[R-sig-hpc] Parallel linear model
Dirk Eddelbuettel
edd at debian.org
Thu Aug 23 15:16:33 CEST 2012
On 23 August 2012 at 12:12, Patrik Waldmann wrote:
| Here's a comparison on Windows based on 8 cores (excluding foreach):
| > y<-rnorm(1000)
| > x<-matrix(rnorm(1000*10000),ncol=10000)
| > dimx<-dim(x)
| > library(rbenchmark)
| > benchmark(pval<-apply(x,2, function(x)summary(lm(y~x))$coeff[2,4]),
| replications=1)
|
| 1 pval <- apply(x, 2, function(x) summary(lm(y ~ x))$coeff[2, 4])
| test replications elapsed relative user.self sys.self user.child sys.child
| 1 1 25.16 1 20.46 2.17 NA NA
|
| > library(parallel)
| > cores<-detectCores()
| > cl <- makeCluster(cores, methods=FALSE)
| > clusterExport(cl,"y")
| > benchmark(pval<-parApply(cl, x,2, function(x)summary(lm(y~x))$coeff[2,4]),
| replications=1)
| 1 pval <- parApply(cl, x, 2, function(x) summary(lm(y ~ x))$coeff[2, 4])
| test replications elapsed relative user.self sys.self user.child sys.child
| 1 1 5.52 1 0.74 0.28 NA NA
|
| > stopCluster(cl)
|
| # More fair
|
| > benchmark({cores<-detectCores()
| + cl <- makeCluster(cores, methods=FALSE)
| + clusterExport(cl,"y")
| + pval<-parApply(cl, x,2, function(x)summary(lm(y~x))$coeff[2,4])},
| replications=1)
| test replications elapsed relative user.self sys.self user.child sys.child
| 1 { 1 7.11 1 0.65 0.37 NA NA
| Warning messages:
| 1: closing unused connection 10 (<-patwa-PC:10187)
| 2: closing unused connection 9 (<-patwa-PC:10187)
| 3: closing unused connection 8 (<-patwa-PC:10187)
| 4: closing unused connection 7 (<-patwa-PC:10187)
| 5: closing unused connection 6 (<-patwa-PC:10187)
| 6: closing unused connection 5 (<-patwa-PC:10187)
| 7: closing unused connection 4 (<-patwa-PC:10187)
| 8: closing unused connection 3 (<-patwa-PC:10187)
| > stopCluster(cl)
|
| What does the warnings refer to?
I prefer doing HPC on Linux --- but on Windows, as I recall, this stems from
the socket connections needed to make things do.
Dirk
| Patrik
|
| >>> Dirk Eddelbuettel <edd at debian.org> 23/08/2012 02:53 >>>
|
| The difference between user and elapsed is an old hat. Here is a great
| example (and IIRC first shown here by Simon) with no compute time:
|
| R> system.time(mclapply(1:8, function(x) Sys.sleep(1))) ## 2 cores by
| default
| user system elapsed
| 0.000 0.012 4.014
| R> system.time(mclapply(1:8, function(x) Sys.sleep(1), mc.cores=8))
| user system elapsed
| 0.012 0.020 1.039
| R>
|
| so elapsed time is effectively the one second a Sys.sleep(1) takes, plus
| overhead, if we allow for all eight (hyperthreaded) cores here. By Brian
| Ripley's choice a default of two is baked-in, so clueless users only get a
| small gain. "user time" is roughly the actual system load _summed over all
| processes / threads_.
|
| With that, could I ask any of the participants in the thread to re-try with a
| proper benchmarking package such as rbenchmark or microbenchmark? Either one
| beats to the socks of system.time:
|
| R> library(rbenchmark)
| R> benchmark( mclapply(1:8, function(x) Sys.sleep(1)), mclapply(1:8,
| function(x) Sys.sleep(1), mc.cores=8), replications=1)
| test replications elapsed
| relative user.self sys.self user.child sys.child
| 1 mclapply(1:8, function(x) Sys.sleep(1)) 1
| 4.013 3.89612 0.000 0.008 0.000 0.004
| 2 mclapply(1:8, function(x) Sys.sleep(1), mc.cores = 8) 1
| 1.030 1.00000 0.004 0.008 0.004 0.000
| R>
|
| and
|
| R> library(microbenchmark)
| R> microbenchmark( mclapply(1:8, function(x) Sys.sleep(1)), mclapply(1:8,
| function(x) Sys.sleep(1), mc.cores=8), times=1)
| Unit: seconds
| expr min lq
| median uq max
| 1 mclapply(1:8, function(x) Sys.sleep(1)) 4.01377 4.01377
| 4.01377 4.01377 4.01377
| 2 mclapply(1:8, function(x) Sys.sleep(1), mc.cores = 8) 1.03457 1.03457
| 1.03457 1.03457 1.03457
| R>
|
| (and you normally want to run either with 10 or 100 or ... replications /
| times).
|
| Dirk
|
| --
| Dirk Eddelbuettel | edd at debian.org | http://dirk.eddelbuettel.com
|
| _______________________________________________
| R-sig-hpc mailing list
| R-sig-hpc at r-project.org
| https://stat.ethz.ch/mailman/listinfo/r-sig-hpc
--
Dirk Eddelbuettel | edd at debian.org | http://dirk.eddelbuettel.com
More information about the R-sig-hpc
mailing list