Spectrum LSF

Showing 503 of 15426

Collect I/O per LSF job as a resource to be reported to RTM

As number of CPUs per machine rises, the I/O to disk rises proportionally, but the bandwidth to the disk generally does not scale. We have found that in high-cpu machines, the system can grind to a near halt due to overload of the disk I/O. We can...

over 7 years ago in Spectrum LSF 2 Not under consideration

1 VOTE

GPU Usage for charge back

Chargeback report in LSF explorer reports usage based on slots that are tied to CPU cores. Deep learning experiments use CPUs and GPUs, we need chargeback to include GPU usage so that the target users/departments can be charged appropriately.

almost 4 years ago in Spectrum LSF / Reporting/Explorer 1 Not under consideration

1 VOTE

energy usage reporting (EAS) not working on intel x86_64 servers

During the testing of EAS on my Dell/HPE hardwares running Linux/CentOS - I figured out that various features in EAS is not really working on these hardwares and they may be only designed for IBM Hardware like Power Systems. Worked with Support on...

almost 4 years ago in Spectrum LSF 1 Not under consideration

1 VOTE

host power management 'EAS' feature for x86_64 h/w

almost 4 years ago in Spectrum LSF 1 Not under consideration

1 VOTE

Horizontal Scroll On/Off Default User Cfg

Horizontal Scroll On/Off Default User CfgREF PMR -- TS004718748 New option to enable/disable Horizontal scroll bar Currently defaults to OFF -- No Scroll barThis differs from our current installed usageInstalling FP11 would BREAK all users and for...

almost 4 years ago in Spectrum LSF / Operations/RTM 1 Not under consideration

1 VOTE

lsf_current_status index contains very old entries

have 2,146,965,961 documents in the lsf_current_status index, which is all the documents one shard can contain. When I query the data I see records in this index that date back 6 months. I believe their is some issue that requires fixing to as ent...

almost 4 years ago in Spectrum LSF / Reporting/Explorer 1 Not under consideration

1 VOTE

badmin should not call bhosts, bqueues command

During testing, noted that badmin hclose/qclose, hopen/qopen commands also issues bhosts and bqueues command respectively. bhosts/bqueues query commands should be for end users and LSF binaries should not call this commands like bhosts etc and pul...

almost 4 years ago in Spectrum LSF 1 Not under consideration

2 VOTE

RTM Database - Selective Data storage based up ClusterID

In RTM I have several clusters. They are identified by ClusterID numbers. There is a ClusterID which is Test or LAB cluster. I use that cluster to test LSF configurations. I run several 1000 jobs. As a result in current configuration the Job start...

almost 8 years ago in Spectrum LSF 1 Not under consideration

2 VOTE

Preempt real memory usage reported by a preemptable job

Memory preemption only works with the reserved memory (rusage[mem=X]) by the preemptable job. However, if the preemptable job uses more memory than the reserved, it may happen that a LSF thinks a preemptive job does not fit on the node because of ...

almost 8 years ago in Spectrum LSF 1 Not under consideration

2 VOTE

Alternate Job Collection parameters needed in mobile App.

In the current version of mobile app, all jobs are collected by job groups. Searching by job group or by the already existing search of jobID is unrealistic when you are dealing with jobs in the millions. Xilinx would like to see the addition of m...

about 8 years ago in Spectrum LSF 1 Not under consideration

Please enter your email address

FILTER BY CATEGORY