xref: /freebsd/lib/libpmc/pmc.westmereuc.3 (revision 315ee00f)
1.\" Copyright (c) 2010 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 March 24, 2010
25.Dt PMC.WESTMEREUC 3
26.Os
27.Sh NAME
28.Nm pmc.westmere
29.Nd uncore measurement events for
30.Tn Intel
31.Tn Westmere
32family CPUs
33.Sh LIBRARY
34.Lb libpmc
35.Sh SYNOPSIS
36.In pmc.h
37.Sh DESCRIPTION
38.Tn Intel
39.Tn "Westmere"
40CPUs contain PMCs conforming to version 2 of the
41.Tn Intel
42performance measurement architecture.
43These CPUs contain two classes of PMCs:
44.Bl -tag -width "Li PMC_CLASS_UCP"
45.It Li PMC_CLASS_UCF
46Fixed-function counters that count only one hardware event per counter.
47.It Li PMC_CLASS_UCP
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 Westmere PMCs are documented in
57.Rs
58.%B "Intel(R) 64 and IA-32 Architectures Software Developes Manual"
59.%T "Volume 3B: System Programming Guide, Part 2"
60.%N "Order Number: 253669-033US"
61.%D December 2009
62.%Q "Intel Corporation"
63.Re
64.Ss WESTMERE UNCORE FIXED FUNCTION PMCS
65These PMCs and their supported events are documented in
66.Xr pmc.ucf 3 .
67Not all CPUs in this family implement fixed-function counters.
68.Ss WESTMERE UNCORE 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 \&No
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 \&No
79.It PMC_CAP_TAGGING Ta \&No
80.It PMC_CAP_THRESHOLD Ta Yes
81.It PMC_CAP_USER Ta \&No
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 cmask= Ns Ar value
89Configure the PMC to increment only if the number of configured
90events measured in a cycle is greater than or equal to
91.Ar value .
92.It Li edge
93Configure the PMC to count the number of de-asserted to asserted
94transitions of the conditions expressed by the other qualifiers.
95If specified, the counter will increment only once whenever a
96condition becomes true, irrespective of the number of clocks during
97which the condition remains true.
98.It Li inv
99Invert the sense of comparison when the
100.Dq Li cmask
101qualifier is present, making the counter increment when the number of
102events per cycle is less than the value specified by the
103.Dq Li cmask
104qualifier.
105.El
106.Ss Event Specifiers (Programmable PMCs)
107Westmere uncore programmable PMCs support the following events:
108.Bl -tag -width indent
109.It Li GQ_CYCLES_FULL.READ_TRACKER
110.Pq Event 00H , Umask 01H
111Uncore cycles Global Queue read tracker is full.
112.It Li GQ_CYCLES_FULL.WRITE_TRACKER
113.Pq Event 00H , Umask 02H
114Uncore cycles Global Queue write tracker is full.
115.It Li GQ_CYCLES_FULL.PEER_PROBE_TRACKER
116.Pq Event 00H , Umask 04H
117Uncore cycles Global Queue peer probe tracker is full.
118The peer probe tracker queue tracks snoops from the IOH and remote sockets.
119.It Li GQ_CYCLES_NOT_EMPTY.READ_TRACKER
120.Pq Event 01H , Umask 01H
121Uncore cycles were Global Queue read tracker has at least one valid entry.
122.It Li GQ_CYCLES_NOT_EMPTY.WRITE_TRACKER
123.Pq Event 01H , Umask 02H
124Uncore cycles were Global Queue write tracker has at least one valid entry.
125.It Li GQ_CYCLES_NOT_EMPTY.PEER_PROBE_TRACKER
126.Pq Event 01H , Umask 04H
127Uncore cycles were Global Queue peer probe tracker has at least one valid entry.
128The peer probe tracker queue tracks IOH and remote socket snoops.
129.It Li GQ_OCCUPANCY.READ_TRACKER
130.Pq Event 02H , Umask 01H
131Increments the number of queue entries (code read, data read, and RFOs) in
132the tread tracker.
133The GQ read tracker allocate to deallocate occupancy count is divided by the
134count to obtain the average read tracker latency.
135.It Li GQ_ALLOC.READ_TRACKER
136.Pq Event 03H , Umask 01H
137Counts the number of tread tracker allocate to deallocate entries.
138The GQ read tracker allocate to deallocate occupancy count is divided by
139the count to obtain the average read tracker latency.
140.It Li GQ_ALLOC.RT_L3_MISS
141.Pq Event 03H , Umask 02H
142Counts the number GQ read tracker entries for which a full cache line read
143has missed the L3.
144The GQ read tracker L3 miss to fill occupancy count is divided by this count
145to obtain the average cache line read L3 miss latency.
146The latency represents the time after which the L3 has determined that the
147cache line has missed.
148The time between a GQ read tracker allocation and the L3 determining that
149the cache line has missed is the average L3 hit latency.
150The total L3 cache line read miss latency is the hit latency + L3 miss
151latency.
152.It Li GQ_ALLOC.RT_TO_L3_RESP
153.Pq Event 03H , Umask 04H
154Counts the number of GQ read tracker entries that are allocated in the read
155tracker queue that hit or miss the L3.
156The GQ read tracker L3 hit occupancy count is divided by this count to obtain the average L3 hit latency.
157.It Li GQ_ALLOC.RT_TO_RTID_ACQUIRED
158.Pq Event 03H , Umask 08H
159Counts the number of GQ read tracker entries that are allocated in the read
160tracker, have missed in the L3 and have not acquired a Request Transaction ID.
161The GQ read tracker L3 miss to RTID acquired occupancy count is
162divided by this count to obtain the average latency for a read L3 miss to
163acquire an RTID.
164.It Li GQ_ALLOC.WT_TO_RTID_ACQUIRED
165.Pq Event 03H , Umask 10H
166Counts the number of GQ write tracker entries that are allocated in the
167write tracker, have missed in the L3 and have not acquired a Request
168Transaction ID.
169The GQ write tracker L3 miss to RTID occupancy count is divided by this count
170to obtain the average latency for a write L3 miss to acquire an RTID.
171.It Li GQ_ALLOC.WRITE_TRACKER
172.Pq Event 03H , Umask 20H
173Counts the number of GQ write tracker entries that are allocated in the write
174tracker queue that miss the L3.
175The GQ write tracker occupancy count
176is divided by the this count to obtain the average L3 write miss latency.
177.It Li GQ_ALLOC.PEER_PROBE_TRACKER
178.Pq Event 03H , Umask 40H
179Counts the number of GQ peer probe tracker (snoop) entries that are
180allocated in the peer probe tracker queue that miss the L3.
181The GQ peer probe occupancy count is divided by this count to obtain the average
182L3 peer probe miss latency.
183.It Li GQ_DATA.FROM_QPI
184.Pq Event 04H , Umask 01H
185Cycles Global Queue Quickpath Interface input data port is busy importing
186data from the Quickpath Interface.
187Each cycle the input port can transfer 8 or 16 bytes of data.
188.It Li GQ_DATA.FROM_QMC
189.Pq Event 04H , Umask 02H
190Cycles Global Queue Quickpath Memory Interface input data port is busy
191importing data from the Quickpath Memory Interface.
192Each cycle the input port can transfer 8 or 16 bytes of data.
193.It Li GQ_DATA.FROM_L3
194.Pq Event 04H , Umask 04H
195Cycles GQ L3 input data port is busy importing data from the Last Level Cache.
196Each cycle the input port can transfer 32 bytes of data.
197.It Li GQ_DATA.FROM_CORES_02
198.Pq Event 04H , Umask 08H
199Cycles GQ Core 0 and 2 input data port is busy importing data from processor
200cores 0 and 2.
201Each cycle the input port can transfer 32 bytes of data.
202.It Li GQ_DATA.FROM_CORES_13
203.Pq Event 04H , Umask 10H
204Cycles GQ Core 1 and 3 input data port is busy importing data from processor
205cores 1 and 3.
206Each cycle the input port can transfer 32 bytes of data.
207.It Li GQ_DATA.TO_QPI_QMC
208.Pq Event 05H , Umask 01H
209Cycles GQ QPI and QMC output data port is busy sending data to the Quickpath
210Interface or Quickpath Memory Interface.
211Each cycle the output port can transfer 32 bytes of data.
212.It Li GQ_DATA.TO_L3
213.Pq Event 05H , Umask 02H
214Cycles GQ L3 output data port is busy sending data to the Last Level Cache.
215Each cycle the output port can transfer 32 bytes of data.
216.It Li GQ_DATA.TO_CORES
217.Pq Event 05H , Umask 04H
218Cycles GQ Core output data port is busy sending data to the Cores.
219Each cycle the output port can transfer 32 bytes of data.
220.It Li SNP_RESP_TO_LOCAL_HOME.I_STATE
221.Pq Event 06H , Umask 01H
222Number of snoop responses to the local home that L3 does not have the
223referenced cache line.
224.It Li SNP_RESP_TO_LOCAL_HOME.S_STATE
225.Pq Event 06H , Umask 02H
226Number of snoop responses to the local home that L3 has the referenced line
227cached in the S state.
228.It Li SNP_RESP_TO_LOCAL_HOME.FWD_S_STATE
229.Pq Event 06H , Umask 04H
230Number of responses to code or data read snoops to the local home that the
231L3 has the referenced cache line in the E state.
232The L3 cache line state is changed to the S state and the line is forwarded
233to the local home in the S state.
234.It Li SNP_RESP_TO_LOCAL_HOME.FWD_I_STATE
235.Pq Event 06H , Umask 08H
236Number of responses to read invalidate snoops to the local home that the L3
237has the referenced cache line in the M state.
238The L3 cache line state is invalidated and the line is forwarded to the
239local home in the M state.
240.It Li SNP_RESP_TO_LOCAL_HOME.CONFLICT
241.Pq Event 06H , Umask 10H
242Number of conflict snoop responses sent to the local home.
243.It Li SNP_RESP_TO_LOCAL_HOME.WB
244.Pq Event 06H , Umask 20H
245Number of responses to code or data read snoops to the local home that the
246L3 has the referenced line cached in the M state.
247.It Li SNP_RESP_TO_REMOTE_HOME.I_STATE
248.Pq Event 07H , Umask 01H
249Number of snoop responses to a remote home that L3 does not have the
250referenced cache line.
251.It Li SNP_RESP_TO_REMOTE_HOME.S_STATE
252.Pq Event 07H , Umask 02H
253Number of snoop responses to a remote home that L3 has the referenced line
254cached in the S state.
255.It Li SNP_RESP_TO_REMOTE_HOME.FWD_S_STATE
256.Pq Event 07H , Umask 04H
257Number of responses to code or data read snoops to a remote home that the L3
258has the referenced cache line in the E state.
259The L3 cache line state is changed to the S state and the line is forwarded
260to the remote home in the S state.
261.It Li SNP_RESP_TO_REMOTE_HOME.FWD_I_STATE
262.Pq Event 07H , Umask 08H
263Number of responses to read invalidate snoops to a remote home that the L3
264has the referenced cache line in the M state.
265The L3 cache line state is invalidated and the line is forwarded to the
266remote home in the M state.
267.It Li SNP_RESP_TO_REMOTE_HOME.CONFLICT
268.Pq Event 07H , Umask 10H
269Number of conflict snoop responses sent to the local home.
270.It Li SNP_RESP_TO_REMOTE_HOME.WB
271.Pq Event 07H , Umask 20H
272Number of responses to code or data read snoops to a remote home that the L3
273has the referenced line cached in the M state.
274.It Li SNP_RESP_TO_REMOTE_HOME.HITM
275.Pq Event 07H , Umask 24H
276Number of HITM snoop responses to a remote home.
277.It Li L3_HITS.READ
278.Pq Event 08H , Umask 01H
279Number of code read, data read and RFO requests that hit in the L3.
280.It Li L3_HITS.WRITE
281.Pq Event 08H , Umask 02H
282Number of writeback requests that hit in the L3.
283Writebacks from the cores will always result in L3 hits due to the
284inclusive property of the L3.
285.It Li L3_HITS.PROBE
286.Pq Event 08H , Umask 04H
287Number of snoops from IOH or remote sockets that hit in the L3.
288.It Li L3_HITS.ANY
289.Pq Event 08H , Umask 03H
290Number of reads and writes that hit the L3.
291.It Li L3_MISS.READ
292.Pq Event 09H , Umask 01H
293Number of code read, data read and RFO requests that miss the L3.
294.It Li L3_MISS.WRITE
295.Pq Event 09H , Umask 02H
296Number of writeback requests that miss the L3.
297Should always be zero as writebacks from the cores will always result in L3 hits due to the inclusive
298property of the L3.
299.It Li L3_MISS.PROBE
300.Pq Event 09H , Umask 04H
301Number of snoops from IOH or remote sockets that miss the L3.
302.It Li L3_MISS.ANY
303.Pq Event 09H , Umask 03H
304Number of reads and writes that miss the L3.
305.It Li L3_LINES_IN.M_STATE
306.Pq Event 0AH , Umask 01H
307Counts the number of L3 lines allocated in M state.
308The only time a cache line is allocated in the M state is when the
309line was forwarded in M state is forwarded due to a Snoop Read Invalidate Own request.
310.It Li L3_LINES_IN.E_STATE
311.Pq Event 0AH , Umask 02H
312Counts the number of L3 lines allocated in E state.
313.It Li L3_LINES_IN.S_STATE
314.Pq Event 0AH , Umask 04H
315Counts the number of L3 lines allocated in S state.
316.It Li L3_LINES_IN.F_STATE
317.Pq Event 0AH , Umask 08H
318Counts the number of L3 lines allocated in F state.
319.It Li L3_LINES_IN.ANY
320.Pq Event 0AH , Umask 0FH
321Counts the number of L3 lines allocated in any state.
322.It Li L3_LINES_OUT.M_STATE
323.Pq Event 0BH , Umask 01H
324Counts the number of L3 lines victimized that were in the M state.
325When the victim cache line is in M state, the line is written to its home cache agent
326which can be either local or remote.
327.It Li L3_LINES_OUT.E_STATE
328.Pq Event 0BH , Umask 02H
329Counts the number of L3 lines victimized that were in the E state.
330.It Li L3_LINES_OUT.S_STATE
331.Pq Event 0BH , Umask 04H
332Counts the number of L3 lines victimized that were in the S state.
333.It Li L3_LINES_OUT.I_STATE
334.Pq Event 0BH , Umask 08H
335Counts the number of L3 lines victimized that were in the I state.
336.It Li L3_LINES_OUT.F_STATE
337.Pq Event 0BH , Umask 10H
338Counts the number of L3 lines victimized that were in the F state.
339.It Li L3_LINES_OUT.ANY
340.Pq Event 0BH , Umask 1FH
341Counts the number of L3 lines victimized in any state.
342.It Li GQ_SNOOP.GOTO_S
343.Pq Event 0CH , Umask 01H
344Counts the number of remote snoops that have requested a cache line be set
345to the S state.
346.It Li GQ_SNOOP.GOTO_I
347.Pq Event 0CH , Umask 02H
348Counts the number of remote snoops that have requested a cache line be set
349to the I state.
350.It Li GQ_SNOOP.GOTO_S_HIT_E
351.Pq Event 0CH , Umask 04H
352Counts the number of remote snoops that have requested a cache line be set
353to the S state from E state.
354Requires writing MSR 301H with mask = 2H
355.It Li GQ_SNOOP.GOTO_S_HIT_F
356.Pq Event 0CH , Umask 04H
357Counts the number of remote snoops that have requested a cache line be set
358to the S state from F (forward) state.
359Requires writing MSR 301H with mask = 8H
360.It Li GQ_SNOOP.GOTO_S_HIT_M
361.Pq Event 0CH , Umask 04H
362Counts the number of remote snoops that have requested a cache line be set
363to the S state from M state.
364Requires writing MSR 301H with mask = 1H
365.It Li GQ_SNOOP.GOTO_S_HIT_S
366.Pq Event 0CH , Umask 04H
367Counts the number of remote snoops that have requested a cache line be set
368to the S state from S state.
369Requires writing MSR 301H with mask = 4H
370.It Li GQ_SNOOP.GOTO_I_HIT_E
371.Pq Event 0CH , Umask 08H
372Counts the number of remote snoops that have requested a cache line be set
373to the I state from E state.
374Requires writing MSR 301H with mask = 2H
375.It Li GQ_SNOOP.GOTO_I_HIT_F
376.Pq Event 0CH , Umask 08H
377Counts the number of remote snoops that have requested a cache line be set
378to the I state from F (forward) state.
379Requires writing MSR 301H with mask = 8H
380.It Li GQ_SNOOP.GOTO_I_HIT_M
381.Pq Event 0CH , Umask 08H
382Counts the number of remote snoops that have requested a cache line be set
383to the I state from M state.
384Requires writing MSR 301H with mask = 1H
385.It Li GQ_SNOOP.GOTO_I_HIT_S
386.Pq Event 0CH , Umask 08H
387Counts the number of remote snoops that have requested a cache line be set
388to the I state from S state.
389Requires writing MSR 301H with mask = 4H
390.It Li QHL_REQUESTS.IOH_READS
391.Pq Event 20H , Umask 01H
392Counts number of Quickpath Home Logic read requests from the IOH.
393.It Li QHL_REQUESTS.IOH_WRITES
394.Pq Event 20H , Umask 02H
395Counts number of Quickpath Home Logic write requests from the IOH.
396.It Li QHL_REQUESTS.REMOTE_READS
397.Pq Event 20H , Umask 04H
398Counts number of Quickpath Home Logic read requests from a remote socket.
399.It Li QHL_REQUESTS.REMOTE_WRITES
400.Pq Event 20H , Umask 08H
401Counts number of Quickpath Home Logic write requests from a remote socket.
402.It Li QHL_REQUESTS.LOCAL_READS
403.Pq Event 20H , Umask 10H
404Counts number of Quickpath Home Logic read requests from the local socket.
405.It Li QHL_REQUESTS.LOCAL_WRITES
406.Pq Event 20H , Umask 20H
407Counts number of Quickpath Home Logic write requests from the local socket.
408.It Li QHL_CYCLES_FULL.IOH
409.Pq Event 21H , Umask 01H
410Counts uclk cycles all entries in the Quickpath Home Logic IOH are full.
411.It Li QHL_CYCLES_FULL.REMOTE
412.Pq Event 21H , Umask 02H
413Counts uclk cycles all entries in the Quickpath Home Logic remote tracker
414are full.
415.It Li QHL_CYCLES_FULL.LOCAL
416.Pq Event 21H , Umask 04H
417Counts uclk cycles all entries in the Quickpath Home Logic local tracker are
418full.
419.It Li QHL_CYCLES_NOT_EMPTY.IOH
420.Pq Event 22H , Umask 01H
421Counts uclk cycles all entries in the Quickpath Home Logic IOH is busy.
422.It Li QHL_CYCLES_NOT_EMPTY.REMOTE
423.Pq Event 22H , Umask 02H
424Counts uclk cycles all entries in the Quickpath Home Logic remote tracker is
425busy.
426.It Li QHL_CYCLES_NOT_EMPTY.LOCAL
427.Pq Event 22H , Umask 04H
428Counts uclk cycles all entries in the Quickpath Home Logic local tracker is
429busy.
430.It Li QHL_OCCUPANCY.IOH
431.Pq Event 23H , Umask 01H
432QHL IOH tracker allocate to deallocate read occupancy.
433.It Li QHL_OCCUPANCY.REMOTE
434.Pq Event 23H , Umask 02H
435QHL remote tracker allocate to deallocate read occupancy.
436.It Li QHL_OCCUPANCY.LOCAL
437.Pq Event 23H , Umask 04H
438QHL local tracker allocate to deallocate read occupancy.
439.It Li QHL_ADDRESS_CONFLICTS.2WAY
440.Pq Event 24H , Umask 02H
441Counts number of QHL Active Address Table (AAT) entries that saw a max of 2 conflicts.
442The AAT is a structure that tracks requests that are in conflict.
443The requests themselves are in the home tracker entries.
444The count is reported when an AAT entry deallocates.
445.It Li QHL_ADDRESS_CONFLICTS.3WAY
446.Pq Event 24H , Umask 04H
447Counts number of QHL Active Address Table (AAT) entries that saw a max of 3 conflicts.
448The AAT is a structure that tracks requests that are in conflict.
449The requests themselves are in the home tracker entries.
450The count is reported when an AAT entry deallocates.
451.It Li QHL_CONFLICT_CYCLES.IOH
452.Pq Event 25H , Umask 01H
453Counts cycles the Quickpath Home Logic IOH Tracker contains two or more
454requests with an address conflict.
455A max of 3 requests can be in conflict.
456.It Li QHL_CONFLICT_CYCLES.REMOTE
457.Pq Event 25H , Umask 02H
458Counts cycles the Quickpath Home Logic Remote Tracker contains two or more
459requests with an address conflict.
460A max of 3 requests can be in conflict.
461.It Li QHL_CONFLICT_CYCLES.LOCAL
462.Pq Event 25H , Umask 04H
463Counts cycles the Quickpath Home Logic Local Tracker contains two or more
464requests with an address conflict.
465A max of 3 requests can be in conflict.
466.It Li QHL_TO_QMC_BYPASS
467.Pq Event 26H , Umask 01H
468Counts number or requests to the Quickpath Memory Controller that bypass the
469Quickpath Home Logic.
470All local accesses can be bypassed.
471For remote requests, only read requests can be bypassed.
472.It Li QMC_ISOC_FULL.READ.CH0
473.Pq Event 28H , Umask 01H
474Counts cycles all the entries in the DRAM channel 0 high priority queue are
475occupied with isochronous read requests.
476.It Li QMC_ISOC_FULL.READ.CH1
477.Pq Event 28H , Umask 02H
478Counts cycles all the entries in the DRAM channel 1 high priority queue are
479occupied with isochronous read requests.
480.It Li QMC_ISOC_FULL.READ.CH2
481.Pq Event 28H , Umask 04H
482Counts cycles all the entries in the DRAM channel 2 high priority queue are
483occupied with isochronous read requests.
484.It Li QMC_ISOC_FULL.WRITE.CH0
485.Pq Event 28H , Umask 08H
486Counts cycles all the entries in the DRAM channel 0 high priority queue are
487occupied with isochronous write requests.
488.It Li QMC_ISOC_FULL.WRITE.CH1
489.Pq Event 28H , Umask 10H
490Counts cycles all the entries in the DRAM channel 1 high priority queue are
491occupied with isochronous write requests.
492.It Li QMC_ISOC_FULL.WRITE.CH2
493.Pq Event 28H , Umask 20H
494Counts cycles all the entries in the DRAM channel 2 high priority queue are
495occupied with isochronous write requests.
496.It Li QMC_BUSY.READ.CH0
497.Pq Event 29H , Umask 01H
498Counts cycles where Quickpath Memory Controller has at least 1 outstanding
499read request to DRAM channel 0.
500.It Li QMC_BUSY.READ.CH1
501.Pq Event 29H , Umask 02H
502Counts cycles where Quickpath Memory Controller has at least 1 outstanding
503read request to DRAM channel 1.
504.It Li QMC_BUSY.READ.CH2
505.Pq Event 29H , Umask 04H
506Counts cycles where Quickpath Memory Controller has at least 1 outstanding
507read request to DRAM channel 2.
508.It Li QMC_BUSY.WRITE.CH0
509.Pq Event 29H , Umask 08H
510Counts cycles where Quickpath Memory Controller has at least 1 outstanding
511write request to DRAM channel 0.
512.It Li QMC_BUSY.WRITE.CH1
513.Pq Event 29H , Umask 10H
514Counts cycles where Quickpath Memory Controller has at least 1 outstanding
515write request to DRAM channel 1.
516.It Li QMC_BUSY.WRITE.CH2
517.Pq Event 29H , Umask 20H
518Counts cycles where Quickpath Memory Controller has at least 1 outstanding
519write request to DRAM channel 2.
520.It Li QMC_OCCUPANCY.CH0
521.Pq Event 2AH , Umask 01H
522IMC channel 0 normal read request occupancy.
523.It Li QMC_OCCUPANCY.CH1
524.Pq Event 2AH , Umask 02H
525IMC channel 1 normal read request occupancy.
526.It Li QMC_OCCUPANCY.CH2
527.Pq Event 2AH , Umask 04H
528IMC channel 2 normal read request occupancy.
529.It Li QMC_OCCUPANCY.ANY
530.Pq Event 2AH , Umask 07H
531Normal read request occupancy for any channel.
532.It Li QMC_ISSOC_OCCUPANCY.CH0
533.Pq Event 2BH , Umask 01H
534IMC channel 0 issoc read request occupancy.
535.It Li QMC_ISSOC_OCCUPANCY.CH1
536.Pq Event 2BH , Umask 02H
537IMC channel 1 issoc read request occupancy.
538.It Li QMC_ISSOC_OCCUPANCY.CH2
539.Pq Event 2BH , Umask 04H
540IMC channel 2 issoc read request occupancy.
541.It Li QMC_ISSOC_READS.ANY
542.Pq Event 2BH , Umask 07H
543IMC issoc read request occupancy.
544.It Li QMC_NORMAL_READS.CH0
545.Pq Event 2CH , Umask 01H
546Counts the number of Quickpath Memory Controller channel 0 medium and low
547priority read requests.
548The QMC channel 0 normal read occupancy divided by this count provides the
549average QMC channel 0 read latency.
550.It Li QMC_NORMAL_READS.CH1
551.Pq Event 2CH , Umask 02H
552Counts the number of Quickpath Memory Controller channel 1 medium and low
553priority read requests.
554The QMC channel 1 normal read occupancy divided by this count provides the
555average QMC channel 1 read latency.
556.It Li QMC_NORMAL_READS.CH2
557.Pq Event 2CH , Umask 04H
558Counts the number of Quickpath Memory Controller channel 2 medium and low
559priority read requests.
560The QMC channel 2 normal read occupancy divided by this count provides the
561average QMC channel 2 read latency.
562.It Li QMC_NORMAL_READS.ANY
563.Pq Event 2CH , Umask 07H
564Counts the number of Quickpath Memory Controller medium and low priority read requests.
565The QMC normal read occupancy divided by this count provides the average
566QMC read latency.
567.It Li QMC_HIGH_PRIORITY_READS.CH0
568.Pq Event 2DH , Umask 01H
569Counts the number of Quickpath Memory Controller channel 0 high priority
570isochronous read requests.
571.It Li QMC_HIGH_PRIORITY_READS.CH1
572.Pq Event 2DH , Umask 02H
573Counts the number of Quickpath Memory Controller channel 1 high priority
574isochronous read requests.
575.It Li QMC_HIGH_PRIORITY_READS.CH2
576.Pq Event 2DH , Umask 04H
577Counts the number of Quickpath Memory Controller channel 2 high priority
578isochronous read requests.
579.It Li QMC_HIGH_PRIORITY_READS.ANY
580.Pq Event 2DH , Umask 07H
581Counts the number of Quickpath Memory Controller high priority isochronous
582read requests.
583.It Li QMC_CRITICAL_PRIORITY_READS.CH0
584.Pq Event 2EH , Umask 01H
585Counts the number of Quickpath Memory Controller channel 0 critical priority
586isochronous read requests.
587.It Li QMC_CRITICAL_PRIORITY_READS.CH1
588.Pq Event 2EH , Umask 02H
589Counts the number of Quickpath Memory Controller channel 1 critical priority
590isochronous read requests.
591.It Li QMC_CRITICAL_PRIORITY_READS.CH2
592.Pq Event 2EH , Umask 04H
593Counts the number of Quickpath Memory Controller channel 2 critical priority
594isochronous read requests.
595.It Li QMC_CRITICAL_PRIORITY_READS.ANY
596.Pq Event 2EH , Umask 07H
597Counts the number of Quickpath Memory Controller critical priority
598isochronous read requests.
599.It Li QMC_WRITES.FULL.CH0
600.Pq Event 2FH , Umask 01H
601Counts number of full cache line writes to DRAM channel 0.
602.It Li QMC_WRITES.FULL.CH1
603.Pq Event 2FH , Umask 02H
604Counts number of full cache line writes to DRAM channel 1.
605.It Li QMC_WRITES.FULL.CH2
606.Pq Event 2FH , Umask 04H
607Counts number of full cache line writes to DRAM channel 2.
608.It Li QMC_WRITES.FULL.ANY
609.Pq Event 2FH , Umask 07H
610Counts number of full cache line writes to DRAM.
611.It Li QMC_WRITES.PARTIAL.CH0
612.Pq Event 2FH , Umask 08H
613Counts number of partial cache line writes to DRAM channel 0.
614.It Li QMC_WRITES.PARTIAL.CH1
615.Pq Event 2FH , Umask 10H
616Counts number of partial cache line writes to DRAM channel 1.
617.It Li QMC_WRITES.PARTIAL.CH2
618.Pq Event 2FH , Umask 20H
619Counts number of partial cache line writes to DRAM channel 2.
620.It Li QMC_WRITES.PARTIAL.ANY
621.Pq Event 2FH , Umask 38H
622Counts number of partial cache line writes to DRAM.
623.It Li QMC_CANCEL.CH0
624.Pq Event 30H , Umask 01H
625Counts number of DRAM channel 0 cancel requests.
626.It Li QMC_CANCEL.CH1
627.Pq Event 30H , Umask 02H
628Counts number of DRAM channel 1 cancel requests.
629.It Li QMC_CANCEL.CH2
630.Pq Event 30H , Umask 04H
631Counts number of DRAM channel 2 cancel requests.
632.It Li QMC_CANCEL.ANY
633.Pq Event 30H , Umask 07H
634Counts number of DRAM cancel requests.
635.It Li QMC_PRIORITY_UPDATES.CH0
636.Pq Event 31H , Umask 01H
637Counts number of DRAM channel 0 priority updates.
638A priority update occurs when an ISOC high or critical request is
639received by the QHL and there is a matching request with normal priority
640that has already been issued to the QMC.
641In this instance, the QHL will send a priority update to QMC to
642expedite the request.
643.It Li QMC_PRIORITY_UPDATES.CH1
644.Pq Event 31H , Umask 02H
645Counts number of DRAM channel 1 priority updates.
646A priority update occurs when an ISOC high or critical request is received
647by the QHL and there is a matching request with normal priority that has
648already been issued to the QMC.
649In this instance, the QHL will send a priority update to QMC to expedite the request.
650.It Li QMC_PRIORITY_UPDATES.CH2
651.Pq Event 31H , Umask 04H
652Counts number of DRAM channel 2 priority updates.
653A priority update occurs when an ISOC high or critical request is received
654by the QHL and there is a matching request with normal priority that has
655already been issued to the QMC.
656In this instance, the QHL will send a priority update to QMC to expedite the request.
657.It Li QMC_PRIORITY_UPDATES.ANY
658.Pq Event 31H , Umask 07H
659Counts number of DRAM priority updates.
660A priority update occurs when an ISOC high or critical request is received
661by the QHL and there is a matching request with normal priority that has already
662been issued to the QMC.
663In this instance, the QHL will send a priority update to QMC to expedite the request.
664.It Li IMC_RETRY.CH0
665.Pq Event 32H , Umask 01H
666Counts number of IMC DRAM channel 0 retries.
667DRAM retry only occurs when configured in RAS mode.
668.It Li IMC_RETRY.CH1
669.Pq Event 32H , Umask 02H
670Counts number of IMC DRAM channel 1 retries.
671DRAM retry only occurs when configured in RAS mode.
672.It Li IMC_RETRY.CH2
673.Pq Event 32H , Umask 04H
674Counts number of IMC DRAM channel 2 retries.
675DRAM retry only occurs when configured in RAS mode.
676.It Li IMC_RETRY.ANY
677.Pq Event 32H , Umask 07H
678Counts number of IMC DRAM retries from any channel.
679DRAM retry only occurs when configured in RAS mode.
680.It Li QHL_FRC_ACK_CNFLTS.IOH
681.Pq Event 33H , Umask 01H
682Counts number of Force Acknowledge Conflict messages sent by the Quickpath
683Home Logic to the IOH.
684.It Li QHL_FRC_ACK_CNFLTS.REMOTE
685.Pq Event 33H , Umask 02H
686Counts number of Force Acknowledge Conflict messages sent by the Quickpath
687Home Logic to the remote home.
688.It Li QHL_FRC_ACK_CNFLTS.LOCAL
689.Pq Event 33H , Umask 04H
690Counts number of Force Acknowledge Conflict messages sent by the Quickpath
691Home Logic to the local home.
692.It Li QHL_FRC_ACK_CNFLTS.ANY
693.Pq Event 33H , Umask 07H
694Counts number of Force Acknowledge Conflict messages sent by the Quickpath
695Home Logic.
696.It Li QHL_SLEEPS.IOH_ORDER
697.Pq Event 34H , Umask 01H
698Counts number of occurrences a request was put to sleep due to IOH ordering
699(write after read) conflicts.
700While in the sleep state, the request is not eligible to be scheduled to the QMC.
701.It Li QHL_SLEEPS.REMOTE_ORDER
702.Pq Event 34H , Umask 02H
703Counts number of occurrences a request was put to sleep due to remote socket
704ordering (write after read) conflicts.
705While in the sleep state, the request is not eligible to be scheduled to the QMC.
706.It Li QHL_SLEEPS.LOCAL_ORDER
707.Pq Event 34H , Umask 04H
708Counts number of occurrences a request was put to sleep due to local socket
709ordering (write after read) conflicts.
710While in the sleep state, the request is not eligible to be scheduled to the QMC.
711.It Li QHL_SLEEPS.IOH_CONFLICT
712.Pq Event 34H , Umask 08H
713Counts number of occurrences a request was put to sleep due to IOH address conflicts.
714While in the sleep state, the request is not eligible to be scheduled to the QMC.
715.It Li QHL_SLEEPS.REMOTE_CONFLICT
716.Pq Event 34H , Umask 10H
717Counts number of occurrences a request was put to sleep due to remote socket
718address conflicts.
719While in the sleep state, the request is not eligible to be scheduled to the QMC.
720.It Li QHL_SLEEPS.LOCAL_CONFLICT
721.Pq Event 34H , Umask 20H
722Counts number of occurrences a request was put to sleep due to local socket address conflicts.
723While in the sleep state, the request is not eligible to be scheduled to the QMC.
724.It Li ADDR_OPCODE_MATCH.IOH
725.Pq Event 35H , Umask 01H
726Counts number of requests from the IOH, address/opcode of request is
727qualified by mask value written to MSR 396H.
728The following mask values are supported:
7290: NONE 40000000_00000000H:RSPFWDI 40001A00_00000000H:RSPFWDS
73040001D00_00000000H:RSPIWB
731Match opcode/address by writing MSR 396H with mask supported mask value.
732.It Li ADDR_OPCODE_MATCH.REMOTE
733.Pq Event 35H , Umask 02H
734Counts number of requests from the remote socket, address/opcode of request
735is qualified by mask value written to MSR 396H.
736The following mask values are supported:
7370: NONE 40000000_00000000H:RSPFWDI 40001A00_00000000H:RSPFWDS
73840001D00_00000000H:RSPIWB
739Match opcode/address by writing MSR 396H with mask supported mask value.
740.It Li ADDR_OPCODE_MATCH.LOCAL
741.Pq Event 35H , Umask 04H
742Counts number of requests from the local socket, address/opcode of request
743is qualified by mask value written to MSR 396H.
744The following mask values are supported:
7450: NONE 40000000_00000000H:RSPFWDI 40001A00_00000000H:RSPFWDS
74640001D00_00000000H:RSPIWB
747Match opcode/address by writing MSR 396H with mask supported mask value.
748.It Li QPI_TX_STALLED_SINGLE_FLIT.HOME.LINK_0
749.Pq Event 40H , Umask 01H
750Counts cycles the Quickpath outbound link 0 HOME virtual channel is stalled
751due to lack of a VNA and VN0 credit.
752Note that this event does not filter out when a flit would not have been selected
753for arbitration because another virtual channel is getting arbitrated.
754.It Li QPI_TX_STALLED_SINGLE_FLIT.SNOOP.LINK_0
755.Pq Event 40H , Umask 02H
756Counts cycles the Quickpath outbound link 0 SNOOP virtual channel is stalled
757due to lack of a VNA and VN0 credit.
758Note that this event does not filter out when a flit would not have been selected
759for arbitration because another virtual channel is getting arbitrated.
760.It Li QPI_TX_STALLED_SINGLE_FLIT.NDR.LINK_0
761.Pq Event 40H , Umask 04H
762Counts cycles the Quickpath outbound link 0 non-data response virtual
763channel is stalled due to lack of a VNA and VN0 credit.
764Note that this event does not filter out when a flit would not have been selected
765for arbitration because another virtual channel is getting arbitrated.
766.It Li QPI_TX_STALLED_SINGLE_FLIT.HOME.LINK_1
767.Pq Event 40H , Umask 08H
768Counts cycles the Quickpath outbound link 1 HOME virtual channel is stalled
769due to lack of a VNA and VN0 credit.
770Note that this event does not filter out when a flit would not have been selected
771for arbitration because another virtual channel is getting arbitrated.
772.It Li QPI_TX_STALLED_SINGLE_FLIT.SNOOP.LINK_1
773.Pq Event 40H , Umask 10H
774Counts cycles the Quickpath outbound link 1 SNOOP virtual channel is stalled
775due to lack of a VNA and VN0 credit.
776Note that this event does not filter out when a flit would not have been selected
777for arbitration because another virtual channel is getting arbitrated.
778.It Li QPI_TX_STALLED_SINGLE_FLIT.NDR.LINK_1
779.Pq Event 40H , Umask 20H
780Counts cycles the Quickpath outbound link 1 non-data response virtual
781channel is stalled due to lack of a VNA and VN0 credit.
782Note that this event does not filter out when a flit would not have been selected
783for arbitration because another virtual channel is getting arbitrated.
784.It Li QPI_TX_STALLED_SINGLE_FLIT.LINK_0
785.Pq Event 40H , Umask 07H
786Counts cycles the Quickpath outbound link 0 virtual channels are stalled due
787to lack of a VNA and VN0 credit.
788Note that this event does not filter out when a flit would not have been selected
789for arbitration because another virtual channel is getting arbitrated.
790.It Li QPI_TX_STALLED_SINGLE_FLIT.LINK_1
791.Pq Event 40H , Umask 38H
792Counts cycles the Quickpath outbound link 1 virtual channels are stalled due
793to lack of a VNA and VN0 credit.
794Note that this event does not filter out when a flit would not have been selected
795for arbitration because another virtual channel is getting arbitrated.
796.It Li QPI_TX_STALLED_MULTI_FLIT.DRS.LINK_0
797.Pq Event 41H , Umask 01H
798Counts cycles the Quickpath outbound link 0 Data ResponSe virtual channel is
799stalled due to lack of VNA and VN0 credits.
800Note that this event does not filter out when a flit would not have been selected
801for arbitration because another virtual channel is getting arbitrated.
802.It Li QPI_TX_STALLED_MULTI_FLIT.NCB.LINK_0
803.Pq Event 41H , Umask 02H
804Counts cycles the Quickpath outbound link 0 Non-Coherent Bypass virtual
805channel is stalled due to lack of VNA and VN0 credits.
806Note that this event does not filter out when a flit would not have been selected
807for arbitration because another virtual channel is getting arbitrated.
808.It Li QPI_TX_STALLED_MULTI_FLIT.NCS.LINK_0
809.Pq Event 41H , Umask 04H
810Counts cycles the Quickpath outbound link 0 Non-Coherent Standard virtual
811channel is stalled due to lack of VNA and VN0 credits.
812Note that this event does not filter out when a flit would not have been selected
813for arbitration because another virtual channel is getting arbitrated.
814.It Li QPI_TX_STALLED_MULTI_FLIT.DRS.LINK_1
815.Pq Event 41H , Umask 08H
816Counts cycles the Quickpath outbound link 1 Data ResponSe virtual channel is
817stalled due to lack of VNA and VN0 credits.
818Note that this event does not filter out when a flit would not have been selected
819for arbitration because another virtual channel is getting arbitrated.
820.It Li QPI_TX_STALLED_MULTI_FLIT.NCB.LINK_1
821.Pq Event 41H , Umask 10H
822Counts cycles the Quickpath outbound link 1 Non-Coherent Bypass virtual
823channel is stalled due to lack of VNA and VN0 credits.
824Note that this event does not filter out when a flit would not have been selected
825for arbitration because another virtual channel is getting arbitrated.
826.It Li QPI_TX_STALLED_MULTI_FLIT.NCS.LINK_1
827.Pq Event 41H , Umask 20H
828Counts cycles the Quickpath outbound link 1 Non-Coherent Standard virtual
829channel is stalled due to lack of VNA and VN0 credits.
830Note that this event does not filter out when a flit would not have been selected
831for arbitration because another virtual channel is getting arbitrated.
832.It Li QPI_TX_STALLED_MULTI_FLIT.LINK_0
833.Pq Event 41H , Umask 07H
834Counts cycles the Quickpath outbound link 0 virtual channels are stalled due
835to lack of VNA and VN0 credits.
836Note that this event does not filter out when a flit would not have been selected
837for arbitration because another virtual channel is getting arbitrated.
838.It Li QPI_TX_STALLED_MULTI_FLIT.LINK_1
839.Pq Event 41H , Umask 38H
840Counts cycles the Quickpath outbound link 1 virtual channels are stalled due
841to lack of VNA and VN0 credits.
842Note that this event does not filter out when a flit would not have been selected
843for arbitration because another virtual channel is getting arbitrated.
844.It Li QPI_TX_HEADER.FULL.LINK_0
845.Pq Event 42H , Umask 01H
846Number of cycles that the header buffer in the Quickpath Interface outbound
847link 0 is full.
848.It Li QPI_TX_HEADER.BUSY.LINK_0
849.Pq Event 42H , Umask 02H
850Number of cycles that the header buffer in the Quickpath Interface outbound
851link 0 is busy.
852.It Li QPI_TX_HEADER.FULL.LINK_1
853.Pq Event 42H , Umask 04H
854Number of cycles that the header buffer in the Quickpath Interface outbound
855link 1 is full.
856.It Li QPI_TX_HEADER.BUSY.LINK_1
857.Pq Event 42H , Umask 08H
858Number of cycles that the header buffer in the Quickpath Interface outbound
859link 1 is busy.
860.It Li QPI_RX_NO_PPT_CREDIT.STALLS.LINK_0
861.Pq Event 43H , Umask 01H
862Number of cycles that snoop packets incoming to the Quickpath Interface link
8630 are stalled and not sent to the GQ because the GQ Peer Probe Tracker (PPT)
864does not have any available entries.
865.It Li QPI_RX_NO_PPT_CREDIT.STALLS.LINK_1
866.Pq Event 43H , Umask 02H
867Number of cycles that snoop packets incoming to the Quickpath Interface link
8681 are stalled and not sent to the GQ because the GQ Peer Probe Tracker (PPT)
869does not have any available entries.
870.It Li DRAM_OPEN.CH0
871.Pq Event 60H , Umask 01H
872Counts number of DRAM Channel 0 open commands issued either for read or write.
873To read or write data, the referenced DRAM page must first be opened.
874.It Li DRAM_OPEN.CH1
875.Pq Event 60H , Umask 02H
876Counts number of DRAM Channel 1 open commands issued either for read or write.
877To read or write data, the referenced DRAM page must first be opened.
878.It Li DRAM_OPEN.CH2
879.Pq Event 60H , Umask 04H
880Counts number of DRAM Channel 2 open commands issued either for read or write.
881To read or write data, the referenced DRAM page must first be opened.
882.It Li DRAM_PAGE_CLOSE.CH0
883.Pq Event 61H , Umask 01H
884DRAM channel 0 command issued to CLOSE a page due to page idle timer expiration.
885Closing a page is done by issuing a precharge.
886.It Li DRAM_PAGE_CLOSE.CH1
887.Pq Event 61H , Umask 02H
888DRAM channel 1 command issued to CLOSE a page due to page idle timer expiration.
889Closing a page is done by issuing a precharge.
890.It Li DRAM_PAGE_CLOSE.CH2
891.Pq Event 61H , Umask 04H
892DRAM channel 2 command issued to CLOSE a page due to page idle timer expiration.
893Closing a page is done by issuing a precharge.
894.It Li DRAM_PAGE_MISS.CH0
895.Pq Event 62H , Umask 01H
896Counts the number of precharges (PRE) that were issued to DRAM channel 0
897because there was a page miss.
898A page miss refers to a situation in which a page is currently open and another
899page from the same bank needs to be opened.
900The new page experiences a page miss.
901Closing of the old page is done by issuing a precharge.
902.It Li DRAM_PAGE_MISS.CH1
903.Pq Event 62H , Umask 02H
904Counts the number of precharges (PRE) that were issued to DRAM channel 1
905because there was a page miss.
906A page miss refers to a situation in which a page is currently open and another
907page from the same bank needs to be opened.
908The new page experiences a page miss.
909Closing of the old page is done by issuing a precharge.
910.It Li DRAM_PAGE_MISS.CH2
911.Pq Event 62H , Umask 04H
912Counts the number of precharges (PRE) that were issued to DRAM channel 2
913because there was a page miss.
914A page miss refers to a situation in which a page is currently open and another
915page from the same bank needs to be opened.
916The new page experiences a page miss.
917Closing of the old page is done by issuing a precharge.
918.It Li DRAM_READ_CAS.CH0
919.Pq Event 63H , Umask 01H
920Counts the number of times a read CAS command was issued on DRAM channel 0.
921.It Li DRAM_READ_CAS.AUTOPRE_CH0
922.Pq Event 63H , Umask 02H
923Counts the number of times a read CAS command was issued on DRAM channel 0
924where the command issued used the auto-precharge (auto page close) mode.
925.It Li DRAM_READ_CAS.CH1
926.Pq Event 63H , Umask 04H
927Counts the number of times a read CAS command was issued on DRAM channel 1.
928.It Li DRAM_READ_CAS.AUTOPRE_CH1
929.Pq Event 63H , Umask 08H
930Counts the number of times a read CAS command was issued on DRAM channel 1
931where the command issued used the auto-precharge (auto page close) mode.
932.It Li DRAM_READ_CAS.CH2
933.Pq Event 63H , Umask 10H
934Counts the number of times a read CAS command was issued on DRAM channel 2.
935.It Li DRAM_READ_CAS.AUTOPRE_CH2
936.Pq Event 63H , Umask 20H
937Counts the number of times a read CAS command was issued on DRAM channel 2
938where the command issued used the auto-precharge (auto page close) mode.
939.It Li DRAM_WRITE_CAS.CH0
940.Pq Event 64H , Umask 01H
941Counts the number of times a write CAS command was issued on DRAM channel 0.
942.It Li DRAM_WRITE_CAS.AUTOPRE_CH0
943.Pq Event 64H , Umask 02H
944Counts the number of times a write CAS command was issued on DRAM channel 0
945where the command issued used the auto-precharge (auto page close) mode.
946.It Li DRAM_WRITE_CAS.CH1
947.Pq Event 64H , Umask 04H
948Counts the number of times a write CAS command was issued on DRAM channel 1.
949.It Li DRAM_WRITE_CAS.AUTOPRE_CH1
950.Pq Event 64H , Umask 08H
951Counts the number of times a write CAS command was issued on DRAM channel 1
952where the command issued used the auto-precharge (auto page close) mode.
953.It Li DRAM_WRITE_CAS.CH2
954.Pq Event 64H , Umask 10H
955Counts the number of times a write CAS command was issued on DRAM channel 2.
956.It Li DRAM_WRITE_CAS.AUTOPRE_CH2
957.Pq Event 64H , Umask 20H
958Counts the number of times a write CAS command was issued on DRAM channel 2
959where the command issued used the auto-precharge (auto page close) mode.
960.It Li DRAM_REFRESH.CH0
961.Pq Event 65H , Umask 01H
962Counts number of DRAM channel 0 refresh commands.
963DRAM loses data content over time.
964In order to keep correct data content, the data values have to be
965refreshed periodically.
966.It Li DRAM_REFRESH.CH1
967.Pq Event 65H , Umask 02H
968Counts number of DRAM channel 1 refresh commands.
969DRAM loses data content over time.
970In order to keep correct data content, the data values have to be refreshed periodically.
971.It Li DRAM_REFRESH.CH2
972.Pq Event 65H , Umask 04H
973Counts number of DRAM channel 2 refresh commands.
974DRAM loses data content over time.
975In order to keep correct data content, the data values have to be refreshed periodically.
976.It Li DRAM_PRE_ALL.CH0
977.Pq Event 66H , Umask 01H
978Counts number of DRAM Channel 0 precharge-all (PREALL) commands that close
979all open pages in a rank.
980PREALL is issued when the DRAM needs to be refreshed or needs to go into a power down mode.
981.It Li DRAM_PRE_ALL.CH1
982.Pq Event 66H , Umask 02H
983Counts number of DRAM Channel 1 precharge-all (PREALL) commands that close
984all open pages in a rank.
985PREALL is issued when the DRAM needs to be refreshed or needs to go into a power down mode.
986.It Li DRAM_PRE_ALL.CH2
987.Pq Event 66H , Umask 04H
988Counts number of DRAM Channel 2 precharge-all (PREALL) commands that close
989all open pages in a rank.
990PREALL is issued when the DRAM needs to be refreshed or needs to go into a power down mode.
991.It Li DRAM_THERMAL_THROTTLED
992.Pq Event 67H , Umask 01H
993Uncore cycles DRAM was throttled due to its temperature being above the
994thermal throttling threshold.
995.It Li THERMAL_THROTTLING_TEMP.CORE_0
996.Pq Event 80H , Umask 01H
997Cycles that the PCU records that core 0 is above the thermal throttling
998threshold temperature.
999.It Li THERMAL_THROTTLING_TEMP.CORE_1
1000.Pq Event 80H , Umask 02H
1001Cycles that the PCU records that core 1 is above the thermal throttling
1002threshold temperature.
1003.It Li THERMAL_THROTTLING_TEMP.CORE_2
1004.Pq Event 80H , Umask 04H
1005Cycles that the PCU records that core 2 is above the thermal throttling
1006threshold temperature.
1007.It Li THERMAL_THROTTLING_TEMP.CORE_3
1008.Pq Event 80H , Umask 08H
1009Cycles that the PCU records that core 3 is above the thermal throttling
1010threshold temperature.
1011.It Li THERMAL_THROTTLED_TEMP.CORE_0
1012.Pq Event 81H , Umask 01H
1013Cycles that the PCU records that core 0 is in the power throttled state due
1014to cores temperature being above the thermal throttling threshold.
1015.It Li THERMAL_THROTTLED_TEMP.CORE_1
1016.Pq Event 81H , Umask 02H
1017Cycles that the PCU records that core 1 is in the power throttled state due
1018to cores temperature being above the thermal throttling threshold.
1019.It Li THERMAL_THROTTLED_TEMP.CORE_2
1020.Pq Event 81H , Umask 04H
1021Cycles that the PCU records that core 2 is in the power throttled state due
1022to cores temperature being above the thermal throttling threshold.
1023.It Li THERMAL_THROTTLED_TEMP.CORE_3
1024.Pq Event 81H , Umask 08H
1025Cycles that the PCU records that core 3 is in the power throttled state due
1026to cores temperature being above the thermal throttling threshold.
1027.It Li PROCHOT_ASSERTION
1028.Pq Event 82H , Umask 01H
1029Number of system assertions of PROCHOT indicating the entire processor has
1030exceeded the thermal limit.
1031.It Li THERMAL_THROTTLING_PROCHOT.CORE_0
1032.Pq Event 83H , Umask 01H
1033Cycles that the PCU records that core 0 is a low power state due to the
1034system asserting PROCHOT the entire processor has exceeded the thermal
1035limit.
1036.It Li THERMAL_THROTTLING_PROCHOT.CORE_1
1037.Pq Event 83H , Umask 02H
1038Cycles that the PCU records that core 1 is a low power state due to the
1039system asserting PROCHOT the entire processor has exceeded the thermal
1040limit.
1041.It Li THERMAL_THROTTLING_PROCHOT.CORE_2
1042.Pq Event 83H , Umask 04H
1043Cycles that the PCU records that core 2 is a low power state due to the
1044system asserting PROCHOT the entire processor has exceeded the thermal
1045limit.
1046.It Li THERMAL_THROTTLING_PROCHOT.CORE_3
1047.Pq Event 83H , Umask 08H
1048Cycles that the PCU records that core 3 is a low power state due to the
1049system asserting PROCHOT the entire processor has exceeded the thermal
1050limit.
1051.It Li TURBO_MODE.CORE_0
1052.Pq Event 84H , Umask 01H
1053Uncore cycles that core 0 is operating in turbo mode.
1054.It Li TURBO_MODE.CORE_1
1055.Pq Event 84H , Umask 02H
1056Uncore cycles that core 1 is operating in turbo mode.
1057.It Li TURBO_MODE.CORE_2
1058.Pq Event 84H , Umask 04H
1059Uncore cycles that core 2 is operating in turbo mode.
1060.It Li TURBO_MODE.CORE_3
1061.Pq Event 84H , Umask 08H
1062Uncore cycles that core 3 is operating in turbo mode.
1063.It Li CYCLES_UNHALTED_L3_FLL_ENABLE
1064.Pq Event 85H , Umask 02H
1065Uncore cycles that at least one core is unhalted and all L3 ways are
1066enabled.
1067.It Li CYCLES_UNHALTED_L3_FLL_DISABLE
1068.Pq Event 86H , Umask 01H
1069Uncore cycles that at least one core is unhalted and all L3 ways are
1070disabled.
1071.El
1072.Sh SEE ALSO
1073.Xr pmc 3 ,
1074.Xr pmc.atom 3 ,
1075.Xr pmc.core 3 ,
1076.Xr pmc.corei7 3 ,
1077.Xr pmc.corei7uc 3 ,
1078.Xr pmc.iaf 3 ,
1079.Xr pmc.k7 3 ,
1080.Xr pmc.k8 3 ,
1081.Xr pmc.soft 3 ,
1082.Xr pmc.tsc 3 ,
1083.Xr pmc.ucf 3 ,
1084.Xr pmc.westmere 3 ,
1085.Xr pmc_cpuinfo 3 ,
1086.Xr pmclog 3 ,
1087.Xr hwpmc 4
1088.Sh HISTORY
1089The
1090.Nm pmc
1091library first appeared in
1092.Fx 6.0 .
1093.Sh AUTHORS
1094The
1095.Lb libpmc
1096library was written by
1097.An Joseph Koshy Aq Mt jkoshy@FreeBSD.org .
1098