We're running VTune on a 256 core SGI Linux machine.
In the syslog, we get the following messages when I start VTune:
Kernel : vtss_cmd_start[cpu98]: starting vtss++ collection
Kernel: vtss_cmd_start[cpu98]: starting vtss++ collection continue 1
Kernel: vtss_cmd_start[cpu96]: starting vtss++ collection continue 10
…then all the errors spew out….
Kernel: uhhuh. NMI received for unknown reason 00 on CPU 74
Kernel: do you have a strange power saving mode enabled?
Kernel: dazed and confused, but trying to continue
Same messages spew out for multiple cpus, 74, 32, 35, 107, 226, 43, 227, 105, 70
Mostly 32-35 range
…ends with these messages when vtune is stopped:
Kernel: vtss_procfs_ctrl_write[cpu96]: STOP
Kernel:vtss_task_map_fini[cpu185] : item is busy now, key=72780, usage=1
Kernel: vtss_transport_fini[cpu185]: ‘72773-72780.1’ lost 36 events
Kernel: vtss_transport_fini[cpu185]: ‘72872-72894.1’ lost 419 events
…more of the same lines above with different numbers of events lost….
Kernel: vtss_collection_fini[cpu185]: vtss++ collection stopped
At first we thought that somehow VTune was highlighting a hardware problem but the SGI folks feel that this is an Intel VTune problem.
We also will sometimes get a seg fault but I've neglected to catch the output.
Any suggestions on what to try next?
This last Monday I registered for Premier Support. I received an email stating that there is a 1 hour delay while my account is being "verified". I've continued to check and my account is still incomplete. There's a "Want more help? Contact support." link but it doesn't provide any actual contact information.
Can you suggest a link or email or telephone number that I can use to move this Premier Support process forward?
I'm really sorry about that. Stupid computers. ;)
Did you use the same email address as you are using on the forum? If not, please PM me with that info and I will look into what is going on at IPS.