List of Contributions

YI SUN

Contact Details

My Content

1 to 20 of 50+ total

RE: Can't display Error while submitting the LSF jobs

Posted By YI SUN Mon April 15, 2024 12:06 PM
Found In Egroup: High Performance Computing Group \ view thread
You may check this link see how LSF supports X forwarding job. There are some other job submission (bsub) options to work with SSH tunneling. ------------------------------ YI SUN ------------------------------

RE: Spectrum LSF Community Edition -- cannot download anything on the download page

Posted By YI SUN Tue April 09, 2024 12:52 PM
Found In Egroup: High Performance Computing Group \ view thread
Now it is working on browsers in addition to Chrome. ------------------------------ YI SUN ------------------------------

RE: Spectrum LSF Community Edition -- cannot download anything on the download page

Posted By YI SUN Wed March 27, 2024 05:35 PM
Found In Egroup: High Performance Computing Group \ view thread
Yes.Seems Chrome works, but filefox and Edge fail. ------------------------------ YI SUN ------------------------------

RE: Spectrum LSF Community Edition -- cannot download anything on the download page

Posted By YI SUN Wed March 20, 2024 01:22 PM
Found In Egroup: High Performance Computing Group \ view thread
Hi Mitch, could you share the link you use to access LSF Community Edition? ------------------------------ YI SUN ------------------------------

RE: Ibm spectrum lsf community edition (lsfsce10.2.0.12)

Posted By YI SUN Tue March 12, 2024 12:13 PM
Found In Egroup: High Performance Computing Group \ view thread
Maybe try following to skip authentication Shut down LSF daemons on both hosts set LSF_AUTH=none set LSF_STRICT_CHECKING=N set LSF_AUTH_QUERY_COMMANDS=N Start up LSF daemons on both hosts ------------------------------ YI SUN ------------------------------

RE: Ibm spectrum lsf community edition (lsfsce10.2.0.12)

Posted By YI SUN Mon March 11, 2024 08:45 PM
Found In Egroup: High Performance Computing Group \ view thread
Do you have the same problem on host57? It seems lsfadmin account authentication fails. Are you sure lsfadmin account setting on host58 and host57 are same? ------------------------------ YI SUN ------------------------------

RE: Ibm spectrum lsf community edition (lsfsce10.2.0.12)

Posted By YI SUN Sun March 10, 2024 09:37 PM
Found In Egroup: High Performance Computing Group \ view thread
It seems you enabled the EGO during installation. Let's try following on host58. kill lim process as root run ". profile.lsf" env | grep EGO unset all EGO_* environment varaibles listed in 3) set LSF_ENABLE_EGO=N and comment out LSF_EGO_ENVDIR in lsf.conf run lsadmin limstartup ---------- ...

RE: Ibm spectrum lsf community edition (lsfsce10.2.0.12)

Posted By YI SUN Thu March 07, 2024 08:55 PM
Found In Egroup: High Performance Computing Group \ view thread
log on host58 as root and make sure lim process is not running. 1) source LSF_TOP/conf/profile.lsf 2) env | egrep "LSF | EGO" to list related parameters 3) lsadmin ckconfig -v see if reports any error 4) if no error in 3), run lsadmin limstartup 5) if there is error in 3), append updated lsf.conf ...

RE: Ibm spectrum lsf community edition (lsfsce10.2.0.12)

Posted By YI SUN Wed March 06, 2024 02:06 PM
Found In Egroup: High Performance Computing Group \ view thread
Could you disable LSF_EGO_DAEMON_CONTROL=Y and add LSF_SERVERDIR, LSF_LIBDIR, LSF_BINDIR into lsf.conf file, then stop/start LSF service. ------------------------------ YI SUN ------------------------------

RE: Accessing multiple GPUs on different hosts using LSF

Posted By YI SUN Mon March 04, 2024 04:24 PM
Found In Egroup: High Performance Computing Group \ view thread
You may check this link. https://community.ibm.com/community/user/cloud/discussion/how-to-run-pytorch-ddp-job-on-multi-nodes#bmcfa6563e-098c-4ad0-88f7-a0615a97de40 ------------------------------ YI SUN ------------------------------

RE: Accessing multiple GPUs on different hosts using LSF

Posted By YI SUN Mon March 04, 2024 04:19 PM
Found In Egroup: High Performance Computing Group \ view thread
How did you launch the 2nd task on host1? ------------------------------ YI SUN ------------------------------

RE: Accessing multiple GPUs on different hosts using LSF

Posted By YI SUN Sun March 03, 2024 03:59 PM
Found In Egroup: High Performance Computing Group \ view thread
Seems you should submit test job to queue "gq" rather "cq" ------------------------------ YI SUN ------------------------------

RE: about 10.1.0.14 version

Posted By YI SUN Fri March 01, 2024 03:09 PM
Found In Egroup: High Performance Computing Group \ view thread
LSF 10.1.0.14 adds lsfd-lim, lsfd-res, and lsfd-sbatchd service unit files to help auto restart lim/res/sbatchd if they exits unexpectedly. But this implementation seems not compatible with LSF daemon management through lsadmin/badmin/bctrld commands. You may consider to install following patch. The ...

RE: Ibm spectrum lsf community edition (lsfsce10.2.0.12)

Posted By YI SUN Fri March 01, 2024 03:03 PM
Found In Egroup: High Performance Computing Group \ view thread
After sourcing LSF profile, run env | grep LSF to check if LSF_SERVERDIR, LSF_BINDIR, and LSF_LIBDIR environment variables are set. The error usually indicates those three variables are not set. You also can manually add them in lsf.conf file. ------------------------------ YI SUN --------------- ...

RE: Ibm spectrum lsf community edition (lsfsce10.2.0.12)

Posted By YI SUN Thu February 29, 2024 02:03 PM
Found In Egroup: High Performance Computing Group \ view thread
It seems LSF environment issue. Before run lsid, you can try "source $LSF_TOP/conf/cshrc.lsf" or ". $LSF_TOP/conf/profile.lsf) to set up LSF environment in the shell session. Here LSF_TOP is the top directory of your LSF installation. ------------------------------ YI SUN ------------------------- ...

RE: Ibm spectrum lsf community edition (lsfsce10.2.0.12)

Posted By YI SUN Wed February 28, 2024 01:06 PM
Found In Egroup: High Performance Computing Group \ view thread
You can use chmod u+s on $LSF_BINDIR/bctrld (owned by root account) as workaround. ------------------------------ YI SUN ------------------------------

RE: Accessing multiple GPUs on different hosts using LSF

Posted By YI SUN Thu February 22, 2024 11:53 AM
Found In Egroup: High Performance Computing Group \ view thread
Try following see if one node job can get GPU allocation on host2. On host1, bsub -I -gpu "num=4" -m host2 nvidia-smi If above test is positive, try following see if it works on two nodes. On host1, bsub -n 2 -gpu "num=4/host" -R "type==any span[ptile=1]" blaunch nvdia-smi ...

RE: Ibm spectrum lsf community edition (lsfsce10.2.0.12)

Posted By YI SUN Fri February 16, 2024 11:08 AM
Found In Egroup: High Performance Computing Group \ view thread
It doesn't seem you have resolved host name resolution issue as previously mentioned on host58. As same user I guess no problem for you to submit job and run hosts. ------------------------------ YI SUN ------------------------------

RE: Ibm spectrum lsf community edition (lsfsce10.2.0.12)

Posted By YI SUN Fri February 16, 2024 11:01 AM
Found In Egroup: High Performance Computing Group \ view thread
On host57 use root account to stop and start LSF services. On host58 you need to source LSF profile before running any LSF command ------------------------------ YI SUN ------------------------------

RE: Ibm spectrum lsf community edition (lsfsce10.2.0.12)

Posted By YI SUN Fri February 16, 2024 10:57 AM
Found In Egroup: High Performance Computing Group \ view thread
You can do both on the compute node. ------------------------------ YI SUN ------------------------------