libpfm_intel_knm — support for Intel Knights Mill core PMU
#include <perfmon/pfmlib.h> PMU name: knm PMU desc: Intel Kinghts Mill
The library supports the Intel Kinghts Mill core PMU. It should be noted that this PMU model only covers each core's PMU and not the socket level PMU.
On Knights Mill, the number of generic counters is 4. There is 4-way HyperThreading support. The pfm_get_pmu_info() function returns the maximum number of generic counters in num_cntrs.
The following modifiers are supported on Intel Kinghts Mill processors:
Measure at user level which includes privilege levels 1, 2, 3. This corresponds to PFM_PLM3. This is a boolean modifier.
Measure at kernel level which includes privilege level 0. This corresponds to PFM_PLM0. This is a boolean modifier.
Invert the meaning of the event. The counter will now count cycles in which the event is not occurring. This is a boolean modifier
Enable edge detection, i.e., count only when there is a state transition from no occurrence of the event to at least one occurrence. This modifier must be combined with a counter mask modifier (m) with a value greater or equal to one. This is a boolean modifier.
Set the counter mask value. The mask acts as a threshold. The counter will count the number of cycles in which the number of occurrences of the event is greater or equal to the threshold. This is an integer modifier with values in the range [0:255].
Measure on any of the 4 hyper-threads at the same time assuming hyper-threading is enabled. This is a boolean modifier. This modifier is only available on fixed counters (unhalted_reference_cycles, instructions_retired, unhalted_core_cycles). Depending on the underlying kernel interface, the event may be programmed on a fixed counter or a generic counter, except for unhalted_reference_cycles, in which case, this modifier may be ignored or rejected.
Intel Knights Mill provides two offcore_response events. They are called OFFCORE_RESPONSE_0 and OFFCORE_RESPONSE_1.
Those events need special treatment in the performance monitoring infrastructure because each event uses an extra register to store some settings. Thus, in case multiple offcore_response events are monitored simultaneously, the kernel needs to manage the sharing of that extra register.
The offcore_response events are exposed as normal events by the library. The extra settings are exposed as regular umasks. The library takes care of encoding the events according to the underlying kernel interface.
On Intel Knights Mill, the umasks are divided into 4 categories: request, supplier and snoop and average latency. Offcore_response event has two modes of operations: normal and average latency. In the first mode, the two offcore_respnse events operate independently of each other. The user must provide at least one umask for each of the first 3 categories: request, supplier, snoop. In the second mode, the two offcore_response events are combined to compute an average latency per request type.
For the normal mode, there is a special supplier (response) umask called ANY_RESPONSE. When this umask is used then it overrides any supplier and snoop umasks. In other words, users can specify either ANY_RESPONSE OR any combinations of supplier + snoops. In case no supplier or snoop is specified, the library defaults to using ANY_RESPONSE.
For instance, the following are valid event selections:
But the following is illegal:
In average latency mode, OFFCORE_RESPONSE_0 must be programmed to select the request types of interest, for instance, DMND_DATA_RD, and the OUTSTANDING umask must be set and no others. the library will enforce that restriction as soon as the OUTSTANDING umask is used. Then OFFCORE_RESPONSE_1 must be set with the same request types and the ANY_RESPONSE umask. It should be noted that the library encodes events independently of each other and therefore cannot verify that the requests are matching between the two events. Example of average latency settings:
The average latency for the request(s) is obtained by dividing the counts of OFFCORE_RESPONSE_0 by the count of OFFCORE_RESPONSE_1. The ratio is expressed in core cycles.