Hi,
I was reviewing some test results issues with Michael Petlan. Some of the
testing at Red Hat is done on guest VMs. PAPI and libpfm use the CPU model
number to determine which performance events are available. This works
reasonably well for many of the events. However, there are a couple of
situations where this doesn't work:
- Events that need to program additional offcore MSRs such as PAPI's
PRESET,PAPI_CA_INV,NOT_DERIVED,OFFCORE_RESPONSE_0:HITM
- Events that have additional machines constraints such as (0xb0 events have
note in the Intel software developer's manual "Use only when HTT is off."):
PRESET,PAPI_CA_SHR,NOT_DERIVED,OFFCORE_REQUESTS:ALL_DATA_RD
Any thoughts on how to better determine which events are actually available on
the machine and limit the libpfm/PAPI event lists to usable events?
-Will
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
perfmon2-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
