On Fri, 2 Nov 2007, Graeme Stewart wrote: > Hi Stephen > > Because of the confusion in the way people have used the information > system to date, we publish SMPSize=1 (actually it's 4 (well, 2x2 > cores)), and GlueHostMainMemoryRAMSize=2G (actually it's 8). >... > For the moment it seems to me that what Glasgow are doing ensures > that the right answer is given in both the correct and incorrect JDL > expressions, so I'm not going to change that. This is the "traditional" config Heinz and Jeremy mentioned yesterday. Since _LCG_ were never interested in SMP processing, there wasn't a penalty in making the site look like a larger number of smaller single-CPU boxes. Of course other VOs now might want such resources, but a bigger risk is that with pilot jobs frameworks, the pilot may look at the node it's running on and request work packets suitable for the physical resources it sees without taking into account that these will have to be shared with others! At least one part of biomed seems to be using boinc as a pilot job framework for example (though they may have hacked their client of course). Thanks Henry > On 2 Nov 2007, at 09:52, Burke, S (Stephen) wrote: > > > Hi all, > > > > With reference to the recent discussion, this is what sites are > > currently publishing for the SMPSize, which should be the number of > > CPUs/cores per WN. Can people check that it looks right? > > > > Stephen > > > > > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: lcgce01.gridpp.rl.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: 1000-lcgce01.gridpp.rl.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: 2000-lcgce01.gridpp.rl.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: 700-lcgce01.gridpp.rl.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: 1000-lcgce02.gridpp.rl.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: 700-lcgce02.gridpp.rl.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: lcgce02.gridpp.rl.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: 2000-lcgce02.gridpp.rl.ac.uk > > -- > > GlueHostArchitectureSMPSize: 1 > > GlueSubClusterName: ce.epcc.ed.ac.uk > > -- > > GlueHostArchitectureSMPSize: 4 > > GlueSubClusterName: dgc-grid-44.brunel.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: dgc-grid-35.brunel.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: dgc-grid-40.brunel.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: gw39.hep.ph.ic.ac.uk > > -- > > GlueHostArchitectureSMPSize: 4 > > GlueSubClusterName: ce00.hep.ph.ic.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: hep-ce.cx1.hpc.ic.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: mars-ce2.mars.lesc.doc.ic.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: viking-ce0.viking.lesc.doc.ic.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: ce02.esc.qmul.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: ce1.pp.rhul.ac.uk > > -- > > GlueHostArchitectureSMPSize: 4 > > GlueSubClusterName: gw-2.ccc.ucl.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: pc90.hep.ucl.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: fal-pygrid-18.lancs.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: hepgrid2.ph.liv.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: ce01.tier2.hep.manchester.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: ce02.tier2.hep.manchester.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: lcgce0.shef.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: helmsley.dur.scotgrid.ac.uk > > -- > > GlueHostArchitectureSMPSize: 4 > > GlueSubClusterName: ce.glite.ecdf.ed.ac.uk > > -- > > GlueHostArchitectureSMPSize: 1 > > GlueSubClusterName: svr016.gla.scotgrid.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: epgce1.ph.bham.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: epgce2.ph.bham.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: lcgce01.phy.bris.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: serv03.hep.phy.cam.ac.uk > > -- > > GlueHostArchitectureSMPSize: 8 > > GlueSubClusterName: t2ce03.physics.ox.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: t2ce02.physics.ox.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: heplnx206.pp.rl.ac.uk > > -- > > GlueHostArchitectureSMPSize: 2 > > GlueSubClusterName: heplnx207.pp.rl.ac.uk > > -- > Dr Graeme Stewart - http://wiki.gridpp.ac.uk/wiki/User:Graeme_stewart > ScotGrid - http://www.scotgrid.ac.uk/ http://scotgrid.blogspot.com/ > Dr. Henry Nebrensky [log in to unmask] http://people.brunel.ac.uk/~eesrjjn "The opossum is a very sophisticated animal. It doesn't even get up until 5 or 6 p.m."