Major A writes: > I've just checked, the central file space is 160GB, which is about 50% > full right now. It's shared via NFS, unfortunately, so it's not that > good really. They still have impressive computing power, I've just > checked that they have a 4x800MHz Itanium and a 4x1GHz Alpha. The > Itanium has 4GB of RAM, as many of the SMP systems they have. > > All in all, if it's really just copying data around, then it would be > best to do it on a PC with a lot of (local) disk space. From your > experience, what difference does CPU speed, the number of CPUs, and > the amount of RAM make? As someone else has said, disk space isn't > that expensive, and if we get a donation, you could set up a new > heater^H^H^H^H^H^HPC in your house with a 1TB RAID made of 4 IDE disks > or so.
I appreciate the discussion of donations [to me] :-) but to be fair, right now things are ok. More hardware == faster, but things are ok/fine right now. Seriously, if you have a burning desire to make some sort of donation, please consider donating to someone who might be having trouble just getting enough food, or basic medical attention, or shelter. But, to answer your question CPU speed does definitely help. Generally I'm never memory bound on a 256 machine except for the one time task of splitting up the world land mass data set into tiles... it would have been nice to have 1Gb RAM for that. Even though there is a large data shuffling component, I have found significant advantages to clustering machines and running several tile builders at once, hanging off a single (fast) nfs server. I've never done any time measurements though to see how things scaled. Regards, Curt. -- Curtis Olson IVLab / HumanFIRST Program FlightGear Project Twin Cities [EMAIL PROTECTED] [EMAIL PROTECTED] Minnesota http://www.menet.umn.edu/~curt http://www.flightgear.org _______________________________________________ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
