xref: /freebsd/lib/libpmc/pmc.sandybridgexeon.3 (revision 2e3f49888ec8851bafb22011533217487764fdb0)
1.\" Copyright (c) 2012 Hiren Panchasara <hiren.panchasara@gmail.com>
2.\" All rights reserved.
3.\"
4.\" Redistribution and use in source and binary forms, with or without
5.\" modification, are permitted provided that the following conditions
6.\" are met:
7.\" 1. Redistributions of source code must retain the above copyright
8.\"    notice, this list of conditions and the following disclaimer.
9.\" 2. Redistributions in binary form must reproduce the above copyright
10.\"    notice, this list of conditions and the following disclaimer in the
11.\"    documentation and/or other materials provided with the distribution.
12.\"
13.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
14.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
15.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
16.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
17.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
18.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
19.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
20.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
21.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
22.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
23.\" SUCH DAMAGE.
24.\"
25.Dd October 18, 2012
26.Dt PMC.SANDYBRIDGEXEON 3
27.Os
28.Sh NAME
29.Nm pmc.sandybridgexeon
30.Nd measurement events for
31.Tn Intel
32.Tn Sandy Bridge Xeon
33family CPUs
34.Sh LIBRARY
35.Lb libpmc
36.Sh SYNOPSIS
37.In pmc.h
38.Sh DESCRIPTION
39.Tn Intel
40.Tn "Sandy Bridge Xeon"
41CPUs contain PMCs conforming to version 2 of the
42.Tn Intel
43performance measurement architecture.
44These CPUs may contain up to two classes of PMCs:
45.Bl -tag -width "Li PMC_CLASS_IAP"
46.It Li PMC_CLASS_IAF
47Fixed-function counters that count only one hardware event per counter.
48.It Li PMC_CLASS_IAP
49Programmable counters that may be configured to count one of a defined
50set of hardware events.
51.El
52.Pp
53The number of PMCs available in each class and their widths need to be
54determined at run time by calling
55.Xr pmc_cpuinfo 3 .
56.Pp
57Intel Sandy Bridge Xeon PMCs are documented in
58.Rs
59.%B "Intel(R) 64 and IA-32 Architectures Software Developer's Manual"
60.%T "Volume 3B: System Programming Guide, Part 2"
61.%N "Order Number: 253669-043US"
62.%D August 2012
63.%Q "Intel Corporation"
64.Re
65.Ss SANDYBRIDGE XEON FIXED FUNCTION PMCS
66These PMCs and their supported events are documented in
67.Xr pmc.iaf 3 .
68.Ss SANDYBRIDGE XEON PROGRAMMABLE PMCS
69The programmable PMCs support the following capabilities:
70.Bl -column "PMC_CAP_INTERRUPT" "Support"
71.It Em Capability Ta Em Support
72.It PMC_CAP_CASCADE Ta \&No
73.It PMC_CAP_EDGE Ta Yes
74.It PMC_CAP_INTERRUPT Ta Yes
75.It PMC_CAP_INVERT Ta Yes
76.It PMC_CAP_READ Ta Yes
77.It PMC_CAP_PRECISE Ta \&No
78.It PMC_CAP_SYSTEM Ta Yes
79.It PMC_CAP_TAGGING Ta \&No
80.It PMC_CAP_THRESHOLD Ta Yes
81.It PMC_CAP_USER Ta Yes
82.It PMC_CAP_WRITE Ta Yes
83.El
84.Ss Event Qualifiers
85Event specifiers for these PMCs support the following common
86qualifiers:
87.Bl -tag -width indent
88.It Li rsp= Ns Ar value
89Configure the Off-core Response bits.
90.Bl -tag -width indent
91.It Li REQ_DMND_DATA_RD
92Counts the number of demand and DCU prefetch data reads of full and partial
93cachelines as well as demand data page table entry cacheline reads.
94Does not count L2 data read prefetches or instruction fetches.
95.It Li REQ_DMND_RFO
96Counts the number of demand and DCU prefetch reads for ownership (RFO)
97requests generated by a write to data cacheline.
98Does not count L2 RFO prefetches.
99.It Li REQ_DMND_IFETCH
100Counts the number of demand and DCU prefetch instruction cacheline reads.
101Does not count L2 code read prefetches.
102.It Li REQ_WB
103Counts the number of writeback (modified to exclusive) transactions.
104.It Li REQ_PF_DATA_RD
105Counts the number of data cacheline reads generated by L2 prefetchers.
106.It Li REQ_PF_RFO
107Counts the number of RFO requests generated by L2 prefetchers.
108.It Li REQ_PF_IFETCH
109Counts the number of code reads generated by L2 prefetchers.
110.It Li REQ_PF_LLC_DATA_RD
111L2 prefetcher to L3 for loads.
112.It Li REQ_PF_LLC_RFO
113RFO requests generated by L2 prefetcher
114.It Li REQ_PF_LLC_IFETCH
115L2 prefetcher to L3 for instruction fetches.
116.It Li REQ_BUS_LOCKS
117Bus lock and split lock requests.
118.It Li REQ_STRM_ST
119Streaming store requests.
120.It Li REQ_OTHER
121Any other request that crosses IDI, including I/O.
122.It Li RES_ANY
123Catch all value for any response types.
124.It Li RES_SUPPLIER_NO_SUPP
125No Supplier Information available.
126.It Li RES_SUPPLIER_LLC_HITM
127M-state initial lookup stat in L3.
128.It Li RES_SUPPLIER_LLC_HITE
129E-state.
130.It Li RES_SUPPLIER_LLC_HITS
131S-state.
132.It Li RES_SUPPLIER_LLC_HITF
133F-state.
134.It Li RES_SUPPLIER_LOCAL
135Local DRAM Controller.
136.It Li RES_SNOOP_SNP_NONE
137No details on snoop-related information.
138.It Li RES_SNOOP_SNP_NO_NEEDED
139No snoop was needed to satisfy the request.
140.It Li RES_SNOOP_SNP_MISS
141A snoop was needed and it missed all snooped caches:
142-For LLC Hit, ReslHitl was returned by all cores
143-For LLC Miss, Rspl was returned by all sockets and data was returned from
144DRAM.
145.It Li RES_SNOOP_HIT_NO_FWD
146A snoop was needed and it hits in at least one snooped cache.
147Hit denotes a cache-line was valid before snoop effect.
148This includes:
149-Snoop Hit w/ Invalidation (LLC Hit, RFO)
150-Snoop Hit, Left Shared (LLC Hit/Miss, IFetch/Data_RD)
151-Snoop Hit w/ Invalidation and No Forward (LLC Miss, RFO Hit S)
152In the LLC Miss case, data is returned from DRAM.
153.It Li RES_SNOOP_HIT_FWD
154A snoop was needed and data was forwarded from a remote socket.
155This includes:
156-Snoop Forward Clean, Left Shared (LLC Hit/Miss, IFetch/Data_RD/RFT).
157.It Li RES_SNOOP_HITM
158A snoop was needed and it HitM-ed in local or remote cache.
159HitM denotes a cache-line was in modified state before effect as a results of snoop.
160This includes:
161-Snoop HitM w/ WB (LLC miss, IFetch/Data_RD)
162-Snoop Forward Modified w/ Invalidation (LLC Hit/Miss, RFO)
163-Snoop MtoS (LLC Hit, IFetch/Data_RD).
164.It Li RES_NON_DRAM
165Target was non-DRAM system address.
166This includes MMIO transactions.
167.El
168.It Li cmask= Ns Ar value
169Configure the PMC to increment only if the number of configured
170events measured in a cycle is greater than or equal to
171.Ar value .
172.It Li edge
173Configure the PMC to count the number of de-asserted to asserted
174transitions of the conditions expressed by the other qualifiers.
175If specified, the counter will increment only once whenever a
176condition becomes true, irrespective of the number of clocks during
177which the condition remains true.
178.It Li inv
179Invert the sense of comparison when the
180.Dq Li cmask
181qualifier is present, making the counter increment when the number of
182events per cycle is less than the value specified by the
183.Dq Li cmask
184qualifier.
185.It Li os
186Configure the PMC to count events happening at processor privilege
187level 0.
188.It Li usr
189Configure the PMC to count events occurring at privilege levels 1, 2
190or 3.
191.El
192.Pp
193If neither of the
194.Dq Li os
195or
196.Dq Li usr
197qualifiers are specified, the default is to enable both.
198.Ss Event Specifiers (Programmable PMCs)
199Sandy Bridge Xeon programmable PMCs support the following events:
200.Bl -tag -width indent
201.It Li LD_BLOCKS.DATA_UNKNOWN
202.Pq Event 03H , Umask 01H
203blocked loads due to store buffer blocks with unknown data.
204.It Li LD_BLOCKS.STORE_FORWARD
205.Pq Event 03H , Umask 02H
206loads blocked by overlapping with store buffer that cannot
207be forwarded .
208.It Li LD_BLOCKS.NO_SR
209.Pq Event 03H , Umask 08H
210# of Split loads blocked due to resource not available.
211.It Li LD_BLOCKS.ALL_BLOCK
212.Pq Event 03H , Umask 10H
213Number of cases where any load is blocked but has no
214DCU miss.
215.It Li MISALIGN_MEM_REF.LOADS
216.Pq Event 05H , Umask 01H
217Speculative cache-line split load uops dispatched to
218L1D.
219.It Li MISALIGN_MEM_REF.STORES
220.Pq Event 05H , Umask 02H
221Speculative cache-line split Store- address uops
222dispatched to L1D.
223.It Li LD_BLOCKS_PARTIAL.ADDRESS_ALIAS
224.Pq Event 07H , Umask 01H
225False dependencies in MOB due to partial compare on
226address.
227.It Li LD_BLOCKS_PARTIAL.ALL_STALL_BLOCK
228.Pq Event 07H , Umask 08H
229The number of times that load operations are temporarily
230blocked because of older stores, with addresses that are
231not yet known.
232A load operation may incur more than one block of this type.
233.It Li TLB_LOAD_MISSES.MISS_CAUSES_A_WALK
234.Pq Event 08H , Umask 01H
235Misses in all TLB levels that cause a page walk of any
236page size.
237.It Li TLB_LOAD_MISSES.WALK_COMPLETED
238.Pq Event 08H , Umask 02H
239Misses in all TLB levels that caused page walk completed
240of any size.
241.It Li DTLB_LOAD_MISSES.WALK_DURATION
242.Pq Event 08H , Umask 04H
243Cycle PMH is busy with a walk.
244.It Li DTLB_LOAD_MISSES.STLB_HIT
245.Pq Event 08H , Umask 10H
246Number of cache load STLB hits.
247No page walk.
248.It Li INT_MISC.RECOVERY_CYCLES
249.Pq Event 0DH , Umask 03H
250Cycles waiting to recover after Machine Clears or EClear.
251Set Cmask= 1.
252.It Li INT_MISC.RAT_STALL_CYCLES
253.Pq Event 0DH , Umask 40H
254Cycles RAT external stall is sent to IDQ for this thread.
255.It Li UOPS_ISSUED.ANY
256.Pq Event 0EH , Umask 01H
257Increments each cycle the # of Uops issued by the
258RAT to RS.
259Set Cmask = 1, Inv = 1, Any= 1to count stalled cycles
260of this core.
261.It Li FP_COMP_OPS_EXE.X87
262.Pq Event 10H , Umask 01H
263Counts number of X87 uops executed.
264.It Li FP_COMP_OPS_EXE.SSE_FP_PACKED_DOUBLE
265.Pq Event 10H , Umask 10H
266Counts number of SSE* double precision FP packed
267uops executed.
268.It Li FP_COMP_OPS_EXE.SSE_FP_SCALAR_SINGLE
269.Pq Event 10H , Umask 20H
270Counts number of SSE* single precision FP scalar
271uops executed.
272.It Li FP_COMP_OPS_EXE.SSE_PACKED_SINGLE
273.Pq Event 10H , Umask 40H
274Counts number of SSE* single precision FP packed
275uops executed.
276.It Li FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE
277.Pq Event 10H , Umask 80H
278Counts number of SSE* double precision FP scalar
279uops executed.
280.It Li SIMD_FP_256.PACKED_SINGLE
281.Pq Event 11H , Umask 01H
282Counts 256-bit packed single-precision floating-
283point instructions.
284.It Li SIMD_FP_256.PACKED_DOUBLE
285.Pq Event 11H , Umask 02H
286Counts 256-bit packed double-precision floating-
287point instructions.
288.It Li ARITH.FPU_DIV_ACTIVE
289.Pq Event 14H , Umask 01H
290Cycles that the divider is active, includes INT and FP.
291Set 'edge =1, cmask=1' to count the number of
292divides.
293.It Li INSTS_WRITTEN_TO_IQ.INSTS
294.Pq Event 17H , Umask 01H
295Counts the number of instructions written into the
296IQ every cycle.
297.It Li L2_RQSTS.DEMAND_DATA_RD_HIT
298.Pq Event 24H , Umask 01H
299Demand Data Read requests that hit L2 cache.
300.It Li L2_RQSTS.ALL_DEMAND_DATA_RD
301.Pq Event 24H , Umask 03H
302Counts any demand and L1 HW prefetch data load
303requests to L2.
304.It Li L2_RQSTS.RFO_HITS
305.Pq Event 24H , Umask 04H
306Counts the number of store RFO requests that
307hit the L2 cache.
308.It Li L2_RQSTS.RFO_MISS
309.Pq Event 24H , Umask 08H
310Counts the number of store RFO requests that
311miss the L2 cache.
312.It Li L2_RQSTS.ALL_RFO
313.Pq Event 24H , Umask 0CH
314Counts all L2 store RFO requests.
315.It Li L2_RQSTS.CODE_RD_HIT
316.Pq Event 24H , Umask 10H
317Number of instruction fetches that hit the L2
318cache.
319.It Li L2_RQSTS.CODE_RD_MISS
320.Pq Event 24H , Umask 20H
321Number of instruction fetches that missed the L2
322cache.
323.It Li L2_RQSTS.ALL_CODE_RD
324.Pq Event 24H , Umask 30H
325Counts all L2 code requests.
326.It Li L2_RQSTS.PF_HIT
327.Pq Event 24H , Umask 40H
328Requests from L2 Hardware prefetcher that hit L2.
329.It Li L2_RQSTS.PF_MISS
330.Pq Event 24H , Umask 80H
331Requests from L2 Hardware prefetcher that missed
332L2.
333.It Li L2_RQSTS.ALL_PF
334.Pq Event 24H , Umask C0H
335Any requests from L2 Hardware prefetchers.
336.It Li L2_STORE_LOCK_RQSTS.MISS
337.Pq Event 27H , Umask 01H
338ROs that miss cache lines.
339.It Li L2_STORE_LOCK_RQSTS.HIT_E
340.Pq Event 27H , Umask 04H
341RFOs that hit cache lines in E state.
342.It Li L2_STORE_LOCK_RQSTS.HIT_M
343.Pq Event 27H , Umask 08H
344RFOs that hit cache lines in M state.
345.It Li L2_STORE_LOCK_RQSTS.ALL
346.Pq Event 27H , Umask 0FH
347RFOs that access cache lines in any state.
348.It Li L2_L1D_WB_RQSTS.MISS
349.Pq Event 28H , Umask 01H
350Not rejected writebacks from L1D to L2 cache lines
351that missed L2.
352.It Li L2_L1D_WB_RQSTS.HIT_S
353.Pq Event 28H , Umask 02H
354Not rejected writebacks from L1D to L2 cache lines
355in S state.
356.It Li L2_L1D_WB_RQSTS.HIT_E
357.Pq Event 28H , Umask 04H
358Not rejected writebacks from L1D to L2 cache lines
359in E state.
360.It Li L2_L1D_WB_RQSTS.HIT_M
361.Pq Event 28H , Umask 08H
362Not rejected writebacks from L1D to L2 cache lines
363in M state.
364.It Li L2_L1D_WB_RQSTS.ALL
365.Pq Event 28H , Umask 0FH
366Not rejected writebacks from L1D to L2 cache.
367.It Li LONGEST_LAT_CACHE.REFERENCE
368.Pq Event 2EH , Umask 4FH
369This event counts requests originating from the
370core that reference
371a cache line in the last level cache.
372.It Li LONGEST_LAT_CACHE.MISS
373.Pq Event 2EH , Umask 41H
374This event counts each cache miss condition for
375references to the last level cache.
376.It Li CPU_CLK_UNHALTED.THREAD_P
377.Pq Event 3CH , Umask 00H
378Counts the number of thread cycles while the
379thread is not in a halt state.
380The thread enters the halt state when it is running the HLT
381instruction.
382The core frequency may change from time to time due to power or thermal throttling.
383.It Li CPU_CLK_THREAD_UNHALTED.REF_XCLK
384.Pq Event 3CH , Umask 01H
385Increments at the frequency of XCLK (100 MHz)
386when not halted.
387.It Li L1D_PEND_MISS.PENDING
388.Pq Event 48H , Umask 01H
389Increments the number of outstanding L1D misses
390every cycle.
391Set Cmaks = 1 and Edge =1 to count occurrences.
392.It Li DTLB_STORE_MISSES.MISS_CAUSES_A_WALK
393.Pq Event 49H , Umask 01H
394Miss in all TLB levels causes an page walk of
395any page size (4K/2M/4M/1G).
396.It Li DTLB_STORE_MISSES.WALK_COMPLETED
397.Pq Event 49H , Umask 02H
398Miss in all TLB levels causes a page walk that
399completes of any page size (4K/2M/4M/1G).
400.It Li DTLB_STORE_MISSES.WALK_DURATION
401.Pq Event 49H , Umask 04H
402Cycles PMH is busy with this walk.
403.It Li DTLB_STORE_MISSES.STLB_HIT
404.Pq Event 49H , Umask 10H
405Store operations that miss the first TLB level
406but hit the second and do not cause page walks.
407.It Li LOAD_HIT_PRE.SW_PF
408.Pq Event 4CH , Umask 01H
409Not SW-prefetch load dispatches that hit fill
410buffer allocated for S/W prefetch.
411.It Li LOAD_HIT_PER.HW_PF
412.Pq Event 4CH , Umask 02H
413Not SW-prefetch load dispatches that hit fill
414buffer allocated for H/W prefetch.
415.It Li HW_PRE_REQ.DL1_MISS
416.Pq Event 4EH , Umask 02H
417Hardware Prefetch requests that miss the L1D cache.
418A request is being counted each time it access the cache
419& miss it, including if a block is applicable or if hit the Fill
420Buffer for example.
421.It Li L1D.REPLACEMENT
422.Pq Event 51H , Umask 01H
423Counts the number of lines brought into the
424L1 data cache.
425.It Li L1D.ALLOCATED_IN_M
426.Pq Event 51H , Umask 02H
427Counts the number of allocations of modified
428L1D cache lines.
429.It Li L1D.EVICTION
430.Pq Event 51H , Umask 04H
431Counts the number of modified lines evicted
432from the L1 data cache due to replacement.
433.It Li L1D.ALL_M_REPLACEMENT
434.Pq Event 51H , Umask 08H
435Cache lines in M state evicted out of L1D due
436to Snoop HitM or dirty line replacement.
437.It Li PARTIAL_RAT_STALLS.FLAGS_MERGE_UOP
438.Pq Event 59H , Umask 0CH
439Increments the number of flags-merge uops in
440flight each cycle.
441Set Cmask = 1 to count cycles.
442.It Li PARTIAL_RAT_STALLS.SLOW_LEA_WINDOW
443.Pq Event 59H , Umask 0FH
444Cycles with at least one slow LEA uop allocated.
445.It Li PARTIAL_RAT_STALLS.MUL_SINGLE_UOP
446.Pq Event 59H , Umask 40H
447Number of Multiply packed/scalar single precision
448uops allocated.
449.It Li RESOURCE_STALLS2.ALL_FL_EMPTY
450.Pq Event 5BH , Umask 0CH
451Cycles stalled due to free list empty.
452.It Li RESOURCE_STALLS2.ALL_PRF_CONTROL
453.Pq Event 5BH , Umask 0FH
454Cycles stalled due to control structures full for
455physical registers.
456.It Li RESOURCE_STALLS2.BOB_FULL
457.Pq Event 5BH , Umask 40H
458Cycles Allocator is stalled due Branch Order Buffer.
459.It Li RESOURCE_STALLS2.OOO_RSRC
460.Pq Event 5BH , Umask 4FH
461Cycles stalled due to out of order resources full.
462.It Li CPL_CYCLES.RING0
463.Pq Event 5CH , Umask 01H
464Unhalted core cycles when the thread is in ring 0.
465.It Li CPL_CYCLES.RING123
466.Pq Event 5CH , Umask 02H
467Unhalted core cycles when the thread is not in ring
4680.
469.It Li RS_EVENTS.EMPTY_CYCLES
470.Pq Event 5EH , Umask 01H
471Cycles the RS is empty for the thread.
472.It Li OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD
473.Pq Event 60H , Umask 01H
474Offcore outstanding Demand Data Read
475transactions in SQ to uncore.
476Set Cmask=1 to count cycles.
477.It Li OFFCORE_REQUESTS_OUTSTANDING.DEMAND_RFO
478.Pq Event 60H , Umask 04H
479Offcore outstanding RFO store transactions in SQ to
480uncore.
481Set Cmask=1 to count cycles.
482.It Li OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD
483.Pq Event 60H , Umask 08H
484Offcore outstanding cacheable data read
485transactions in SQ to uncore.
486Set Cmask=1 to count cycles.
487.It Li LOCK_CYCLES.SPLIT_LOCK_UC_LOCK_DURATION
488.Pq Event 63H , Umask 01H
489Cycles in which the L1D and L2 are locked, due to a
490UC lock or split lock.
491.It Li LOCK_CYCLES.CACHE_LOCK_DURATION
492.Pq Event 63H , Umask 02H
493Cycles in which the L1D is locked.
494.It Li IDQ.EMPTY
495.Pq Event 79H , Umask 02H
496Counts cycles the IDQ is empty.
497.It Li IDQ.MITE_UOPS
498.Pq Event 79H , Umask 04H
499Increment each cycle # of uops delivered to IDQ
500from MITE path.
501Set Cmask = 1 to count cycles.
502.It Li IDQ.DSB_UOPS
503.Pq Event 79H , Umask 08H
504Increment each cycle. # of uops delivered to IDQ
505from DSB path.
506Set Cmask = 1 to count cycles.
507.It Li IDQ.MS_DSB_UOPS
508.Pq Event 79H , Umask 10H
509Increment each cycle # of uops delivered to IDQ
510when MS busy by DSB.
511Set Cmask = 1 to count cycles MS is busy.
512Set Cmask=1 and Edge =1 to count MS activations.
513.It Li IDQ.MS_MITE_UOPS
514.Pq Event 79H , Umask 20H
515Increment each cycle # of uops delivered to IDQ
516when MS is busy by MITE.
517Set Cmask = 1 to count cycles.
518.It Li IDQ.MS_UOPS
519.Pq Event 79H , Umask 30H
520Increment each cycle # of uops delivered to IDQ
521from MS by either DSB or MITE.
522Set Cmask = 1 to count cycles.
523.It Li ICACHE.MISSES
524.Pq Event 80H , Umask 02H
525Number of Instruction Cache, Streaming Buffer and
526Victim Cache Misses.
527Includes UC accesses.
528.It Li ITLB_MISSES.MISS_CAUSES_A_WALK
529.Pq Event 85H , Umask 01H
530Misses in all ITLB levels that cause page walks.
531.It Li ITLB_MISSES.WALK_COMPLETED
532.Pq Event 85H , Umask 02H
533Misses in all ITLB levels that cause completed page
534walks.
535.It Li ITLB_MISSES.WALK_DURATION
536.Pq Event 85H , Umask 04H
537Cycle PMH is busy with a walk.
538.It Li ITLB_MISSES.STLB_HIT
539.Pq Event 85H , Umask 10H
540Number of cache load STLB hits.
541No page walk.
542.It Li ILD_STALL.LCP
543.Pq Event 87H , Umask 01H
544Stalls caused by changing prefix length of the
545instruction.
546.It Li ILD_STALL.IQ_FULL
547.Pq Event 87H , Umask 04H
548Stall cycles due to IQ is full.
549.It Li BR_INST_EXEC.NONTAKEN_COND
550.Pq Event 88H , Umask 41H
551Count conditional near branch instructions that were executed (but not
552necessarily retired) and not taken.
553.It Li BR_INST_EXEC.TAKEN_COND
554.Pq Event 88H , Umask 81H
555Count conditional near branch instructions that were executed (but not
556necessarily retired) and taken.
557.It Li BR_INST_EXEC.DIRECT_JMP
558.Pq Event 88H , Umask 82H
559Count all unconditional near branch instructions excluding calls and
560indirect branches.
561.It Li BR_INST_EXEC.INDIRECT_JMP_NON_CALL_RET
562.Pq Event 88H , Umask 84H
563Count executed indirect near branch instructions that are not calls nor
564returns.
565.It Li BR_INST_EXEC.RETURN_NEAR
566.Pq Event 88H , Umask 88H
567Count indirect near branches that have a return mnemonic.
568.It Li BR_INST_EXEC.DIRECT_NEAR_CALL
569.Pq Event 88H , Umask 90H
570Count unconditional near call branch instructions, excluding non call
571branch, executed.
572.It Li BR_INST_EXEC.INDIRECT_NEAR_CALL
573.Pq Event 88H , Umask A0H
574Count indirect near calls, including both register and memory indirect,
575executed.
576.It Li BR_INST_EXEC.ALL_BRANCHES
577.Pq Event 88H , Umask FFH
578Counts all near executed branches (not necessarily retired).
579.It Li BR_MISP_EXEC.NONTAKEN_COND
580.Pq Event 89H , Umask 41H
581Count conditional near branch instructions mispredicted as nontaken.
582.It Li BR_MISP_EXEC.TAKEN_COND
583.Pq Event 89H , Umask 81H
584Count conditional near branch instructions mispredicted as taken.
585.It Li BR_MISP_EXEC.INDIRECT_JMP_NON_CALL_RET
586.Pq Event 89H , Umask 84H
587Count mispredicted indirect near branch instructions that are not calls
588nor returns.
589.It Li BR_MISP_EXEC.RETURN_NEAR
590.Pq Event 89H , Umask 88H
591Count mispredicted indirect near branches that have a return mnemonic.
592.It Li BR_MISP_EXEC.DIRECT_NEAR_CALL
593.Pq Event 89H , Umask 90H
594Count mispredicted unconditional near call branch instructions, excluding
595non call branch, executed.
596.It Li BR_MISP_EXEC.INDIRECT_NEAR_CALL
597.Pq Event 89H , Umask A0H
598Count mispredicted indirect near calls, including both register and memory
599indirect, executed.
600.It Li BR_MISP_EXEC.ALL_BRANCHES
601.Pq Event 89H , Umask FFH
602Counts all mispredicted near executed branches (not necessarily retired).
603.It Li IDQ_UOPS_NOT_DELIVERED.CORE
604.Pq Event 9CH , Umask 01H
605Count number of non-delivered uops to RAT per
606thread.
607.It Li UOPS_DISPATCHED_PORT.PORT_0
608.Pq Event A1H , Umask 01H
609Cycles which a Uop is dispatched on port 0.
610.It Li UOPS_DISPATCHED_PORT.PORT_1
611.Pq Event A1H , Umask 02H
612Cycles which a Uop is dispatched on port 1.
613.It Li UOPS_DISPATCHED_PORT.PORT_2_LD
614.Pq Event A1H , Umask 04H
615Cycles which a load uop is dispatched on port 2.
616.It Li UOPS_DISPATCHED_PORT.PORT_2_STA
617.Pq Event A1H , Umask 08H
618Cycles which a store address uop is dispatched on
619port 2.
620.It Li UOPS_DISPATCHED_PORT.PORT_2
621.Pq Event A1H , Umask 0CH
622Cycles which a Uop is dispatched on port 2.
623.It Li UOPS_DISPATCHED_PORT.PORT_3_LD
624.Pq Event A1H , Umask 10H
625Cycles which a load uop is dispatched on port 3.
626.It Li UOPS_DISPATCHED_PORT.PORT_3_STA
627.Pq Event A1H , Umask 20H
628Cycles which a store address uop is dispatched on
629port 3.
630.It Li UOPS_DISPATCHED_PORT.PORT_3
631.Pq Event A1H , Umask 30H
632Cycles which a Uop is dispatched on port 3.
633.It Li UOPS_DISPATCHED_PORT.PORT_4
634.Pq Event A1H , Umask 40H
635Cycles which a Uop is dispatched on port 4.
636.It Li UOPS_DISPATCHED_PORT.PORT_5
637.Pq Event A1H , Umask 80H
638Cycles which a Uop is dispatched on port 5.
639.It Li RESOURCE_STALLS.ANY
640.Pq Event A2H , Umask 01H
641Cycles Allocation is stalled due to Resource Related
642reason.
643.It Li RESOURCE_STALLS.LB
644.Pq Event A2H , Umask 01H
645Counts the cycles of stall due to lack of load buffers.
646.It Li RESOURCE_STALLS.RS
647.Pq Event A2H , Umask 04H
648Cycles stalled due to no eligible RS entry available.
649.It Li RESOURCE_STALLS.SB
650.Pq Event A2H , Umask 08H
651Cycles stalled due to no store buffers available.
652(not including draining form sync).
653.It Li RESOURCE_STALLS.ROB
654.Pq Event A2H , Umask 10H
655Cycles stalled due to re-order buffer full.
656.It Li RESOURCE_STALLS.FCSW
657.Pq Event A2H , Umask 20H
658Cycles stalled due to writing the FPU control word.
659.It Li RESOURCE_STALLS.MXCSR
660.Pq Event A2H , Umask 40H
661Cycles stalled due to the MXCSR register rename
662occurring to close to a previous MXCSR rename.
663.It Li RESOURCE_STALLS.OTHER
664.Pq Event A2H , Umask 80H
665Cycles stalled while execution was stalled due to
666other resource issues.
667.It Li CYCLE_ACTIVITY.CYCLES_L2_PENDING
668.Pq Event A3H , Umask 01H
669Cycles with pending L2 miss loads.
670Set AnyThread to count per core.
671.It Li CYCLE_ACTIVITY.CYCLES_L1D_PENDING
672.Pq Event A3H , Umask 02H
673Cycles with pending L1 cache miss loads.
674Set AnyThread to count per core.
675.It Li CYCLE_ACTIVITY.CYCLES_NO_DISPATCH
676.Pq Event A3H , Umask 04H
677Cycles of dispatch stalls.
678Set AnyThread to count per core.
679.It Li DSB2MITE_SWITCHES.COUNT
680.Pq Event ABH , Umask 01H
681Number of DSB to MITE switches.
682.It Li DSB2MITE_SWITCHES.PENALTY_CYCLES
683.Pq Event ABH , Umask 02H
684Cycles DSB to MITE switches caused delay.
685.It Li DSB_FILL.OTHER_CANCEL
686.Pq Event ACH , Umask 02H
687Cases of cancelling valid DSB fill not because of
688exceeding way limit.
689.It Li DSB_FILL.EXCEED_DSB_LINES
690.Pq Event ACH , Umask 08H
691DSB Fill encountered > 3 DSB lines.
692.It Li DSB_FILL.ALL_CANCEL
693.Pq Event ACH , Umask 0AH
694Cases of cancelling valid Decode Stream Buffer
695(DSB) fill not because of exceeding way limit.
696.It Li ITLB.ITLB_FLUSH
697.Pq Event AEH , Umask 01H
698Counts the number of ITLB flushes, includes
6994k/2M/4M pages.
700.It Li OFFCORE_REQUESTS.DEMAND_DATA_RD
701.Pq Event B0H , Umask 01H
702Demand data read requests sent to uncore.
703.It Li OFFCORE_REQUESTS.DEMAND_RFO
704.Pq Event B0H , Umask 04H
705Demand RFO read requests sent to uncore, including
706regular RFOs, locks, ItoM.
707.It Li OFFCORE_REQUESTS.ALL_DATA_RD
708.Pq Event B0H , Umask 08H
709Data read requests sent to uncore (demand and
710prefetch).
711.It Li UOPS_DISPATCHED.THREAD
712.Pq Event B1H , Umask 01H
713Counts total number of uops to be dispatched per-
714thread each cycle.
715Set Cmask = 1, INV =1 to count stall cycles.
716.It Li UOPS_DISPATCHED.CORE
717.Pq Event B1H , Umask 02H
718Counts total number of uops to be dispatched per-
719core each cycle.
720.It Li OFFCORE_REQUESTS_BUFFER.SQ_FULL
721.Pq Event B2H , Umask 01H
722Offcore requests buffer cannot take more entries
723for this thread core.
724.It Li AGU_BYPASS_CANCEL.COUNT
725.Pq Event B6H , Umask 01H
726Counts executed load operations with all the
727following traits: 1. addressing of the format [base +
728offset], 2. the offset is between 1 and 2047, 3. the
729address specified in the base register is in one page
730and the address [base+offset] is in another page.
731.It Li OFF_CORE_RESPONSE_0
732.Pq Event B7H , Umask 01H
733(Event B7H, Umask 01H) Off-core Response Performance
734Monitoring; PMC0 only.
735Requires programming MSR 01A6H
736.It Li OFF_CORE_RESPONSE_1
737.Pq Event BBH , Umask 01H
738(Event BBH, Umask 01H) Off-core Response Performance
739Monitoring; PMC3 only.
740Requires programming MSR 01A7H
741.It Li TLB_FLUSH.DTLB_THREAD
742.Pq Event BDH , Umask 01H
743DTLB flush attempts of the thread-specific entries.
744.It Li TLB_FLUSH.STLB_ANY
745.Pq Event BDH , Umask 20H
746Count number of STLB flush attempts.
747.It Li L1D_BLOCKS.BANK_CONFLICT_CYCLES
748.Pq Event BFH , Umask 05H
749Cycles when dispatched loads are cancelled due to
750L1D bank conflicts with other load ports.
751.It Li INST_RETIRED.ANY_P
752.Pq Event C0H , Umask 00H
753Number of instructions at retirement.
754.It Li INST_RETIRED.ALL
755.Pq Event C0H , Umask 01H
756Precise instruction retired event with HW to reduce
757effect of PEBS shadow in IP distribution.
758.It Li OTHER_ASSISTS.ITLB_MISS_RETIRED
759.Pq Event C1H , Umask 02H
760Instructions that experienced an ITLB miss.
761.It Li OTHER_ASSISTS.AVX_STORE
762.Pq Event C1H , Umask 08H
763Number of assists associated with 256-bit AVX
764store operations.
765.It Li OTHER_ASSISTS.AVX_TO_SSE
766.Pq Event C1H , Umask 10H
767Number of transitions from AVX-256 to legacy SSE
768when penalty applicable.
769.It Li OTHER_ASSISTS.SSE_TO_AVX
770.Pq Event C1H , Umask 20H
771Number of transitions from SSE to AVX-256 when
772penalty applicable.
773.It Li UOPS_RETIRED.ALL
774.Pq Event C2H , Umask 01H
775Counts the number of micro-ops retired, Use
776cmask=1 and invert to count active cycles or stalled
777cycles.
778.It Li UOPS_RETIRED.RETIRE_SLOTS
779.Pq Event C2H , Umask 02H
780Counts the number of retirement slots used each
781cycle.
782.It Li MACHINE_CLEARS.MEMORY_ORDERING
783.Pq Event C3H , Umask 02H
784Counts the number of machine clears due to
785memory order conflicts.
786.It Li MACHINE_CLEARS.SMC
787.Pq Event C3H , Umask 04H
788Counts the number of times that a program writes
789to a code section.
790.It Li MACHINE_CLEARS.MASKMOV
791.Pq Event C3H , Umask 20H
792Counts the number of executed AVX masked load
793operations that refer to an illegal address range
794with the mask bits set to 0.
795.It Li BR_INST_RETIRED.ALL_BRANCH
796.Pq Event C4H , Umask 00H
797Branch instructions at retirement.
798.It Li BR_INST_RETIRED.CONDITIONAL
799.Pq Event C4H , Umask 01H
800Counts the number of conditional branch
801instructions retired.
802.It Li BR_INST_RETIRED.NEAR_CALL
803.Pq Event C4H , Umask 02H
804Direct and indirect near call instructions retired.
805.It Li BR_INST_RETIRED.ALL_BRANCHES
806.Pq Event C4H , Umask 04H
807Counts the number of branch instructions retired.
808.It Li BR_INST_RETIRED.NEAR_RETURN
809.Pq Event C4H , Umask 08H
810Counts the number of near return instructions
811retired.
812.It Li BR_INST_RETIRED.NOT_TAKEN
813.Pq Event C4H , Umask 10H
814Counts the number of not taken branch instructions
815retired.
816.It Li BR_INST_RETIRED.NEAR_TAKEN
817.Pq Event C4H , Umask 20H
818Number of near taken branches retired.
819.It Li BR_INST_RETIRED.FAR_BRANCH
820.Pq Event C4H , Umask 40H
821Number of far branches retired.
822.It Li BR_MISP_RETIRED.ALL_BRANCHES
823.Pq Event C5H , Umask 00H
824Mispredicted branch instructions at retirement.
825.It Li BR_MISP_RETIRED.CONDITIONAL
826.Pq Event C5H , Umask 01H
827Mispredicted conditional branch instructions retired.
828.It Li BR_MISP_RETIRED.NEAR_CALL
829.Pq Event C5H , Umask 02H
830Direct and indirect mispredicted near call
831instructions retired.
832.It Li BR_MISP_RETIRED.ALL_BRANCHES
833.Pq Event C5H , Umask 04H
834Mispredicted macro branch instructions retired.
835.It Li BR_MISP_RETIRED.NOT_TAKEN
836.Pq Event C5H , Umask 10H
837Mispredicted not taken branch instructions retired.
838.It Li BR_MISP_RETIRED.TAKEN
839.Pq Event C5H , Umask 20H
840Mispredicted taken branch instructions retired.
841.It Li FP_ASSIST.X87_OUTPUT
842.Pq Event CAH , Umask 02H
843Number of X87 assists due to output value.
844.It Li FP_ASSIST.X87_INPUT
845.Pq Event CAH , Umask 04H
846Number of X87 assists due to input value.
847.It Li FP_ASSIST.SIMD_OUTPUT
848.Pq Event CAH , Umask 08H
849 Number of SIMD FP assists due to output values.
850.It Li FP_ASSIST.SIMD_INPUT
851.Pq Event CAH , Umask 10H
852Number of SIMD FP assists due to input values.
853.It Li FP_ASSIST.ANY 1EH
854.Pq Event CAH , Umask
855Cycles with any input/output SSE* or FP assists.
856.It Li ROB_MISC_EVENTS.LBR_INSERTS
857.Pq Event CCH , Umask 20H
858Count cases of saving new LBR records by
859hardware.
860.It Li MEM_TRANS_RETIRED.LOAD_LATENCY
861.Pq Event CDH , Umask 01H
862Sample loads with specified latency threshold.
863PMC3 only.
864.It Li MEM_TRANS_RETIRED.PRECISE_STORE
865.Pq Event CDH , Umask 02H
866Sample stores and collect precise store operation
867via PEBS record.
868PMC3 only.
869.It Li MEM_UOP_RETIRED.LOADS
870.Pq Event D0H , Umask 10H
871Qualify retired memory uops that are loads.
872Combine with umask 10H, 20H, 40H, 80H.
873.It Li MEM_UOP_RETIRED.STORES
874.Pq Event D0H , Umask 02H
875Qualify retired memory uops that are stores.
876Combine with umask 10H, 20H, 40H, 80H.
877.It Li MEM_UOP_RETIRED.STLB_MISS
878.Pq Event D0H , Umask
879Qualify retired memory uops with STLB miss.
880Must combine with umask 01H, 02H, to produce counts.
881.It Li MEM_UOP_RETIRED.LOCK
882.Pq Event D0H , Umask
883Qualify retired memory uops with lock.
884Must combine with umask 01H, 02H, to produce counts.
885.It Li MEM_UOP_RETIRED.SPLIT
886.Pq Event D0H , Umask
887Qualify retired memory uops with line split.
888Must combine with umask 01H, 02H, to produce counts.
889.It Li MEM_UOP_RETIRED_ALL
890.Pq Event D0H , Umask
891Qualify any retired memory uops.
892Must combine with umask 01H, 02H, to produce counts.
893.It Li MEM_LOAD_UOPS_RETIRED.L1_HIT
894.Pq Event D1H , Umask 01H
895Retired load uops with L1 cache hits as data
896sources.
897.It Li MEM_LOAD_UOPS_RETIRED.L2_HIT
898.Pq Event D1H , Umask 02H
899Retired load uops with L2 cache hits as data
900sources.
901.It Li MEM_LOAD_UOPS_RETIRED.LLC_HIT
902.Pq Event D1H , Umask 04H
903Retired load uops which data sources were data hits
904in LLC without snoops required.
905.It Li MEM_LOAD_UOPS_RETIRED.LLC_MISS
906.Pq Event D1H , Umask 20H
907Retired load uops which data sources were data
908missed LLC (excluding unknown data source).
909.It Li MEM_LOAD_UOPS_RETIRED.HIT_LFB
910.Pq Event D1H , Umask 40H
911Retired load uops which data sources were load
912uops missed L1 but hit FB due to preceding miss to
913the same cache line with data not ready.
914.It Li MEM_LOAD_UOPS_MISC_RETIRED.LLC_MISS
915.Pq Event D4H , Umask 02H
916Retired load uops with unknown information as data
917source in cache serviced the load.
918.It Li BACLEARS.ANY
919.Pq Event E6H , Umask 01H
920Counts the number of times the front end is re-
921steered, mainly when the BPU cannot provide a
922correct prediction and this is corrected by other
923branch handling mechanisms at the front end.
924.It Li L2_TRANS.DEMAND_DATA_RD
925.Pq Event F0H , Umask 01H
926Demand Data Read requests that access L2 cache.
927.It Li L2_TRANS.RFO
928.Pq Event F0H , Umask 02H
929RFO requests that access L2 cache.
930.It Li L2_TRANS.CODE_RD
931.Pq Event F0H , Umask 04H
932L2 cache accesses when fetching instructions.
933.It Li L2_TRANS.ALL_PF
934.Pq Event F0H , Umask 08H
935L2 or LLC HW prefetches that access L2 cache.
936.It Li L2_TRANS.L1D_WB
937.Pq Event F0H , Umask 10H
938L1D writebacks that access L2 cache.
939.It Li L2_TRANS.L2_FILL
940.Pq Event F0H , Umask 20H
941L2 fill requests that access L2 cache.
942.It Li L2_TRANS.L2_WB
943.Pq Event F0H , Umask 40H
944L2 writebacks that access L2 cache.
945.It Li L2_TRANS.ALL_REQUESTS
946.Pq Event F0H , Umask 80H
947Transactions accessing L2 pipe.
948.It Li L2_LINES_IN.I
949.Pq Event F1H , Umask 01H
950L2 cache lines in I state filling L2.
951.It Li L2_LINES_IN.S
952.Pq Event F1H , Umask 02H
953L2 cache lines in S state filling L2.
954.It Li L2_LINES_IN.E
955.Pq Event F1H , Umask 04H
956L2 cache lines in E state filling L2.
957.It Li L2_LINES-IN.ALL
958.Pq Event F1H , Umask 07H
959L2 cache lines filling L2.
960.It Li L2_LINES_OUT.DEMAND_CLEAN
961.Pq Event F2H , Umask 01H
962Clean L2 cache lines evicted by demand.
963.It Li L2_LINES_OUT.DEMAND_DIRTY
964.Pq Event F2H , Umask 02H
965Dirty L2 cache lines evicted by demand.
966.It Li L2_LINES_OUT.PF_CLEAN
967.Pq Event F2H , Umask 04H
968Clean L2 cache lines evicted by L2 prefetch.
969.It Li L2_LINES_OUT.PF_DIRTY
970.Pq Event F2H , Umask 08H
971Dirty L2 cache lines evicted by L2 prefetch.
972.It Li L2_LINES_OUT.DIRTY_ALL
973.Pq Event F2H , Umask 0AH
974Dirty L2 cache lines filling the L2.
975.It Li SQ_MISC.SPLIT_LOCK
976.Pq Event F4H , Umask 10H
977Split locks in SQ.
978.El
979.Sh SEE ALSO
980.Xr pmc 3 ,
981.Xr pmc.amd 3 ,
982.Xr pmc.atom 3 ,
983.Xr pmc.core 3 ,
984.Xr pmc.corei7 3 ,
985.Xr pmc.corei7uc 3 ,
986.Xr pmc.haswelluc 3 ,
987.Xr pmc.iaf 3 ,
988.Xr pmc.ivybridge 3 ,
989.Xr pmc.ivybridgexeon 3 ,
990.Xr pmc.sandybridge 3 ,
991.Xr pmc.sandybridgeuc 3 ,
992.Xr pmc.soft 3 ,
993.Xr pmc.tsc 3 ,
994.Xr pmc.ucf 3 ,
995.Xr pmc.westmere 3 ,
996.Xr pmc.westmereuc 3 ,
997.Xr pmc_cpuinfo 3 ,
998.Xr pmclog 3 ,
999.Xr hwpmc 4
1000.Sh HISTORY
1001The
1002.Nm pmc
1003library first appeared in
1004.Fx 6.0 .
1005.Sh AUTHORS
1006.An -nosplit
1007The
1008.Lb libpmc
1009library was written by
1010.An Joseph Koshy Aq Mt jkoshy@FreeBSD.org .
1011The support for the Sandy Bridge Xeon
1012microarchitecture was written by
1013.An Hiren Panchasara Aq Mt hiren.panchasara@gmail.com .
1014