matrixStats.benchmarks

colLogSumExps() and rowLogSumExps() benchmarks on subsetted computation

This report benchmark the performance of colLogSumExps() and rowLogSumExps() 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  5224956 279.1    8529671 455.6  8529671 455.6
Vcells 10071502  76.9   31876688 243.2 60562128 462.1
> colStats <- microbenchmark(colLogSumExps_X_S = colLogSumExps(X_S, na.rm = FALSE), `colLogSumExps(X, rows, cols)` = colLogSumExps(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colLogSumExps(X[rows, cols])` = colLogSumExps(X[rows, 
+     cols], na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
           used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells  5216600 278.6    8529671 455.6  8529671 455.6
Vcells 10044274  76.7   31876688 243.2 60562128 462.1
> rowStats <- microbenchmark(rowLogSumExps_X_S = rowLogSumExps(X_S, na.rm = FALSE), `rowLogSumExps(X, cols, rows)` = rowLogSumExps(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowLogSumExps(X[cols, rows])` = rowLogSumExps(X[cols, 
+     rows], na.rm = FALSE), unit = "ms")

Table: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(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
1	colLogSumExps_X_S	0.003153	0.0032055	0.0042499	0.003255	0.003335	0.097414
2	colLogSumExps(X, rows, cols)	0.003507	0.0036035	0.0037238	0.003647	0.003753	0.006518
3	colLogSumExps(X[rows, cols])	0.004060	0.0042820	0.0045311	0.004370	0.004523	0.013022

	expr	min	lq	mean	median	uq	max
1	colLogSumExps_X_S	1.000000	1.000000	1.000000	1.00000	1.000000	1.0000000
2	colLogSumExps(X, rows, cols)	1.112274	1.124162	0.876208	1.12043	1.125337	0.0669103
3	colLogSumExps(X[rows, cols])	1.287662	1.335829	1.066158	1.34255	1.356222	0.1336769

Table: Benchmarking of rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(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
1	rowLogSumExps_X_S	0.003180	0.0032400	0.0033692	0.0032845	0.0033800	0.006432
2	rowLogSumExps(X, cols, rows)	0.003584	0.0036505	0.0047052	0.0037270	0.0038125	0.096998
3	rowLogSumExps(X[cols, rows])	0.004095	0.0043450	0.0044783	0.0044225	0.0045425	0.006256

	expr	min	lq	mean	median	uq	max
1	rowLogSumExps_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowLogSumExps(X, cols, rows)	1.127044	1.126697	1.396514	1.134724	1.127959	15.0805348
3	rowLogSumExps(X[cols, rows])	1.287736	1.341049	1.329196	1.346476	1.343935	0.9726368

Figure: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(X[rows, cols])() on 10x10 data as well as rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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
1	colLogSumExps_X_S	3.153	3.2055	4.24995	3.2550	3.335	97.414
2	rowLogSumExps_X_S	3.180	3.2400	3.36921	3.2845	3.380	6.432

	expr	min	lq	mean	median	uq	max
1	colLogSumExps_X_S	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
2	rowLogSumExps_X_S	1.008563	1.010763	0.7927646	1.009063	1.013493	0.0660275

Figure: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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 5215243 278.6    8529671 455.6  8529671 455.6
Vcells 9714948  74.2   31876688 243.2 60562128 462.1
> colStats <- microbenchmark(colLogSumExps_X_S = colLogSumExps(X_S, na.rm = FALSE), `colLogSumExps(X, rows, cols)` = colLogSumExps(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colLogSumExps(X[rows, cols])` = colLogSumExps(X[rows, 
+     cols], na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5215219 278.6    8529671 455.6  8529671 455.6
Vcells 9725001  74.2   31876688 243.2 60562128 462.1
> rowStats <- microbenchmark(rowLogSumExps_X_S = rowLogSumExps(X_S, na.rm = FALSE), `rowLogSumExps(X, cols, rows)` = rowLogSumExps(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowLogSumExps(X[cols, rows])` = rowLogSumExps(X[cols, 
+     rows], na.rm = FALSE), unit = "ms")

Table: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(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	colLogSumExps_X_S	0.090567	0.0967505	0.1061635	0.1031355	0.1137425	0.138596
2	colLogSumExps(X, rows, cols)	0.093520	0.0998940	0.1083768	0.1058630	0.1177440	0.137605
3	colLogSumExps(X[rows, cols])	0.102416	0.1090815	0.1198994	0.1162185	0.1291190	0.177860

	expr	min	lq	mean	median	uq	max
1	colLogSumExps_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colLogSumExps(X, rows, cols)	1.032606	1.032491	1.020849	1.026446	1.035180	0.9928497
3	colLogSumExps(X[rows, cols])	1.130831	1.127452	1.129385	1.126853	1.135187	1.2832982

Table: Benchmarking of rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(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	rowLogSumExps_X_S	0.092120	0.098513	0.1101811	0.1080375	0.1198965	0.140728
2	rowLogSumExps(X, cols, rows)	0.093270	0.105276	0.1114590	0.1091665	0.1173100	0.163011
3	rowLogSumExps(X[cols, rows])	0.103799	0.112263	0.1219352	0.1211630	0.1300495	0.157524

	expr	min	lq	mean	median	uq	max
1	rowLogSumExps_X_S	1.000000	1.000000	1.000000	1.00000	1.0000000	1.000000
2	rowLogSumExps(X, cols, rows)	1.012484	1.068651	1.011598	1.01045	0.9784272	1.158341
3	rowLogSumExps(X[cols, rows])	1.126780	1.139576	1.106679	1.12149	1.0846814	1.119351

Figure: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(X[rows, cols])() on 100x100 data as well as rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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	colLogSumExps_X_S	90.567	96.7505	106.1635	103.1355	113.7425	138.596
2	rowLogSumExps_X_S	92.120	98.5130	110.1811	108.0375	119.8965	140.728

	expr	min	lq	mean	median	uq	max
1	colLogSumExps_X_S	1.000000	1.000000	1.000000	1.00000	1.000000	1.000000
2	rowLogSumExps_X_S	1.017148	1.018217	1.037844	1.04753	1.054105	1.015383

Figure: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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 5215984 278.6    8529671 455.6  8529671 455.6
Vcells 9718996  74.2   31876688 243.2 60562128 462.1
> colStats <- microbenchmark(colLogSumExps_X_S = colLogSumExps(X_S, na.rm = FALSE), `colLogSumExps(X, rows, cols)` = colLogSumExps(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colLogSumExps(X[rows, cols])` = colLogSumExps(X[rows, 
+     cols], na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5215960 278.6    8529671 455.6  8529671 455.6
Vcells 9729049  74.3   31876688 243.2 60562128 462.1
> rowStats <- microbenchmark(rowLogSumExps_X_S = rowLogSumExps(X_S, na.rm = FALSE), `rowLogSumExps(X, cols, rows)` = rowLogSumExps(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowLogSumExps(X[cols, rows])` = rowLogSumExps(X[cols, 
+     rows], na.rm = FALSE), unit = "ms")

Table: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(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	colLogSumExps_X_S	0.089180	0.0951845	0.1021562	0.0983575	0.1092525	0.125407
2	colLogSumExps(X, rows, cols)	0.090603	0.0967450	0.1049659	0.1027125	0.1104635	0.133387
3	colLogSumExps(X[rows, cols])	0.098990	0.1050460	0.1150448	0.1118800	0.1242225	0.171096

	expr	min	lq	mean	median	uq	max
1	colLogSumExps_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colLogSumExps(X, rows, cols)	1.015956	1.016395	1.027504	1.044277	1.011084	1.063633
3	colLogSumExps(X[rows, cols])	1.110002	1.103604	1.126166	1.137483	1.137022	1.364326

Table: Benchmarking of rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(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	rowLogSumExps_X_S	0.090638	0.1002485	0.1075956	0.104025	0.1145445	0.184519
2	rowLogSumExps(X, cols, rows)	0.091644	0.0975770	0.1099123	0.105153	0.1159690	0.200519
3	rowLogSumExps(X[cols, rows])	0.102591	0.1124205	0.1244612	0.119713	0.1337920	0.216149

	expr	min	lq	mean	median	uq	max
1	rowLogSumExps_X_S	1.000000	1.0000000	1.000000	1.000000	1.000000	1.000000
2	rowLogSumExps(X, cols, rows)	1.011099	0.9733512	1.021532	1.010844	1.012436	1.086712
3	rowLogSumExps(X[cols, rows])	1.131876	1.1214183	1.156750	1.150810	1.168035	1.171419

Figure: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(X[rows, cols])() on 1000x10 data as well as rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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	colLogSumExps_X_S	89.180	95.1845	102.1562	98.3575	109.2525	125.407
2	rowLogSumExps_X_S	90.638	100.2485	107.5956	104.0250	114.5445	184.519

	expr	min	lq	mean	median	uq	max
1	colLogSumExps_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowLogSumExps_X_S	1.016349	1.053202	1.053246	1.057621	1.048438	1.471361

Figure: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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 5216188 278.6    8529671 455.6  8529671 455.6
Vcells 9719923  74.2   31876688 243.2 60562128 462.1
> colStats <- microbenchmark(colLogSumExps_X_S = colLogSumExps(X_S, na.rm = FALSE), `colLogSumExps(X, rows, cols)` = colLogSumExps(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colLogSumExps(X[rows, cols])` = colLogSumExps(X[rows, 
+     cols], na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5216164 278.6    8529671 455.6  8529671 455.6
Vcells 9729976  74.3   31876688 243.2 60562128 462.1
> rowStats <- microbenchmark(rowLogSumExps_X_S = rowLogSumExps(X_S, na.rm = FALSE), `rowLogSumExps(X, cols, rows)` = rowLogSumExps(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowLogSumExps(X[cols, rows])` = rowLogSumExps(X[cols, 
+     rows], na.rm = FALSE), unit = "ms")

Table: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(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
1	colLogSumExps_X_S	0.102946	0.1111695	0.1262989	0.1261630	0.137209	0.186332
2	colLogSumExps(X, rows, cols)	0.107643	0.1187910	0.1294139	0.1267515	0.136894	0.193177
3	colLogSumExps(X[rows, cols])	0.116994	0.1310735	0.1416186	0.1394095	0.150752	0.185741

	expr	min	lq	mean	median	uq	max
1	colLogSumExps_X_S	1.000000	1.000000	1.000000	1.000000	1.0000000	1.0000000
2	colLogSumExps(X, rows, cols)	1.045626	1.068558	1.024664	1.004665	0.9977042	1.0367355
3	colLogSumExps(X[rows, cols])	1.136460	1.179042	1.121297	1.104995	1.0987034	0.9968282

Table: Benchmarking of rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(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
1	rowLogSumExps_X_S	0.106905	0.1176905	0.1297395	0.1253305	0.1380135	0.181775
2	rowLogSumExps(X, cols, rows)	0.112772	0.1228200	0.1352408	0.1333800	0.1414230	0.208652
3	rowLogSumExps(X[cols, rows])	0.118833	0.1296295	0.1429067	0.1386620	0.1534275	0.193409

	expr	min	lq	mean	median	uq	max
1	rowLogSumExps_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	rowLogSumExps(X, cols, rows)	1.054881	1.043585	1.042403	1.064226	1.024704	1.147859
3	rowLogSumExps(X[cols, rows])	1.111576	1.101444	1.101490	1.106371	1.111685	1.064002

Figure: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(X[rows, cols])() on 10x1000 data as well as rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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	rowLogSumExps_X_S	106.905	117.6905	129.7395	125.3305	138.0135	181.775
1	colLogSumExps_X_S	102.946	111.1695	126.2989	126.1630	137.2090	186.332

	expr	min	lq	mean	median	uq	max
2	rowLogSumExps_X_S	1.0000000	1.000000	1.0000000	1.000000	1.0000000	1.00000
1	colLogSumExps_X_S	0.9629671	0.944592	0.9734809	1.006642	0.9941709	1.02507

Figure: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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 5216401 278.6    8529671 455.6  8529671 455.6
Vcells 9764656  74.5   31876688 243.2 60562128 462.1
> colStats <- microbenchmark(colLogSumExps_X_S = colLogSumExps(X_S, na.rm = FALSE), `colLogSumExps(X, rows, cols)` = colLogSumExps(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colLogSumExps(X[rows, cols])` = colLogSumExps(X[rows, 
+     cols], na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5216377 278.6    8529671 455.6  8529671 455.6
Vcells 9864709  75.3   31876688 243.2 60562128 462.1
> rowStats <- microbenchmark(rowLogSumExps_X_S = rowLogSumExps(X_S, na.rm = FALSE), `rowLogSumExps(X, cols, rows)` = rowLogSumExps(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowLogSumExps(X[cols, rows])` = rowLogSumExps(X[cols, 
+     rows], na.rm = FALSE), unit = "ms")

Table: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(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	colLogSumExps_X_S	0.781934	0.8967560	0.9067277	0.8979115	0.9010825	1.456129
2	colLogSumExps(X, rows, cols)	0.805189	0.9229225	0.9228147	0.9256165	0.9292315	1.360920
3	colLogSumExps(X[rows, cols])	0.880037	1.0061050	1.0225432	1.0087370	1.0196910	1.341322

	expr	min	lq	mean	median	uq	max
1	colLogSumExps_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
2	colLogSumExps(X, rows, cols)	1.029740	1.029179	1.017742	1.030855	1.031239	0.934615
3	colLogSumExps(X[rows, cols])	1.125462	1.121938	1.127729	1.123426	1.131629	0.921156

Table: Benchmarking of rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(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
1	rowLogSumExps_X_S	0.827333	0.9208120	0.9433978	0.923690	0.9504715	1.451742
2	rowLogSumExps(X, cols, rows)	0.849239	0.9464205	0.9543103	0.948168	0.9749475	1.389180
3	rowLogSumExps(X[cols, rows])	0.929286	1.0340740	1.0589153	1.043933	1.0746860	1.459369

	expr	min	lq	mean	median	uq	max
1	rowLogSumExps_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowLogSumExps(X, cols, rows)	1.026478	1.027811	1.011567	1.026500	1.025751	0.9569056
3	rowLogSumExps(X[cols, rows])	1.123231	1.123002	1.122448	1.130177	1.130687	1.0052537

Figure: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(X[rows, cols])() on 100x1000 data as well as rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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	colLogSumExps_X_S	781.934	896.756	906.7277	897.9115	901.0825	1456.129
2	rowLogSumExps_X_S	827.333	920.812	943.3978	923.6900	950.4715	1451.742

	expr	min	lq	mean	median	uq	max
1	colLogSumExps_X_S	1.00000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowLogSumExps_X_S	1.05806	1.026826	1.040442	1.028709	1.054811	0.9969872

Figure: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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 5216611 278.6    8529671 455.6  8529671 455.6
Vcells 9765483  74.6   31876688 243.2 60562128 462.1
> colStats <- microbenchmark(colLogSumExps_X_S = colLogSumExps(X_S, na.rm = FALSE), `colLogSumExps(X, rows, cols)` = colLogSumExps(X, 
+     rows = rows, cols = cols, na.rm = FALSE), `colLogSumExps(X[rows, cols])` = colLogSumExps(X[rows, 
+     cols], na.rm = FALSE), unit = "ms")
> X <- t(X)
> X_S <- t(X_S)
> gc()
          used  (Mb) gc trigger  (Mb) max used  (Mb)
Ncells 5216587 278.6    8529671 455.6  8529671 455.6
Vcells 9865536  75.3   31876688 243.2 60562128 462.1
> rowStats <- microbenchmark(rowLogSumExps_X_S = rowLogSumExps(X_S, na.rm = FALSE), `rowLogSumExps(X, cols, rows)` = rowLogSumExps(X, 
+     rows = cols, cols = rows, na.rm = FALSE), `rowLogSumExps(X[cols, rows])` = rowLogSumExps(X[cols, 
+     rows], na.rm = FALSE), unit = "ms")

Table: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(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	colLogSumExps_X_S	0.745673	0.8376020	0.8636380	0.8539335	0.858835	1.246552
2	colLogSumExps(X, rows, cols)	0.764129	0.8517975	0.8834998	0.8768770	0.883215	1.180537
3	colLogSumExps(X[rows, cols])	0.839221	0.9145985	0.9506828	0.9604135	0.969277	1.420172

	expr	min	lq	mean	median	uq	max
1	colLogSumExps_X_S	1.000000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	colLogSumExps(X, rows, cols)	1.024751	1.016948	1.022998	1.026868	1.028387	0.9470419
3	colLogSumExps(X[rows, cols])	1.125455	1.091925	1.100788	1.124694	1.128595	1.1392802

Table: Benchmarking of rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(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
2	rowLogSumExps(X, cols, rows)	0.783550	0.8507085	0.8992607	0.875669	0.9011415	1.434717
1	rowLogSumExps_X_S	0.785798	0.8754500	0.8941442	0.885309	0.9071225	1.172982
3	rowLogSumExps(X[cols, rows])	0.887504	0.9777690	1.0092014	1.000114	1.0253060	1.411025

	expr	min	lq	mean	median	uq	max
2	rowLogSumExps(X, cols, rows)	1.000000	1.000000	1.0000000	1.000000	1.000000	1.0000000
1	rowLogSumExps_X_S	1.002869	1.029083	0.9943103	1.011009	1.006637	0.8175703
3	rowLogSumExps(X[cols, rows])	1.132670	1.149358	1.1222569	1.142114	1.137786	0.9834866

Figure: Benchmarking of colLogSumExps_X_S(), colLogSumExps(X, rows, cols)() and colLogSumExps(X[rows, cols])() on 1000x100 data as well as rowLogSumExps_X_S(), rowLogSumExps(X, cols, rows)() and rowLogSumExps(X[cols, rows])() on the same data transposed. Outliers are displayed as crosses. Times are in milliseconds.

Table: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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	colLogSumExps_X_S	745.673	837.602	863.6380	853.9335	858.8350	1246.552
2	rowLogSumExps_X_S	785.798	875.450	894.1441	885.3090	907.1225	1172.982

	expr	min	lq	mean	median	uq	max
1	colLogSumExps_X_S	1.00000	1.000000	1.000000	1.000000	1.000000	1.0000000
2	rowLogSumExps_X_S	1.05381	1.045186	1.035323	1.036742	1.056224	0.9409812

Figure: Benchmarking of colLogSumExps_X_S() and rowLogSumExps_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_0.60.1     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] rappdirs_0.3.3          startup_0.15.0          labeling_0.4.2         
[67] bitops_1.0-7            base64enc_0.1-3         boot_1.3-28            
[70] gtable_0.3.0            DBI_1.1.1               markdown_1.1           
[73] R6_2.5.1                lpSolveAPI_5.5.2.0-17.7 rle_0.9.2              
[76] dplyr_1.0.7             fastmap_1.1.0           bit_4.0.4              
[79] utf8_1.2.2              parallel_4.1.1          Rcpp_1.0.7             
[82] vctrs_0.3.8             png_0.1-7               DEoptimR_1.0-9         
[85] tidyselect_1.1.1        xfun_0.25               coda_0.19-4            

Total processing time was 12.18 secs.

Reproducibility

To reproduce this report, do:

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