R package to send function calls as jobs on LSF, SGE, Slurm, PBS/Torque, or each via SSH
NA
(#325) @luwidmerenv()
now visibly lists environment if called without argumentsBiocParallel
did not export required objects (#302) @bhayete-empresssend()
now reports a call identifier that current()
tracksmax_calls_worker
was not respected (#322) @bhayete-empressmulticore
, multiprocess
) again uses local IP (#321) @wlandauinfo()
now also returns current worker and number of calls-DZMQ_BUILD_DRAFT_API=1
) is now optional (#317)Rf_error
lead to improper cleanup of resources (#311)-Werror
flag led to compilation issues on M1 MacCMQMaster
now needs to know add_pending_workers(n)
info()
now displays properlyn_jobs
and max_calls_worker
provide insufficient call slots (#258)gc()
may be in different column (#240)job_id
to qdel
in PBS (#265)purrr
packageRcpp
in preparation for v0.9
(#151)multiprocess
backend via callr
instead of forking (#142, #197)multicore
, multiprocess
schedulers now support logging (#169)clustermq.host
can specify host IP or network interface name (#170)127.0.0.1
instead of localhost
(#192)foreach
now have their environment stripped (#200)log_worker=T/F
argument is rescindedclustermq.ssh.timeout
for SSH proxy startup (#157) @brendanfclustermq.worker.timeout
for delay before worker shutdown (#188)clustermq.data.warning
(#189)Q
, Q_rows
have new arguments verbose
(#111) and pkgs
(#144)foreach
backend now uses its dedicated API where possible (#143, #144)infuser
packageQ
now has max_calls_worker
argument to avoid walltime (#110)cores
per job field (#123)foreach
now treats .export
(#124) and .combine
(#126) correctlyclustermq.error.timeout
to not wait for clean shutdown (#134)@keywords internal