Multicore and/or as.POSIXct seems to have problems with MacOS Lion.
The following program (from the pvec help) runs fine:
dates <- sprintf('%04d-%02d-%02d', as.integer(2000+rnorm(1e5)),
as.integer(runif(1e5,1,12)), as.integer(runif(1e5,1,28)))
system.time(a <- as.POSIXct(dates))
but
system.time(b <- pvec(dates, as.POSIXct))
blocks all cpu cores and does not terminate on both my Apple machines. I
tested that on MacOS Lion with multicore 1.5 and 1.6 under R 2.13.1 and R
2.12.2. On snow leopard everything ran smooth.
Is this a problem with R libraries or with the OS? Are there any known
workarounds?
Thank you very much,
Sebastian
Multicore and as.POSIXct on MacOS Lion
2 messages · Sebastian Schmidt, Simon Urbanek
1 day later
On Aug 9, 2011, at 12:34 PM, Sebastian Schmidt wrote:
Multicore and/or as.POSIXct seems to have problems with MacOS Lion.
The following program (from the pvec help) runs fine:
dates <- sprintf('%04d-%02d-%02d', as.integer(2000+rnorm(1e5)),
as.integer(runif(1e5,1,12)), as.integer(runif(1e5,1,28)))
system.time(a <- as.POSIXct(dates))
but
system.time(b <- pvec(dates, as.POSIXct))
blocks all cpu cores and does not terminate on both my Apple machines. I
tested that on MacOS Lion with multicore 1.5 and 1.6 under R 2.13.1 and R
2.12.2. On snow leopard everything ran smooth.
Is this a problem with R libraries or with the OS? Are there any known
workarounds?
I had a report from someone (ah, Sebastian, now I see it was you :)) that there are certain system calls that block on Lion - and I think they were date/time related. It seems to be a bug in Lion. I did have time to have a closer look yet (mainly because I avoid Lion until it gets more stable so I can use it on my regular machines), but I doubt there is much we can do - more likely Apple needs to fix it. Cheers, Simon