Cardio is the HPC facility at the CEU.
scontrol show part # cpu-o-1 for interactive jobs with no limit on cpus per user expecting fair usage # cpu-o-[2-5] for long jobs with maximum cpus per user set to 40 and a maximum wall time of 7 days # cpu-o-7 for short jobs with a maximum wall time of 24 hours # Interactive jobs #SBATCH -A CARDIO-SL0-CPU #SBATCH -p cardio_intr #SBATCH --qos=cardio_intr # Long jobs #SBATCH -A CARDIO-SL0-CPU #SBATCH -p cardio #SBATCH --qos=cardio # Short jobs #SBATCH -A CARDIO-SL0-CPU #SBATCH -p cardio_short #SBATCH --qos=cardio_short
Automatic login can be enabled with
It is possible to fetch
scp cardio-login.hpc.cam.ac.uk:/home/$USER/my-file-on-Cardio .
sftp, noting the
-r option for both could be very useful. More generally, it is preferable to use
Old version on Cardio, rsync 3.0.6, gives errors,
rsync -av --partial mydir/ login-cpu.hpc.cam.ac.uk:/rds/user/$USER/hpc-work/mydir Errors after successfully sending a number of files : rsync: writefd_unbuffered failed to write 4 bytes to socket [sender]: Broken pipe (32) rsync: connection unexpectedly closed (604 bytes received so far) [sender] rsync error: error in rsync protocol data stream (code 12) at io.c(600) [sender=3.0.6]
Bram Prins (firstname.lastname@example.org) has the latest version (3.1.3) that doesn't give this error here:
Note /scratch at Cardio is now /DO-NOT-MODIFY-SCRATCH – an example is jp.sh.
R package reinstallations
Package redeployment is illustrated with R below for building R package list from /home/$USER/R at Cardio to be resintalled to /rds/user/$USER/hpc-work/R at CSD3.
We can use screen copy of package list from Cardio since users do not have write permission.
home <- Sys.getenv("HOME") from <- paste0(home,"/R") pkgs <- unname(installed.packages(lib.loc = from)[, "Package"]) edit(pkgs)
pkgs <- "mark and copy the list as given in c() above with Shift+Ins" user <- Sys.getenv("USER") to <- paste0("/rds/user/",user,"/hpc-work/R") install.packages(pkgs, lib=to, repos="https://cran.r-project.org")
If a package list is already built one can reinstall them as follows.
user <- Sys.getenv("USER") location <- paste0("/rds/user/",user,"/hpc-work/R") pkgs <- unname(installed.packages(lib.loc = location)[, "Package"]) install.packages(pkgs, lib=location, repos="https://cran.r-project.org")
A version by Scott Ritchie (email@example.com), reinstall_r_pkgs.R, also touches upon Bioconductor, whose package installations and updates are described at https://bioconductor.org/install/.