xref: /freebsd/lib/libpmc/pmc.ivybridge.3 (revision d37eb51047221dc3322b34db1038ff3aa533883f)
1.\" Copyright (c) 2012 Fabien Thomas.  All rights reserved.
2.\"
3.\" Redistribution and use in source and binary forms, with or without
4.\" modification, are permitted provided that the following conditions
5.\" are met:
6.\" 1. Redistributions of source code must retain the above copyright
7.\"    notice, this list of conditions and the following disclaimer.
8.\" 2. Redistributions in binary form must reproduce the above copyright
9.\"    notice, this list of conditions and the following disclaimer in the
10.\"    documentation and/or other materials provided with the distribution.
11.\"
12.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
13.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
14.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
15.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
16.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
17.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
18.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
19.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
20.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
21.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
22.\" SUCH DAMAGE.
23.\"
24.Dd October 19, 2012
25.Dt PMC.IVYBRIDGE 3
26.Os
27.Sh NAME
28.Nm pmc.ivybridge
29.Nd measurement events for
30.Tn Intel
31.Tn Ivy Bridge
32family CPUs
33.Sh LIBRARY
34.Lb libpmc
35.Sh SYNOPSIS
36.In pmc.h
37.Sh DESCRIPTION
38.Tn Intel
39.Tn "Ivy Bridge"
40CPUs contain PMCs conforming to version 2 of the
41.Tn Intel
42performance measurement architecture.
43These CPUs may contain up to three classes of PMCs:
44.Bl -tag -width "Li PMC_CLASS_IAP"
45.It Li PMC_CLASS_IAF
46Fixed-function counters that count only one hardware event per counter.
47.It Li PMC_CLASS_IAP
48Programmable counters that may be configured to count one of a defined
49set of hardware events.
50.El
51.Pp
52The number of PMCs available in each class and their widths need to be
53determined at run time by calling
54.Xr pmc_cpuinfo 3 .
55.Pp
56Intel Ivy Bridge PMCs are documented in
57.Rs
58.%B "Intel(R) 64 and IA-32 Architectures Software Developer's Manual"
59.%T "Volume 3B: System Programming Guide, Part 2"
60.%N "Order Number: 253669-043US"
61.%D May 2012
62.%Q "Intel Corporation"
63.Re
64.Ss IVYBRIDGE FIXED FUNCTION PMCS
65These PMCs and their supported events are documented in
66.Xr pmc.iaf 3 .
67.Ss IVYBRIDGE PROGRAMMABLE PMCS
68The programmable PMCs support the following capabilities:
69.Bl -column "PMC_CAP_INTERRUPT" "Support"
70.It Em Capability Ta Em Support
71.It PMC_CAP_CASCADE Ta \&No
72.It PMC_CAP_EDGE Ta Yes
73.It PMC_CAP_INTERRUPT Ta Yes
74.It PMC_CAP_INVERT Ta Yes
75.It PMC_CAP_READ Ta Yes
76.It PMC_CAP_PRECISE Ta \&No
77.It PMC_CAP_SYSTEM Ta Yes
78.It PMC_CAP_TAGGING Ta \&No
79.It PMC_CAP_THRESHOLD Ta Yes
80.It PMC_CAP_USER Ta Yes
81.It PMC_CAP_WRITE Ta Yes
82.El
83.Ss Event Qualifiers
84Event specifiers for these PMCs support the following common
85qualifiers:
86.Bl -tag -width indent
87.It Li rsp= Ns Ar value
88Configure the Off-core Response bits.
89.Bl -tag -width indent
90.It Li REQ_DMND_DATA_RD
91Counts the number of demand and DCU prefetch data reads of full and partial
92cachelines as well as demand data page table entry cacheline reads.
93Does not count L2 data read prefetches or instruction fetches.
94.It Li REQ_DMND_RFO
95Counts the number of demand and DCU prefetch reads for ownership (RFO)
96requests generated by a write to data cacheline.
97Does not count L2 RFO prefetches.
98.It Li REQ_DMND_IFETCH
99Counts the number of demand and DCU prefetch instruction cacheline reads.
100Does not count L2 code read prefetches.
101.It Li REQ_WB
102Counts the number of writeback (modified to exclusive) transactions.
103.It Li REQ_PF_DATA_RD
104Counts the number of data cacheline reads generated by L2 prefetchers.
105.It Li REQ_PF_RFO
106Counts the number of RFO requests generated by L2 prefetchers.
107.It Li REQ_PF_IFETCH
108Counts the number of code reads generated by L2 prefetchers.
109.It Li REQ_PF_LLC_DATA_RD
110L2 prefetcher to L3 for loads.
111.It Li REQ_PF_LLC_RFO
112RFO requests generated by L2 prefetcher
113.It Li REQ_PF_LLC_IFETCH
114L2 prefetcher to L3 for instruction fetches.
115.It Li REQ_BUS_LOCKS
116Bus lock and split lock requests.
117.It Li REQ_STRM_ST
118Streaming store requests.
119.It Li REQ_OTHER
120Any other request that crosses IDI, including I/O.
121.It Li RES_ANY
122Catch all value for any response types.
123.It Li RES_SUPPLIER_NO_SUPP
124No Supplier Information available.
125.It Li RES_SUPPLIER_LLC_HITM
126M-state initial lookup stat in L3.
127.It Li RES_SUPPLIER_LLC_HITE
128E-state.
129.It Li RES_SUPPLIER_LLC_HITS
130S-state.
131.It Li RES_SUPPLIER_LLC_HITF
132F-state.
133.It Li RES_SUPPLIER_LOCAL
134Local DRAM Controller.
135.It Li RES_SNOOP_SNP_NONE
136No details on snoop-related information.
137.It Li RES_SNOOP_SNP_NO_NEEDED
138No snoop was needed to satisfy the request.
139.It Li RES_SNOOP_SNP_MISS
140A snoop was needed and it missed all snooped caches:
141-For LLC Hit, ReslHitl was returned by all cores
142-For LLC Miss, Rspl was returned by all sockets and data was returned from
143DRAM.
144.It Li RES_SNOOP_HIT_NO_FWD
145A snoop was needed and it hits in at least one snooped cache.
146Hit denotes a cache-line was valid before snoop effect.
147This includes:
148-Snoop Hit w/ Invalidation (LLC Hit, RFO)
149-Snoop Hit, Left Shared (LLC Hit/Miss, IFetch/Data_RD)
150-Snoop Hit w/ Invalidation and No Forward (LLC Miss, RFO Hit S)
151In the LLC Miss case, data is returned from DRAM.
152.It Li RES_SNOOP_HIT_FWD
153A snoop was needed and data was forwarded from a remote socket.
154This includes:
155-Snoop Forward Clean, Left Shared (LLC Hit/Miss, IFetch/Data_RD/RFT).
156.It Li RES_SNOOP_HITM
157A snoop was needed and it HitM-ed in local or remote cache.
158HitM denotes a cache-line was in modified state before effect as a results of snoop.
159This includes:
160-Snoop HitM w/ WB (LLC miss, IFetch/Data_RD)
161-Snoop Forward Modified w/ Invalidation (LLC Hit/Miss, RFO)
162-Snoop MtoS (LLC Hit, IFetch/Data_RD).
163.It Li RES_NON_DRAM
164Target was non-DRAM system address.
165This includes MMIO transactions.
166.El
167.It Li cmask= Ns Ar value
168Configure the PMC to increment only if the number of configured
169events measured in a cycle is greater than or equal to
170.Ar value .
171.It Li edge
172Configure the PMC to count the number of de-asserted to asserted
173transitions of the conditions expressed by the other qualifiers.
174If specified, the counter will increment only once whenever a
175condition becomes true, irrespective of the number of clocks during
176which the condition remains true.
177.It Li inv
178Invert the sense of comparison when the
179.Dq Li cmask
180qualifier is present, making the counter increment when the number of
181events per cycle is less than the value specified by the
182.Dq Li cmask
183qualifier.
184.It Li os
185Configure the PMC to count events happening at processor privilege
186level 0.
187.It Li usr
188Configure the PMC to count events occurring at privilege levels 1, 2
189or 3.
190.El
191.Pp
192If neither of the
193.Dq Li os
194or
195.Dq Li usr
196qualifiers are specified, the default is to enable both.
197.Ss Event Specifiers (Programmable PMCs)
198Ivy Bridge programmable PMCs support the following events:
199.Bl -tag -width indent
200.It Li LD_BLOCKS.STORE_FORWARD
201.Pq Event 03H , Umask 02H
202loads blocked by overlapping with store buffer that cannot be forwarded .
203.It Li MISALIGN_MEM_REF.LOADS
204.Pq Event 05H , Umask 01H
205Speculative cache-line split load uops dispatched to L1D.
206.It Li MISALIGN_MEM_REF.STORES
207.Pq Event 05H , Umask 02H
208Speculative cache-line split Store- address uops dispatched to L1D.
209.It Li LD_BLOCKS_PARTIAL.ADDRESS_ALIAS
210.Pq Event 07H , Umask 01H
211False dependencies in MOB due to partial compare on address.
212.It Li DTLB_LOAD_MISSES.DEMAND_LD_MISS_CAUSES_A_WALK
213.Pq Event 08H , Umask 81H
214Misses in all TLB levels that cause a page walk of any page size from demand loads.
215.It Li DTLB_LOAD_MISSES.DEMAND_LD_WALK_COMPLETED
216.Pq Event 08H , Umask 82H
217Misses in all TLB levels that caused page walk completed of any size by demand loads.
218.It Li DTLB_LOAD_MISSES.DEMAND_LD_WALK_DURATION
219.Pq Event 08H , Umask 84H
220Cycle PMH is busy with a walk due to demand loads.
221.It Li UOPS_ISSUED.ANY
222.Pq Event 0EH , Umask 01H
223Increments each cycle the # of Uops issued by the RAT to RS.
224Set Cmask = 1, Inv = 1to count stalled cycles.
225Set Cmask = 1, Inv = 1, Any= 1to count stalled cycles of this core.
226.It Li UOPS_ISSUED.FLAGS_MERGE
227.Pq Event 0EH , Umask 10H
228Number of flags-merge uops allocated.
229Such uops adds delay.
230.It Li UOPS_ISSUED.SLOW_LEA
231.Pq Event 0EH , Umask 20H
232Number of slow LEA or similar uops allocated.
233Such uop has 3 sources (e.g. 2 sources + immediate) regardless if as a result of LEA instruction or not.
234.It Li UOPS_ISSUED.SINGLE_MUL
235.Pq Event 0EH , Umask 40H
236Number of multiply packed/scalar single precision uops allocated.
237.It Li ARITH.FPU_DIV_ACTIVE
238.Pq Event 14H , Umask 01H
239Cycles that the divider is active, includes INT and FP.
240Set 'edge =1, cmask=1' to count the number of divides.
241.It Li L2_RQSTS.DEMAND_DATA_RD_HIT
242.Pq Event 24H , Umask 01H
243Demand Data Read requests that hit L2 cache.
244.It Li L2_RQSTS.ALL_DEMAND_DATA_RD
245.Pq Event 24H , Umask 03H
246Counts any demand and L1 HW prefetch data load requests to L2.
247.It Li L2_RQSTS.RFO_HITS
248.Pq Event 24H , Umask 04H
249Counts the number of store RFO requests that hit the L2 cache.
250.It Li L2_RQSTS.RFO_MISS
251.Pq Event 24H , Umask 08H
252Counts the number of store RFO requests that miss the L2 cache.
253.It Li L2_RQSTS.ALL_RFO
254.Pq Event 24H , Umask 0CH
255Counts all L2 store RFO requests.
256.It Li L2_RQSTS.CODE_RD_HIT
257.Pq Event 24H , Umask 10H
258Number of instruction fetches that hit the L2 cache.
259.It Li L2_RQSTS.CODE_RD_MISS
260.Pq Event 24H , Umask 20H
261Number of instruction fetches that missed the L2 cache.
262.It Li L2_RQSTS.ALL_CODE_RD
263.Pq Event 24H , Umask 30H
264Counts all L2 code requests.
265.It Li L2_RQSTS.PF_HIT
266.Pq Event 24H , Umask 40H
267Counts all L2 HW prefetcher requests that hit L2.
268.It Li L2_RQSTS.PF_MISS
269.Pq Event 24H , Umask 80H
270Counts all L2 HW prefetcher requests that missed L2.
271.It Li L2_RQSTS.ALL_PF
272.Pq Event 24H , Umask C0H
273Counts all L2 HW prefetcher requests.
274.It Li L2_STORE_LOCK_RQSTS.MISS
275.Pq Event 27H , Umask 01H
276RFOs that miss cache lines.
277.It Li L2_STORE_LOCK_RQSTS.HIT_M
278.Pq Event 27H , Umask 08H
279RFOs that hit cache lines in M state.
280.It Li L2_STORE_LOCK_RQSTS.ALL
281.Pq Event 27H , Umask 0FH
282RFOs that access cache lines in any state.
283.It Li L2_L1D_WB_RQSTS.MISS
284.Pq Event 28H , Umask 01H
285Not rejected writebacks that missed LLC.
286.It Li L2_L1D_WB_RQSTS.HIT_E
287.Pq Event 28H , Umask 04H
288Not rejected writebacks from L1D to L2 cache lines in E state.
289.It Li L2_L1D_WB_RQSTS.HIT_M
290.Pq Event 28H , Umask 08H
291Not rejected writebacks from L1D to L2 cache lines in M state.
292.It Li L2_L1D_WB_RQSTS.ALL
293.Pq Event 28H , Umask 0FH
294Not rejected writebacks from L1D to L2 cache lines in any state.
295.It Li LONGEST_LAT_CACHE.REFERENCE
296.Pq Event 2EH , Umask 4FH
297This event counts requests originating from the core that reference a cache
298line in the last level cache.
299.It Li LONGEST_LAT_CACHE.MISS
300.Pq Event 2EH , Umask 41H
301This event counts each cache miss condition for references to the last level
302cache.
303.It Li CPU_CLK_UNHALTED.THREAD_P
304.Pq Event 3CH , Umask 00H
305Counts the number of thread cycles while the thread is not in a halt state.
306The thread enters the halt state when it is running the HLT instruction.
307The core frequency may change from time to time due to power or thermal
308throttling.
309.It Li CPU_CLK_THREAD_UNHALTED.REF_XCLK
310.Pq Event 3CH , Umask 01H
311Increments at the frequency of XCLK (100 MHz) when not halted.
312.It Li L1D_PEND_MISS.PENDING
313.Pq Event 48H , Umask 01H
314Increments the number of outstanding L1D misses every cycle.
315Set Cmaks = 1 and Edge =1 to count occurrences.
316Counter 2 only.
317Set Cmask = 1 to count cycles.
318.It Li DTLB_STORE_MISSES.MISS_CAUSES_A_WALK
319.Pq Event 49H , Umask 01H
320Miss in all TLB levels causes an page walk of any page size (4K/2M/4M/1G).
321.It Li DTLB_STORE_MISSES.WALK_COMPLETED
322.Pq Event 49H , Umask 02H
323Miss in all TLB levels causes a page walk that completes of any page size
324(4K/2M/4M/1G).
325.It Li DTLB_STORE_MISSES.WALK_DURATION
326.Pq Event 49H , Umask 04H
327Cycles PMH is busy with this walk.
328.It Li DTLB_STORE_MISSES.STLB_HIT
329.Pq Event 49H , Umask 10H
330Store operations that miss the first TLB level but hit the second and do not
331cause page walks.
332.It Li LOAD_HIT_PRE.SW_PF
333.Pq Event 4CH , Umask 01H
334Non-SW-prefetch load dispatches that hit fill buffer allocated for S/W prefetch.
335.It Li LOAD_HIT_PRE.HW_PF
336.Pq Event 4CH , Umask 02H
337Non-SW-prefetch load dispatches that hit fill buffer allocated for H/W prefetch.
338.It Li L1D.REPLACEMENT
339.Pq Event 51H , Umask 01H
340Counts the number of lines brought into the L1 data cache.
341.It Li MOVE_ELIMINATION.INT_NOT_ELIMINATED
342.Pq Event 58H , Umask 01H
343Number of integer Move Elimination candidate uops that were not eliminated.
344.It Li MOVE_ELIMINATION.SIMD_NOT_ELIMINATED
345.Pq Event 58H , Umask 02H
346Number of SIMD Move Elimination candidate uops that were not eliminated.
347.It Li MOVE_ELIMINATION.INT_ELIMINATED
348.Pq Event 58H , Umask 04H
349Number of integer Move Elimination candidate uops that were eliminated.
350.It Li MOVE_ELIMINATION.SIMD_ELIMINATED
351.Pq Event 58H , Umask 08H
352Number of SIMD Move Elimination candidate uops that were eliminated.
353.It Li CPL_CYCLES.RING0
354.Pq Event 5CH , Umask 01H
355Unhalted core cycles when the thread is in ring 0.
356Use Edge to count transition.
357.It Li CPL_CYCLES.RING123
358.Pq Event 5CH , Umask 02H
359Unhalted core cycles when the thread is not in ring 0.
360.It Li RS_EVENTS.EMPTY_CYCLES
361.Pq Event 5EH , Umask 01H
362Cycles the RS is empty for the thread.
363.It Li TLB_ACCESS.LOAD_STLB_HIT
364.Pq Event 5FH , Umask 01H
365Counts load operations that missed 1st level DTLB but hit the 2nd level.
366.It Li OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD
367.Pq Event 60H , Umask 01H
368Offcore outstanding Demand Data Read transactions in SQ to uncore.
369Set Cmask=1 to count cycles.
370.It Li OFFCORE_REQUESTS_OUTSTANDING.DEMAND_CODE_RD
371.Pq Event 60H , Umask 02H
372Offcore outstanding Demand Code Read transactions in SQ to uncore.
373Set Cmask=1 to count cycles.
374.It Li OFFCORE_REQUESTS_OUTSTANDING.DEMAND_RFO
375.Pq Event 60H , Umask 04H
376Offcore outstanding RFO store transactions in SQ to uncore.
377Set Cmask=1 to count cycles.
378.It Li OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD
379.Pq Event 60H , Umask 08H
380Offcore outstanding cacheable data read transactions in SQ to uncore.
381Set Cmask=1 to count cycles.
382.It Li LOCK_CYCLES.SPLIT_LOCK_UC_LOCK_DURATION
383.Pq Event 63H , Umask 01H
384Cycles in which the L1D and L2 are locked, due to a UC lock or split lock.
385.It Li LOCK_CYCLES.CACHE_LOCK_DURATION
386.Pq Event 63H , Umask 02H
387Cycles in which the L1D is locked.
388.It Li IDQ.EMPTY
389.Pq Event 79H , Umask 02H
390Counts cycles the IDQ is empty.
391.It Li IDQ.MITE_UOPS
392.Pq Event 79H , Umask 04H
393Increment each cycle # of uops delivered to IDQ from MITE path.
394Can combine Umask 04H and 20H.
395Set Cmask = 1 to count cycles.
396.It Li IDQ.DSB_UOPS
397.Pq Event 79H , Umask 08H
398Increment each cycle. # of uops delivered to IDQ from DSB path.
399Can combine Umask 08H and 10H
400Set Cmask = 1 to count cycles.
401.It Li IDQ.MS_DSB_UOPS
402.Pq Event 79H , Umask 10H
403Increment each cycle # of uops delivered to IDQ when MS_busy by DSB.
404Set Cmask = 1 to count cycles.
405Add Edge=1 to count # of delivery.
406Can combine Umask 04H, 08H.
407.It Li IDQ.MS_MITE_UOPS
408.Pq Event 79H , Umask 20H
409Increment each cycle # of uops delivered to IDQ when MS_busy by MITE.
410Set Cmask = 1 to count cycles.
411Can combine Umask 04H, 08H.
412.It Li IDQ.MS_UOPS
413.Pq Event 79H , Umask 30H
414Increment each cycle # of uops delivered to IDQ from MS by either DSB or
415MITE.
416Set Cmask = 1 to count cycles.
417Can combine Umask 04H, 08H.
418.It Li IDQ.ALL_DSB_CYCLES_ANY_UOPS
419.Pq Event 79H , Umask 18H
420Counts cycles DSB is delivered at least one uops.
421Set Cmask = 1.
422.It Li IDQ.ALL_DSB_CYCLES_4_UOPS
423.Pq Event 79H , Umask 18H
424Counts cycles DSB is delivered four uops.
425Set Cmask = 4.
426.It Li IDQ.ALL_MITE_CYCLES_ANY_UOPS
427.Pq Event 79H , Umask 24H
428Counts cycles MITE is delivered at least one uops.
429Set Cmask = 1.
430.It Li IDQ.ALL_MITE_CYCLES_4_UOPS
431.Pq Event 79H , Umask 24H
432Counts cycles MITE is delivered four uops.
433Set Cmask = 4.
434.It Li IDQ.MITE_ALL_UOPS
435.Pq Event 79H , Umask 3CH
436# of uops delivered to IDQ from any path.
437.It Li ICACHE.MISSES
438.Pq Event 80H , Umask 02H
439Number of Instruction Cache, Streaming Buffer and Victim Cache Misses.
440Includes UC accesses.
441.It Li ITLB_MISSES.MISS_CAUSES_A_WALK
442.Pq Event 85H , Umask 01H
443Misses in all ITLB levels that cause page walks.
444.It Li ITLB_MISSES.WALK_COMPLETED
445.Pq Event 85H , Umask 02H
446Misses in all ITLB levels that cause completed page walks.
447.It Li ITLB_MISSES.WALK_DURATION
448.Pq Event 85H , Umask 04H
449Cycle PMH is busy with a walk.
450.It Li ITLB_MISSES.STLB_HIT
451.Pq Event 85H , Umask 10H
452Number of cache load STLB hits.
453No page walk.
454.It Li ILD_STALL.LCP
455.Pq Event 87H , Umask 01H
456Stalls caused by changing prefix length of the instruction.
457.It Li ILD_STALL.IQ_FULL
458.Pq Event 87H , Umask 04H
459Stall cycles due to IQ is full.
460.It Li BR_INST_EXEC.NONTAKEN_COND
461.Pq Event 88H , Umask 41H
462Count conditional near branch instructions that were executed (but not
463necessarily retired) and not taken.
464.It Li BR_INST_EXEC.TAKEN_COND
465.Pq Event 88H , Umask 81H
466Count conditional near branch instructions that were executed (but not
467necessarily retired) and taken.
468.It Li BR_INST_EXEC.DIRECT_JMP
469.Pq Event 88H , Umask 82H
470Count all unconditional near branch instructions excluding calls and
471indirect branches.
472.It Li BR_INST_EXEC.INDIRECT_JMP_NON_CALL_RET
473.Pq Event 88H , Umask 84H
474Count executed indirect near branch instructions that are not calls nor
475returns.
476.It Li BR_INST_EXEC.RETURN_NEAR
477.Pq Event 88H , Umask 88H
478Count indirect near branches that have a return mnemonic.
479.It Li BR_INST_EXEC.DIRECT_NEAR_CALL
480.Pq Event 88H , Umask 90H
481Count unconditional near call branch instructions, excluding non call
482branch, executed.
483.It Li BR_INST_EXEC.INDIRECT_NEAR_CALL
484.Pq Event 88H , Umask A0H
485Count indirect near calls, including both register and memory indirect,
486executed.
487.It Li BR_INST_EXEC.ALL_BRANCHES
488.Pq Event 88H , Umask FFH
489Counts all near executed branches (not necessarily retired).
490.It Li BR_MISP_EXEC.NONTAKEN_COND
491.Pq Event 89H , Umask 41H
492Count conditional near branch instructions mispredicted as nontaken.
493.It Li BR_MISP_EXEC.TAKEN_COND
494.Pq Event 89H , Umask 81H
495Count conditional near branch instructions mispredicted as taken.
496.It Li BR_MISP_EXEC.INDIRECT_JMP_NON_CALL_RET
497.Pq Event 89H , Umask 84H
498Count mispredicted indirect near branch instructions that are not calls
499nor returns.
500.It Li BR_MISP_EXEC.RETURN_NEAR
501.Pq Event 89H , Umask 88H
502Count mispredicted indirect near branches that have a return mnemonic.
503.It Li BR_MISP_EXEC.DIRECT_NEAR_CALL
504.Pq Event 89H , Umask 90H
505Count mispredicted unconditional near call branch instructions, excluding
506non call branch, executed.
507.It Li BR_MISP_EXEC.INDIRECT_NEAR_CALL
508.Pq Event 89H , Umask A0H
509Count mispredicted indirect near calls, including both register and memory
510indirect, executed.
511.It Li BR_MISP_EXEC.ALL_BRANCHES
512.Pq Event 89H , Umask FFH
513Counts all mispredicted near executed branches (not necessarily retired).
514.It Li IDQ_UOPS_NOT_DELIVERED.CORE
515.Pq Event 9CH , Umask 01H
516Count number of non-delivered uops to RAT per thread.
517Use Cmask to qualify uop b/w.
518.It Li UOPS_DISPATCHED_PORT.PORT_0
519.Pq Event A1H , Umask 01H
520Cycles which a Uop is dispatched on port 0.
521.It Li UOPS_DISPATCHED_PORT.PORT_1
522.Pq Event A1H , Umask 02H
523Cycles which a Uop is dispatched on port 1.
524.It Li UOPS_DISPATCHED_PORT.PORT_2_LD
525.Pq Event A1H , Umask 04H
526Cycles which a load uop is dispatched on port 2.
527.It Li UOPS_DISPATCHED_PORT.PORT_2_STA
528.Pq Event A1H , Umask 08H
529Cycles which a store address uop is dispatched on port 2.
530.It Li UOPS_DISPATCHED_PORT.PORT_2
531.Pq Event A1H , Umask 0CH
532Cycles which a Uop is dispatched on port 2.
533.It Li UOPS_DISPATCHED_PORT.PORT_3_LD
534.Pq Event A1H , Umask 10H
535Cycles which a load uop is dispatched on port 3.
536.It Li UOPS_DISPATCHED_PORT.PORT_3_STA
537.Pq Event A1H , Umask 20H
538Cycles which a store address uop is dispatched on port 3.
539.It Li UOPS_DISPATCHED_PORT.PORT_3
540.Pq Event A1H , Umask 30H
541Cycles which a Uop is dispatched on port 3.
542.It Li UOPS_DISPATCHED_PORT.PORT_4
543.Pq Event A1H , Umask 40H
544Cycles which a Uop is dispatched on port 4.
545.It Li UOPS_DISPATCHED_PORT.PORT_5
546.Pq Event A1H , Umask 80H
547Cycles which a Uop is dispatched on port 5.
548.It Li RESOURCE_STALLS.ANY
549.Pq Event A2H , Umask 01H
550Cycles Allocation is stalled due to Resource Related reason.
551.It Li RESOURCE_STALLS.RS
552.Pq Event A2H , Umask 04H
553Cycles stalled due to no eligible RS entry available.
554.It Li RESOURCE_STALLS.SB
555.Pq Event A2H , Umask 08H
556Cycles stalled due to no store buffers available. (not including draining
557form sync).
558.It Li RESOURCE_STALLS.ROB
559.Pq Event A2H , Umask 10H
560Cycles stalled due to re-order buffer full.
561.It Li DSB2MITE_SWITCHES.COUNT
562.Pq Event ABH , Umask 01H
563Number of DSB to MITE switches.
564.It Li DSB2MITE_SWITCHES.PENALTY_CYCLES
565.Pq Event ABH , Umask 02H
566Cycles DSB to MITE switches caused delay.
567.It Li DSB_FILL.EXCEED_DSB_LINES
568.Pq Event ACH , Umask 08H
569DSB Fill encountered > 3 DSB lines.
570.It Li ITLB.ITLB_FLUSH
571.Pq Event AEH , Umask 01H
572Counts the number of ITLB flushes, includes 4k/2M/4M pages.
573.It Li OFFCORE_REQUESTS.DEMAND_DATA_RD
574.Pq Event B0H , Umask 01H
575Demand data read requests sent to uncore.
576.It Li OFFCORE_REQUESTS.DEMAND_CODE_RD
577.Pq Event B0H , Umask 02H
578Demand code read requests sent to uncore.
579.It Li OFFCORE_REQUESTS.DEMAND_RFO
580.Pq Event B0H , Umask 04H
581Demand RFO read requests sent to uncore, including regular RFOs, locks,
582ItoM.
583.It Li OFFCORE_REQUESTS.ALL_DATA_RD
584.Pq Event B0H , Umask 08H
585Data read requests sent to uncore (demand and prefetch).
586.It Li UOPS_EXECUTED.THREAD
587.Pq Event B1H , Umask 01H
588Counts total number of uops to be executed per-thread each cycle.
589Set Cmask = 1, INV =1 to count stall cycles.
590.It Li UOPS_EXECUTED.CORE
591.Pq Event B1H , Umask 02H
592Counts total number of uops to be executed per-core each cycle.
593Do not need to set ANY.
594.It Li OFF_CORE_RESPONSE_0
595.Pq Event B7H , Umask 01H
596Off-core Response Performance Monitoring.
597PMC0 only.
598Requires programming MSR 01A6H.
599.It Li OFF_CORE_RESPONSE_1
600.Pq Event BBH , Umask 01H
601Off-core Response Performance Monitoring.
602PMC3 only.
603Requires programming MSR 01A7H.
604.It Li TLB_FLUSH.DTLB_THREAD
605.Pq Event BDH , Umask 01H
606DTLB flush attempts of the thread- specific entries.
607.It Li TLB_FLUSH.STLB_ANY
608.Pq Event BDH , Umask 20H
609Count number of STLB flush attempts.
610.It Li INST_RETIRED.ANY_P
611.Pq Event C0H , Umask 00H
612Number of instructions at retirement.
613.It Li INST_RETIRED.ALL
614.Pq Event C0H , Umask 01H
615Precise instruction retired event with HW to reduce effect of PEBS shadow in
616IP distribution.
617PMC1 only.
618Must quiesce other PMCs.
619.It Li OTHER_ASSISTS.AVX_STORE
620.Pq Event C1H , Umask 08H
621Number of assists associated with 256-bit AVX store operations.
622.It Li OTHER_ASSISTS.AVX_TO_SSE
623.Pq Event C1H , Umask 10H
624Number of transitions from AVX- 256 to legacy SSE when penalty applicable.
625.It Li OTHER_ASSISTS.SSE_TO_AVX
626.Pq Event C1H , Umask 20H
627Number of transitions from SSE to AVX-256 when penalty applicable.
628.It Li UOPS_RETIRED.ALL
629.Pq Event C2H , Umask 01H
630Counts the number of micro-ops retired, Use cmask=1 and invert to count
631active cycles or stalled cycles.
632Supports PEBS, use Any=1 for core granular.
633.It Li UOPS_RETIRED.RETIRE_SLOTS
634.Pq Event C2H , Umask 02H
635Counts the number of retirement slots used each cycle.
636.It Li MACHINE_CLEARS.MEMORY_ORDERING
637.Pq Event C3H , Umask 02H
638Counts the number of machine clears due to memory order conflicts.
639.It Li MACHINE_CLEARS.SMC
640.Pq Event C3H , Umask 04H
641Number of self-modifying-code machine clears detected.
642.It Li MACHINE_CLEARS.MASKMOV
643.Pq Event C3H , Umask 20H
644Counts the number of executed AVX masked load operations that refer to an
645illegal address range with the mask bits set to 0.
646.It Li BR_INST_RETIRED.ALL_BRANCHES
647.Pq Event C4H , Umask 00H
648Branch instructions at retirement.
649.It Li BR_INST_RETIRED.CONDITIONAL
650.Pq Event C4H , Umask 01H
651Counts the number of conditional branch instructions retired.
652Supports PEBS.
653.It Li BR_INST_RETIRED.NEAR_CALL
654.Pq Event C4H , Umask 02H
655Direct and indirect near call instructions retired.
656.It Li BR_INST_RETIRED.ALL_BRANCHES
657.Pq Event C4H , Umask 04H
658Counts the number of branch instructions retired.
659.It Li BR_INST_RETIRED.NEAR_RETURN
660.Pq Event C4H , Umask 08H
661Counts the number of near return instructions retired.
662.It Li BR_INST_RETIRED.NOT_TAKEN
663.Pq Event C4H , Umask 10H
664Counts the number of not taken branch instructions retired.
665.It Li BR_INST_RETIRED.NEAR_TAKEN
666.Pq Event C4H , Umask 20H
667Number of near taken branches retired.
668.It Li BR_INST_RETIRED.FAR_BRANCH
669.Pq Event C4H , Umask 40H
670Number of far branches retired.
671.It Li BR_MISP_RETIRED.ALL_BRANCHES
672.Pq Event C5H , Umask 00H
673Mispredicted branch instructions at retirement.
674.It Li BR_MISP_RETIRED.CONDITIONAL
675.Pq Event C5H , Umask 01H
676Mispredicted conditional branch instructions retired.
677Supports PEBS.
678.It Li BR_MISP_RETIRED.NEAR_CALL
679.Pq Event C5H , Umask 02H
680Direct and indirect mispredicted near call instructions retired.
681.It Li BR_MISP_RETIRED.ALL_BRANCHES
682.Pq Event C5H , Umask 04H
683Mispredicted macro branch instructions retired.
684.It Li BR_MISP_RETIRED.NOT_TAKEN
685.Pq Event C5H , Umask 10H
686Mispredicted not taken branch instructions retired.
687.It Li BR_MISP_RETIRED.TAKEN
688.Pq Event C5H , Umask 20H
689Mispredicted taken branch instructions retired.
690.It Li FP_ASSIST.X87_OUTPUT
691.Pq Event CAH , Umask 02H
692Number of X87 FP assists due to Output values.
693.It Li FP_ASSIST.X87_INPUT
694.Pq Event CAH , Umask 04H
695Number of X87 FP assists due to input values.
696.It Li FP_ASSIST.SIMD_OUTPUT
697.Pq Event CAH , Umask 08H
698Number of SIMD FP assists due to Output values.
699.It Li FP_ASSIST.SIMD_INPUT
700.Pq Event CAH , Umask 10H
701Number of SIMD FP assists due to input values.
702.It Li FP_ASSIST.ANY
703.Pq Event CAH , Umask 1EH
704Cycles with any input/output SSE* or FP assists.
705.It Li ROB_MISC_EVENTS.LBR_INSERTS
706.Pq Event CCH , Umask 20H
707Count cases of saving new LBR records by hardware.
708.It Li MEM_TRANS_RETIRED.LOAD_LATENCY
709.Pq Event CDH , Umask 01H
710Sample loads with specified latency threshold.
711PMC3 only.
712Specify threshold in MSR 0x3F6.
713.It Li MEM_TRANS_RETIRED.PRECISE_STORE
714.Pq Event CDH , Umask 02H
715Sample stores and collect precise store operation via PEBS record.
716PMC3 only.
717.It Li MEM_UOPS_RETIRED.STLB_MISS_LOADS
718.Pq Event D0H , Umask 11H
719Count retired load uops that missed the STLB.
720.It Li MEM_UOPS_RETIRED.STLB_MISS_STORES
721.Pq Event D0H , Umask 12H
722Count retired store uops that missed the STLB.
723.It Li MEM_UOPS_RETIRED.SPLIT_LOADS
724.Pq Event D0H , Umask 41H
725Count retired load uops that were split across a cache line.
726.It Li MEM_UOPS_RETIRED.SPLIT_STORES
727.Pq Event D0H , Umask 42H
728Count retired store uops that were split across a cache line.
729.It Li MEM_UOPS_RETIRED.ALL_LOADS
730.Pq Event D0H , Umask 81H
731Count all retired load uops.
732.It Li MEM_UOPS_RETIRED.ALL_STORES
733.Pq Event D0H , Umask 82H
734Count all retired store uops.
735.It Li MEM_LOAD_UOPS_RETIRED.L1_HIT
736.Pq Event D1H , Umask 01H
737Retired load uops with L1 cache hits as data sources.
738Supports PEBS.
739.It Li MEM_LOAD_UOPS_RETIRED.L2_HIT
740.Pq Event D1H , Umask 02H
741Retired load uops with L2 cache hits as data sources.
742.It Li MEM_LOAD_UOPS_RETIRED.LLC_HIT
743.Pq Event D1H , Umask 04H
744Retired load uops with LLC cache hits as data sources.
745.It Li MEM_LOAD_UOPS_RETIRED.HIT_LFB
746.Pq Event D1H , Umask 40H
747Retired load uops which data sources were load uops missed L1 but hit FB due
748to preceding miss to the same cache line with data not ready.
749.It Li MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_MISS
750.Pq Event D2H , Umask 01H
751Retired load uops which data sources were LLC hit and cross-core snoop
752missed in on-pkg core cache.
753Supports PEBS.
754.It Li MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_HIT
755.Pq Event D2H , Umask 02H
756Retired load uops which data sources were LLC and cross-core snoop hits in
757on-pkg core cache.
758Supports PEBS.
759.It Li MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_HITM
760.Pq Event D2H , Umask 04H
761Retired load uops which data sources were HitM responses from shared LLC.
762.It Li MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_NONE
763.Pq Event D2H , Umask 08H
764Retired load uops which data sources were hits in LLC without snoops
765required.
766.It Li MEM_LOAD_UOPS_LLC_MISS_RETIRED.LOCAL_DRAM
767.Pq Event D3H , Umask 01H
768Retired load uops which data sources missed LLC but serviced from local
769dram.
770Supports PEBS.
771.It Li L2_TRANS.DEMAND_DATA_RD
772.Pq Event F0H , Umask 01H
773Demand Data Read requests that access L2 cache.
774.It Li L2_TRANS.RFO
775.Pq Event F0H , Umask 02H
776RFO requests that access L2 cache.
777.It Li L2_TRANS.CODE_RD
778.Pq Event F0H , Umask 04H
779L2 cache accesses when fetching instructions.
780.It Li L2_TRANS.ALL_PF
781.Pq Event F0H , Umask 08H
782Any MLC or LLC HW prefetch accessing L2, including rejects.
783.It Li L2_TRANS.L1D_WB
784.Pq Event F0H , Umask 10H
785L1D writebacks that access L2 cache.
786.It Li L2_TRANS.L2_FILL
787.Pq Event F0H , Umask 20H
788L2 fill requests that access L2 cache.
789.It Li L2_TRANS.L2_WB
790.Pq Event F0H , Umask 40H
791L2 writebacks that access L2 cache.
792.It Li L2_TRANS.ALL_REQUESTS
793.Pq Event F0H , Umask 80H
794Transactions accessing L2 pipe.
795.It Li L2_LINES_IN.I
796.Pq Event F1H , Umask 01H
797L2 cache lines in I state filling L2.
798Counting does not cover rejects.
799.It Li L2_LINES_IN.S
800.Pq Event F1H , Umask 02H
801L2 cache lines in S state filling L2.
802Counting does not cover rejects.
803.It Li L2_LINES_IN.E
804.Pq Event F1H , Umask 04H
805L2 cache lines in E state filling L2.
806Counting does not cover rejects.
807.It Li L2_LINES_IN.ALL
808.Pq Event F1H , Umask 07H
809L2 cache lines filling L2.
810Counting does not cover rejects.
811.It Li L2_LINES_OUT.DEMAND_CLEAN
812.Pq Event F2H , Umask 01H
813Clean L2 cache lines evicted by demand.
814.It Li L2_LINES_OUT.DEMAND_DIRTY
815.Pq Event F2H , Umask 02H
816Dirty L2 cache lines evicted by demand.
817.It Li L2_LINES_OUT.PF_CLEAN
818.Pq Event F2H , Umask 04H
819Clean L2 cache lines evicted by the MLC prefetcher.
820.It Li L2_LINES_OUT.PF_DIRTY
821.Pq Event F2H , Umask 08H
822Dirty L2 cache lines evicted by the MLC prefetcher.
823.El
824.Sh SEE ALSO
825.Xr pmc 3 ,
826.Xr pmc.amd 3 ,
827.Xr pmc.atom 3 ,
828.Xr pmc.core 3 ,
829.Xr pmc.corei7 3 ,
830.Xr pmc.corei7uc 3 ,
831.Xr pmc.iaf 3 ,
832.Xr pmc.ivybridgexeon 3 ,
833.Xr pmc.sandybridge 3 ,
834.Xr pmc.sandybridgeuc 3 ,
835.Xr pmc.sandybridgexeon 3 ,
836.Xr pmc.soft 3 ,
837.Xr pmc.tsc 3 ,
838.Xr pmc.ucf 3 ,
839.Xr pmc.westmere 3 ,
840.Xr pmc.westmereuc 3 ,
841.Xr pmc_cpuinfo 3 ,
842.Xr pmclog 3 ,
843.Xr hwpmc 4
844.Sh HISTORY
845The
846.Nm pmc
847library first appeared in
848.Fx 6.0 .
849.Sh AUTHORS
850.An -nosplit
851The
852.Lb libpmc
853library was written by
854.An Joseph Koshy Aq Mt jkoshy@FreeBSD.org .
855The support for the Ivy Bridge
856microarchitecture was written by
857.An Fabien Thomas Aq Mt fabient@FreeBSD.org .
858