[torqueusers] TORQUE 2.1.6 and 2.0.0p11 are released
Andrew J Caird
acaird at umich.edu
Tue Oct 31 07:26:11 MST 2006
On Mon, 30 Oct 2006, Garrick Staples wrote:
> On Mon, Oct 30, 2006 at 01:37:52PM -0500, Andrew J Caird alleged:
>> Garrick - if I recall correctly you have a G5 cluster - is
>> this working OK for you?
> (This is said while wearing my USC hat, not my CRI hat)
> It was turned off quite some time ago and noone really noticed.
> So we just left it off.
> Turns out it was pretty useless for USC's users. We didn't
> have any software for OSX that we didn't have on Linux, and
> everything ran faster on the Linux Intel/AMD boxes anyways.
> We might have seen a performance improvement if we had the
> fancy compilers, but noone wanted to buy them.
> Of course, none of this matters an iota anymore now that Macs
> are Intel.
So much for another simple test case. We certainly understand
your decision to retire the G5s, but we're trying to squeeze out
the last few cycles from them for the next couple years. A lot
of the code here is home grown, so the usage is actually pretty
good. It's likely that we'll continue to be a good test site for
OS X/G5 Torque and Maui builds for a while.
To the topic at hand, though: is there a simple way to see if a
job has a TM interface open with the MOM?
All of that said, after a few jobs have gone through the
system, this seems to be affecting only one set of them, which
suggests to me that they may not be acting as good TM citizens.
More information about the torqueusers