NAME

pmc.haswelluc —

uncore measurement events for Intel Haswell family CPUs

LIBRARY

Performance Counters Library (libpmc, -lpmc)

SYNOPSIS

#include <pmc.h>

DESCRIPTION

Intel Haswell CPUs contain PMCs conforming to version 3 of the Intel performance measurement architecture. These CPUs contain two classes of PMCs:

PMC_CLASS_UCF: Fixed-function counters that count only one hardware event per counter.
PMC_CLASS_UCP: Programmable counters that may be configured to count one of a defined set of hardware events.

The number of PMCs available in each class and their widths need to be determined at run time by calling pmc_cpuinfo(3).

Intel Haswell PMCs are documented in Combined Volumes: 1, 2A, 2B, 2C, 3A, 3B and 3C, Intel(R) 64 and IA-32 Architectures Software Developers Manual, Order Number: 325462-045US, Intel Corporation, January 2013.

HASWELL UNCORE FIXED FUNCTION PMCS

These PMCs and their supported events are documented in pmc.ucf(3). Not all CPUs in this family implement fixed-function counters.

HASWELL UNCORE PROGRAMMABLE PMCS

The programmable PMCs support the following capabilities:

Capability	Support
PMC_CAP_CASCADE	No
PMC_CAP_EDGE	Yes
PMC_CAP_INTERRUPT	No
PMC_CAP_INVERT	Yes
PMC_CAP_READ	Yes
PMC_CAP_PRECISE	No
PMC_CAP_SYSTEM	No
PMC_CAP_TAGGING	No
PMC_CAP_THRESHOLD	Yes
PMC_CAP_USER	No
PMC_CAP_WRITE	Yes

Event Qualifiers

Event specifiers for these PMCs support the following common qualifiers:

cmask=value: Configure the PMC to increment only if the number of configured events measured in a cycle is greater than or equal to value.
edge: Configure the PMC to count the number of de-asserted to asserted transitions of the conditions expressed by the other qualifiers. If specified, the counter will increment only once whenever a condition becomes true, irrespective of the number of clocks during which the condition remains true.
inv: Invert the sense of comparison when the “cmask” qualifier is present, making the counter increment when the number of events per cycle is less than the value specified by the “cmask” qualifier.

Event Specifiers (Programmable PMCs)

Haswell programmable PMCs support the following events:

UNC_CBO_XSNP_RESPONSE.MISS: (Event 22H, Umask 01H) A snoop misses in some processor core.
UNC_CBO_XSNP_RESPONSE.INVAL: (Event 22H, Umask 02H) A snoop invalidates a non-modified line in some processor core.
UNC_CBO_XSNP_RESPONSE.HIT: (Event 22H, Umask 04H) A snoop hits a non-modified line in some processor core.
UNC_CBO_XSNP_RESPONSE.HITM: (Event 22H, Umask 08H) A snoop hits a modified line in some processor core.
UNC_CBO_XSNP_RESPONSE.INVAL_M: (Event 22H, Umask 10H) A snoop invalidates a modified line in some processor core.
UNC_CBO_XSNP_RESPONSE.EXTERNAL_FILTER: (Event 22H, Umask 20H) Filter on cross-core snoops initiated by this Cbox due to external snoop request.
UNC_CBO_XSNP_RESPONSE.XCORE_FILTER: (Event 22H, Umask 40H) Filter on cross-core snoops initiated by this Cbox due to processor core memory request.
UNC_CBO_XSNP_RESPONSE.EVICTION_FILTER: (Event 22H, Umask 80H) Filter on cross-core snoops initiated by this Cbox due to LLC eviction.
UNC_CBO_CACHE_LOOKUP.M: (Event 34H, Umask 01H) LLC lookup request that access cache and found line in M-state.
UNC_CBO_CACHE_LOOKUP.ES: (Event 34H, Umask 06H) LLC lookup request that access cache and found line in E or S state.
UNC_CBO_CACHE_LOOKUP.I: (Event 34H, Umask 08H) LLC lookup request that access cache and found line in I-state.
UNC_CBO_CACHE_LOOKUP.READ_FILTER: (Event 34H, Umask 10H) Filter on processor core initiated cacheable read requests. Must combine with at least one of 01H, 02H, 04H, 08H.
UNC_CBO_CACHE_LOOKUP.WRITE_FILTER: (Event 34H, Umask 20H) Filter on processor core initiated cacheable write requests. Must combine with at least one of 01H, 02H, 04H, 08H.
UNC_CBO_CACHE_LOOKUP.EXTSNP_FILTER: (Event 34H, Umask 40H) Filter on external snoop requests. Must combine with at least one of 01H, 02H, 04H, 08H.
UNC_CBO_CACHE_LOOKUP.ANY_REQUEST_FILTER: (Event 34H, Umask 80H) Filter on any IRQ or IPQ initiated requests including uncacheable, non-coherent requests. Must combine with at least one of 01H, 02H, 04H, 08H.
UNC_ARB_TRK_OCCUPANCY.ALL: (Event 80H, Umask 01H) Counts cycles weighted by the number of requests waiting for data returning from the memory controller. Accounts for coherent and non-coherent requests initiated by IA cores, processor graphic units, or LLC.
UNC_ARB_TRK_REQUEST.ALL: (Event 81H, Umask 01H) Counts the number of coherent and in-coherent requests initiated by IA cores, processor graphic units, or LLC.
UNC_ARB_TRK_REQUEST.WRITES: (Event 81H, Umask 20H) Counts the number of allocated write entries, include full, partial, and LLC evictions.
UNC_ARB_TRK_REQUEST.EVICTIONS: (Event 81H, Umask 80H) Counts the number of LLC evictions allocated.
UNC_ARB_COH, Umask TRK_OCCUPANCY.ALL: (Event 83H, Umask 01H) Cycles weighted by number of requests pending in Coherency Tracker.
UNC_ARB_COH, Umask TRK_REQUEST.ALL: (Event 84H, Umask 01H) Number of requests allocated in Coherency Tracker.

HISTORY

The pmc library first appeared in FreeBSD 6.0.

AUTHORS

The Performance Counters Library (libpmc, -lpmc) library was written by Joseph Koshy <jkoshy@FreeBSD.org>. The support for the Haswell microarchitecture was added by Hiren Panchasara <hiren.panchasara@gmail.com>.