FLYNN%EVALUN11.BITNET@cunyvm.cuny.edu (Mark F. Flynn) (05/08/91)
We're running HP-UX 7.05 on an Apollo 400t, and have just installed the new and wonderful 68040 processor. What we have noticed is that certain jobs spend from 50% to 90% of their time doing system calls, as opposed to user time. This is reported both by "time" and a command we have here called "monitor". I havn't figured out what sort of code actually produces this yet. So far, we have only seen it on Fortran code, but this may just be that we havn't run the right (or wrong) C program yet. Optimization levels have no signifigant effect. Changing the type of cacheing similar does nothing. The HP people here havn't a clue. Has anyone experienced a similar problem? Any suggestions? Thanks in advance. Mark Flynn FLYNN@EVALUN11.(EARN BITNET) Departamto de Fisica Atomica y Nuclear Universidad de Valencia Spain
guru@buhub.bradley.edu (Jerry Whelan) (05/09/91)
In article <26811@adm.brl.mil> FLYNN%EVALUN11.BITNET@cunyvm.cuny.edu (Mark F. Flynn) writes:
-}
-} We're running HP-UX 7.05 on an Apollo 400t, and have just installed the
-} new and wonderful 68040 processor. What we have noticed is that certain
-} jobs spend from 50% to 90% of their time doing system calls, as opposed
-} to user time. This is reported both by "time" and a command we have
Have you tried recompiling?
On the new next machines people have encountered a similar problem.
Now that the FPU is built into the 040, certain transcendental functions
(i.e. sin, cos, tan) work 'differently.' The people at next decided to
handle the difference by providing backwards compatibility such that programs
compiled for an 030 with FPU take the least performance hit possible.
However, this performance hit is still in the 2x-3x range. On the next
they also goofed in that programs specifically compiled for the 040 do
_not_ take advantage of the new stuff and suffer the same 2x-3x slowdown.
It was the general consensus of comp.sys.next that it would not be
particularly hard for next to set things up so that 040 programs ran at
full speed, perhaps HP did their's correctly and a recompile is all
that you need.
If you want more explicit information, I recommend you the
comp.sys.next postings over the last two months or so, there was quite
a furour about it.
-------------------------------------------------------------------------------
"I'm not sure what I mean, so I'm going to listen to what I say."
guru@ (buhub.bradley.edu || bucc1.bradley.edu) || whelan@wiliki.eng.hawaii.edu