Am 23.03.2011 um 12:31 schrieb Mark Dixon: > On Wed, 23 Mar 2011, Reuti wrote: > ... >> Yes, for the SHARETREE, but not the functional policy. Even the parameter >> ACCT_RESERVED_USAGE doesn't say anything about functional policy, only for >> the accounting. > > Ah, apologies - I was conflating "fair share" with "share tree". > >>> It's been an indispensible option for us. I'm not sure why the manual page >>> has it marked as deprecated, when it's so useful. >> >> It's unified with ACCT_RESERVED_USAGE. Most of the time you will set both. > ... > > "unified" is a bit strong isn't it? They are totally independent options. > > We used to set both, but no longer as: > > 1) Sometime between 6.0 and 6.2, ACCT_RESERVED_USAGE also started making > maxvmem show the reserved usage in the accounting file.
True - was this behavior discussed on any list before, whether it's a feature or a bug? When "cpu" and "mem" is already the reserved integral value, you can compute this "maxvmem" easily affterwards. There it no need to set it IMO. > 2) It's interesting to have the actual cpu time in the accounting file, so > that you can see if cpu time / run time ~= 1.0 Yes, as long as the jobs are not reprioritized with different nice values and/or the exechosts are oversubscribed, i.e. you have slots=cores. You would charge for the run time then (in case you do it)? -- Reuti > Best wishes, > > Mark > -- > ----------------------------------------------------------------- > Mark Dixon Email : [email protected] > HPC/Grid Systems Support Tel (int): 35429 > Information Systems Services Tel (ext): +44(0)113 343 5429 > University of Leeds, LS2 9JT, UK > ----------------------------------------------------------------- _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
