Segway reporting impossible CPU times?

[EricR86]

Original report (BitBucket issue) by Rachel Chan (Bitbucket: rcwchan).

---

I was checking through the CPU times in the jobs.tab for one of my runs and noticed some seemingly impossible CPU times:

For example,

#!python

  18737 847359  emt4.1.1472.K562_5_Track.traindir.88a56392194e11e692fcbc305bf153a8  gmtkEMtrain 25  20001   691978240.0000  56.1500 0
  18738 847358  emt5.1.1477.K562_5_Track.traindir.88a56392194e11e692fcbc305bf153a8  gmtkEMtrain 25  20001   1432014848.0000 13.8339 0
  18739 846361  emt3.1.1268.K562_5_Track.traindir.88a56392194e11e692fcbc305bf153a8  gmtkEMtrain 25  20001   3244494848.0000 62077214.6800   0
  18740 846366  emt1.1.1289.K562_5_Track.traindir.88a56392194e11e692fcbc305bf153a8  gmtkEMtrain 25  20001   1727369216.0000 20.6849 0
  18741 848536  emt6.1.1328.K562_5_Track.traindir.88a56392194e11e692fcbc305bf153a8  gmtkEMtrain 25  20001   470630400.0000  16.0756 0
  18742 845030  emt0.1.1603.K562_5_Track.traindir.88a56392194e11e692fcbc305bf153a8  gmtkEMtrain 25  20001   1070817280.0000 13.6149 0
  18743 844380  emt2.1.1652.K562_5_Track.traindir.88a56392194e11e692fcbc305bf153a8  gmtkEMtrain 25  20001   743768064.0000  228.7500    0
  18744 850559  emt9.1.1473.K562_5_Track.traindir.88a56392194e11e692fcbc305bf153a8  gmtkEMtrain 25  20001   529735680.0000  22.2336 0

In particular, 62077214.6800 seconds is 718 days which is impossible. There are a number of other entries like this--in particular, it occurred for every round in this run.

I believe Segway reports the CPU usage directly from SGE (which is in seconds) so this might be an SGE bug when significant CPU resources are used, rather than a Segway bug.

The following SGE thread seems related but I'm unsure if it's exactly the problem we're experiencing: link

[EricR86]

Original comment by Rachel Chan (Bitbucket: rcwchan).

---

• Edited issue description

[EricR86]

Original comment by Rachel Chan (Bitbucket: rcwchan).

---

• Edited issue description