Message boards :
Cafe LHC :
A story of fortran bug
Message board moderation
Author | Message |
---|---|
Send message Joined: 17 Sep 04 Posts: 190 Credit: 649,637 RAC: 0 |
An old AMD 2500 did a great job with lhc last year. Running over months for LHC, actually never got serious problems. Meanwhile connected to other projects, also no problems as well. At the hardware and the framework attitudes was not changed anything. After LHC reopened, the AMD again got work from LHC. Unfortunately, the host captured an overdose of Fortran errors. The Temperature was O.K., even running upper the speficications. Reduced and more reduced the "bus-speed/cpu-speed" settings, Fortran error still there Finally, it's stable now, fortran has gone. Obviously, the cause was to agressive Memory timings. But why was it runnig so well last year? |
©2024 CERN