Jobs submitted to the general farm of batch systems at SLAC will be scheduled to run according to a cross-queue user-based fairshare priority system. There is more information here www.slac.stanford.edu/comp/unix/package/lsf/currdoc/lsf_admin/index.htm about this kind of scheduling. The priority of a user's job is dynamically determined based on various factors such as the number of jobs the user has across all the queues in the fairshare, their resource consumption, the number of shares belonging to the user's fairshare group, etc. The general farm fairshare queue structure can be viewed with the command /usr/local/bin/bqueues -l short. This is an example of what the relevant part of the output from that command looks like at the time of this writing:
renata@rhel6-64g $ bqueues -l short
QUEUE: short
-- This is the default queue.
.
SCHEDULING POLICIES: FAIRSHARE BACKFILL
FAIRSHARE_QUEUES: short medium long atlasq atl-analq
USER_SHARES: [atlasgrp, 37875] [babarAll, 7859] [cdmsALL, 2000] [exoprodgrp, 1500] [exousergrp, 550] [geantgrp, 3874] [glastdata, 854] [glastgrp, 366] [glastusers, 25320] [hpsprodgrp, 1000] [lcd, 600] [lcdprodgrp, 1100] [luxlz, 3500] [rpgrp, 500] [theorygrp, 4257] [AllUsers, 16086]
SHARE_INFO_FOR: short/
| USER/GROUP | SHARES | PRIORITY | STARTED | RESERVED | CPU_TIME | RUN_TIME | ADJUST |
|---|---|---|---|---|---|---|---|
| lcdprodgrp | 1100 | 278.667 | 0 | 0 | 0 | 0 | 0.000 |
| exoprodgrp | 1500 | 231.000 | 0 | 0 | 0 | 0 | 0.000 |
| geantgrp | 3874 | 220.469 | 0 | 0 | 0 | 0 | 0.000 |
| luxlz | 3500 | 212.363 | 0 | 0 | 0 | 0 | 0.000 |
| hpsprodgrp | 1000 | 200.000 | 0 | 0 | 0 | 0 | 0.000 |
| lcd | 600 | 139.333 | 0 | 0 | 0 | 0 | 0.000 |
| rdgrp | 342 | 114.000 | 0 | 0 | 0 | 0 | 0.000 |
| glastdata | 854 | 13.352 | 0 | 0 | 481047.4 | 0 | 0.000 |
| glastgrp | 366 | 13.086 | 0 | 0 | 143741.5 | 0 | 0.000 |
| babarAll | 7859 | 2.497 | 7 | 0 | 54117976.0 | 3353425 | 0.000 |
| glastusers | 25320 | 1.540 | 1822 | 0 | 86034584.0 | 35653321 | 0.000 |
| atlasgrp | 37875 | 0.532 | 4414 | 0 | 263242448.0 | 106437853 | 0.000 |
| rpgrp | 500 | 0.439 | 62 | 0 | 3924898.2 | 6360861 | 0.000 |
| cdmsALL | 2000 | 0.418 | 472 | 0 | 5018088.5 | 19465549 | 0.000 |
| exousergrp | 550 | 0.414 | 36 | 0 | 4399336.0 | 287801 | 0.000 |
| AllUsers | 16086 | 0.414 | 2104 | 0 | 252849040.0 | 19525638 | 0.000 |
theorygrp | 4257 | 0.413 | 91 | 0 | 11024314.0 | 498370 | 0.000 |
USERS: all
HOSTS: genfells+1 hequfarm+1 dolefarm+1 kisofarm+1 bulletfarm/
Jobs submitted by a user that is not a member of any usergoup will get the priority of the AllUsers group. Jobs submitted by a user that is a member of a fairshare group may also get the priority of AllUsers if the priority of AllUsers happens to be higher than that of his own group at the time his job is submitted.