I think that it would be productive for Cameron to describe what he wants this to do. My guess (based on knowing Cameron and he area of work) is that he doesn't want to do serious computational science (aka heavy number crunching). Rather he wants to explore this area on a mainframe, demo what the mainframe does when including this in the mix and train students on these aspects of a mainframe environment.
--henry schaffer On Fri, Jan 27, 2012 at 3:04 PM, John Campbell <soup...@gmail.com> wrote: > Cameron Seay wrote: >> What would I have to do to get a distro to run on z/VM? I want to run >> Scientific Linux, but there does not appear to be a version for VM. Do I >> need to make modifications in the source code and recompile it? > > The greatest weakness I see to a "Scientific Linux" on a mainframe > (especially under z/VM) is that the workload if heavy on computation-- > the resource most shared-- rather than hammering I/O (which mainframes > have optimized for bandwidth and connectivity). > > Now I've done some heavy number crunching wayyyyy in the past (when a > 120MFLOP array processor was still "hot stuff") and it can hammer a > system flat (all right, so I was working w/ a UNIVAC 1100/80 w/ a > DataWest Array Processor... and I handled microcode and I/O management > to do 3D FFT benchmarks in efforts to sell to oil companies for > exploration) and thinking about sharing CPUs with an instance doing > anything like that kind of workload kinda makes me break out into > hives. > > Remember, a mainframe's CPU is optimized for "best reliable > single-thread performance" (think about balancing B-Trees, for > instance, or the "Merge" phase of sort/merge) and the impact of making > the CPU (or "CP") reliable imposes its own burden (Ref: Appendix "A" > of "Linux for the S/390"). > > Yeah, you can do it. Maybe you'd be best looking (memory failure, > distro name not handy!!!) at the distro that downloads source packages > and compiles them as you build up the system (dammit, I can't believe > the hole in my memory). > > (sighs) > > I can't trust my memory because it isn't accurate: I recall looking a > lot younger 30 years ago. > > -soup > > -- > John R. Campbell Speaker to Machines souperb at gmail dot com > MacOS X proved it was easier to make Unix user-friendly than to fix Windows > > ---------------------------------------------------------------------- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > ---------------------------------------------------------------------- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 ---------------------------------------------------------------------- For more information on Linux on System z, visit http://wiki.linuxvm.org/
