[R-sig-hpc] how to remove workers in doSMP

Dirk Eddelbuettel edd at debian.org
Thu Aug 19 15:32:44 CEST 2010


On 19 August 2010 at 09:52, Roman Luštrik wrote:
| Looks like I was sloppy and didn't clean up after myself in previous
| session(s). I'm trying to activate my cores via doSMP package like so:
| 
| > wrk <- startWorkers(4, FORCE = TRUE)  # if I don't use FORCE, it won't
| create the object
| Warning messages:
| 1: In startWorkers(4, FORCE = TRUE) :
|   there is an existing doSMP session using doSMP1
| 2: In startWorkers(4, FORCE = TRUE) :
|   there is an existing doSMP session using doSMP2
| 3: In startWorkers(4, FORCE = TRUE) :
|   there is an existing doSMP session using doSMP3
| 4: In startWorkers(4, FORCE = TRUE) :
|   there is an existing doSMP session using doSMP4
| 5: In startWorkers(4, FORCE = TRUE) :
|   there is an existing doSMP session using doSMP5
| 6: In startWorkers(4, FORCE = TRUE) :
|   there is an existing doSMP session using doSMP6
| 7: In startWorkers(4, FORCE = TRUE) :
|   there is an existing doSMP session using doSMP7
| 8: In startWorkers(4, FORCE = TRUE) :
|   there is an existing doSMP session using doSMP8
| 9: In startWorkers(4, FORCE = TRUE) : unable to recreate task queue doSMP1
| > registerDoSMP(wrk)
| 
| After I registerDoSMP, I get two popup windows that my Rterm has stopped
| responding. If I try to run something in parallel, only one/two cores will
| work. Because I failed to stop workers in my previous sessions (and have
| rebooted the system inbetween), I don't have objects with which I could
| stopWorkers(obj). I tried stopWorkers(wrk) (after using FORCE in
| startWorkers), but R session seems to hang (console unresponsive and not
| much is going with CPU usage. How can I "reboot" cores so that I'll be able
| to register workers?

This is commercially supported software not even released to CRAN so I
recommend you contact the vendor.

Dirk

-- 
Dirk Eddelbuettel | edd at debian.org | http://dirk.eddelbuettel.com



More information about the R-sig-hpc mailing list