+     `anyMissing(x[idxs])` = anyMissing(x[idxs]), unit = "ms")

Table: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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	anyMissing_x_S	0.045269	0.047030	0.0565744	0.055161	0.0605690	0.097646
2	anyMissing(x, idxs)	0.189848	0.196721	0.2237074	0.210139	0.2433385	0.385652
3	anyMissing(x[idxs])	0.222029	0.238814	0.2838194	0.268992	0.2986360	0.765679

	expr	min	lq	mean	median	uq	max
1	anyMissing_x_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	anyMissing(x, idxs)	4.193775	4.182883	3.954212	3.809557	4.017542	3.949491
3	anyMissing(x[idxs])	4.904659	5.077908	5.016743	4.876489	4.930509	7.841376

Figure: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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  5148256 275.0    7916910 422.9  7916910 422.9
Vcells 15230337 116.2   53955392 411.7 50098692 382.3
> stats <- microbenchmark(anyMissing_x_S = anyMissing(x_S), `anyMissing(x, idxs)` = anyMissing(x, idxs = idxs), 
+     `anyMissing(x[idxs])` = anyMissing(x[idxs]), unit = "ms")

Table: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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	anyMissing_x_S	0.454448	0.5216525	0.5786677	0.5436835	0.5899715	1.079747
2	anyMissing(x, idxs)	2.487519	2.7614475	3.4705498	3.3085560	3.8648375	7.685493
3	anyMissing(x[idxs])	3.827613	4.4859705	5.4221502	4.9909670	5.9684330	10.399041

	expr	min	lq	mean	median	uq	max
1	anyMissing_x_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	anyMissing(x, idxs)	5.473715	5.293653	5.997483	6.085445	6.550889	7.117865
3	anyMissing(x[idxs])	8.422554	8.599538	9.370058	9.179913	10.116477	9.630998

Figure: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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  5148328 275.0    7916910 422.9  7916910 422.9
Vcells 21530590 164.3   53955392 411.7 50658010 386.5
> stats <- microbenchmark(anyMissing_x_S = anyMissing(x_S), `anyMissing(x, idxs)` = anyMissing(x, idxs = idxs), 
+     `anyMissing(x[idxs])` = anyMissing(x[idxs]), unit = "ms")

Table: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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	anyMissing_x_S	6.390179	7.663246	8.921722	8.455134	10.36377	13.00473
2	anyMissing(x, idxs)	87.748499	111.039786	115.822160	115.685753	122.27734	141.40928
3	anyMissing(x[idxs])	119.607946	135.674981	140.726406	139.743561	144.71224	168.76377

	expr	min	lq	mean	median	uq	max
1	anyMissing_x_S	1.00000	1.00000	1.00000	1.00000	1.00000	1.00000
2	anyMissing(x, idxs)	13.73177	14.48992	12.98204	13.68231	11.79853	10.87368
3	anyMissing(x[idxs])	18.71746	17.70464	15.77346	16.52766	13.96328	12.97711

Figure: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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  5148409 275.0    7916910 422.9  7916910 422.9
Vcells 20087457 153.3   53955392 411.7 53032641 404.7
> stats <- microbenchmark(anyMissing_x_S = anyMissing(x_S), `anyMissing(x, idxs)` = anyMissing(x, idxs = idxs), 
+     `anyMissing(x[idxs])` = anyMissing(x[idxs]), unit = "ms")

Table: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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
1	anyMissing_x_S	0.001356	0.0014020	0.0014909	0.0014710	0.0015125	0.002507
2	anyMissing(x, idxs)	0.003008	0.0030575	0.0032146	0.0030975	0.0031555	0.009546
3	anyMissing(x[idxs])	0.003724	0.0039190	0.0047963	0.0039910	0.0041330	0.049398

	expr	min	lq	mean	median	uq	max
1	anyMissing_x_S	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000
2	anyMissing(x, idxs)	2.218289	2.180813	2.156126	2.10571	2.086281	3.807738
3	anyMissing(x[idxs])	2.746313	2.795292	3.217042	2.71312	2.732562	19.704029

Figure: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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  5148472 275.0    7916910 422.9  7916910 422.9
Vcells 20097246 153.4   53955392 411.7 53032641 404.7
> stats <- microbenchmark(anyMissing_x_S = anyMissing(x_S), `anyMissing(x, idxs)` = anyMissing(x, idxs = idxs), 
+     `anyMissing(x[idxs])` = anyMissing(x[idxs]), unit = "ms")

Table: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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
1	anyMissing_x_S	0.006850	0.0072490	0.0094397	0.0075730	0.0132905	0.016769
2	anyMissing(x, idxs)	0.023116	0.0235495	0.0286511	0.0238020	0.0282000	0.049185
3	anyMissing(x[idxs])	0.026394	0.0276825	0.0337686	0.0282005	0.0330330	0.065500

	expr	min	lq	mean	median	uq	max
1	anyMissing_x_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	anyMissing(x, idxs)	3.374598	3.248655	3.035167	3.143008	2.121816	2.933091
3	anyMissing(x[idxs])	3.853139	3.818803	3.577295	3.723822	2.485460	3.906017

Figure: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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  5148544 275.0    7916910 422.9  7916910 422.9
Vcells 20192161 154.1   53955392 411.7 53032641 404.7
> stats <- microbenchmark(anyMissing_x_S = anyMissing(x_S), `anyMissing(x, idxs)` = anyMissing(x, idxs = idxs), 
+     `anyMissing(x[idxs])` = anyMissing(x[idxs]), unit = "ms")

Table: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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
1	anyMissing_x_S	0.040801	0.0458760	0.0535020	0.0495030	0.0560180	0.115687
2	anyMissing(x, idxs)	0.177440	0.1994325	0.2273533	0.2125495	0.2316585	0.549579
3	anyMissing(x[idxs])	0.263107	0.2803375	0.3378248	0.2959000	0.3937720	1.195971

	expr	min	lq	mean	median	uq	max
1	anyMissing_x_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	anyMissing(x, idxs)	4.348913	4.347208	4.249433	4.293669	4.135430	4.750568
3	anyMissing(x[idxs])	6.448543	6.110766	6.314242	5.977416	7.029383	10.337990

Figure: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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  5148616 275.0    7916910 422.9  7916910 422.9
Vcells 21137210 161.3   53955392 411.7 53032641 404.7
> stats <- microbenchmark(anyMissing_x_S = anyMissing(x_S), `anyMissing(x, idxs)` = anyMissing(x, idxs = idxs), 
+     `anyMissing(x[idxs])` = anyMissing(x[idxs]), unit = "ms")

Table: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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
1	anyMissing_x_S	0.635010	0.746746	0.7843915	0.762187	0.786418	1.741441
2	anyMissing(x, idxs)	5.124992	5.398494	6.3265125	5.801729	6.654898	11.951174
3	anyMissing(x[idxs])	6.766017	9.684119	10.3313381	10.067975	10.829647	22.629365

	expr	min	lq	mean	median	uq	max
1	anyMissing_x_S	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000
2	anyMissing(x, idxs)	8.070726	7.229358	8.065504	7.61195	8.462292	6.862807
3	anyMissing(x[idxs])	10.654977	12.968425	13.171150	13.20932	13.770853	12.994620

Figure: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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  5148688 275.0    7916910 422.9  7916910 422.9
Vcells 30587693 233.4   53955392 411.7 53339345 407.0
> stats <- microbenchmark(anyMissing_x_S = anyMissing(x_S), `anyMissing(x, idxs)` = anyMissing(x, idxs = idxs), 
+     `anyMissing(x[idxs])` = anyMissing(x[idxs]), unit = "ms")

Table: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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
1	anyMissing_x_S	6.503693	8.108895	10.21231	9.51436	12.53715	16.24656
2	anyMissing(x, idxs)	96.152599	132.862656	148.87533	149.98104	165.32895	194.07407
3	anyMissing(x[idxs])	138.857610	172.913414	180.78766	181.79698	187.64958	222.23205

	expr	min	lq	mean	median	uq	max
1	anyMissing_x_S	1.00000	1.00000	1.00000	1.00000	1.00000	1.00000
2	anyMissing(x, idxs)	14.78431	16.38480	14.57803	15.76365	13.18712	11.94555
3	anyMissing(x[idxs])	21.35058	21.32392	17.70292	19.10764	14.96748	13.67872

Figure: Benchmarking of anyMissing_x_S(), anyMissing(x, idxs)() and anyMissing(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         
[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 1.26 mins.

Reproducibility

To reproduce this report, do:

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

---

Copyright Dongcan Jiang. Last updated on 2021-08-25 22:08:46 (+0200 UTC). Powered by RSP.