There is something seductive about problem solving in technology. We, as technologists, are lured into a sense that it will take just a couple of minutes to solve a pressing issue. OR, we will only give a few minutes to a problem and, if we aren’t e…

Hey Lucas….found out something about the disk/CPW relationship….12 SAS disks do not equal 1200cpw…..you have to mirror, so they only equal 600cpw….per IBM….this is only interactive cpw…

Ed

so…this could be your problem here…

You shouldn’t have to recompile on the power6 machine…supposed to already utilize new functionality…

Let me know if there is any difference, but there shouldn’t be any at all….

sounds like you are losing performance some where in the virtual machine/io….

Both machines are running V6R1. The programs were compiled on the Power 5 machine, though. As far as i’ve read, the machine code itself should automatically be retranslated when restored on a Power 6 machine (the initial delay before starting STROBJCVN on the program libraries seems to agree with this).

Just to make sure, i’ll ask our developers to run a full recompile on the Power 6 blade. This might take some time, though.

Regards,

Lukas

You didn’t state whether during
the benchmark testing you were comparing
V6R1 compiled programs or V5R4…

I have heard that V6r1 on power6 supposedly
runs faster…..

Any thoughts on this??