Message ID | 20200507081436.49071-1-irogers@google.com |
---|---|
Headers | show |
Series | Share events between metrics | expand |
On Thu, May 07, 2020 at 01:14:29AM -0700, Ian Rogers wrote: > Metric groups contain metrics. Metrics create groups of events to > ideally be scheduled together. Often metrics refer to the same events, > for example, a cache hit and cache miss rate. Using separate event > groups means these metrics are multiplexed at different times and the > counts don't sum to 100%. More multiplexing also decreases the > accuracy of the measurement. > > This change orders metrics from groups or the command line, so that > the ones with the most events are set up first. Later metrics see if > groups already provide their events, and reuse them if > possible. Unnecessary events and groups are eliminated. > > RFC because: > - without this change events within a metric may get scheduled > together, after they may appear as part of a larger group and be > multiplexed at different times, lowering accuracy - however, less > multiplexing may compensate for this. > - libbpf's hashmap is used, however, libbpf is an optional > requirement for building perf. > - other things I'm not thinking of. hi, I can't apply this, what branch/commit is this based on? Applying: perf expr: migrate expr ids table to libbpf's hashmap error: patch failed: tools/perf/tests/pmu-events.c:428 error: tools/perf/tests/pmu-events.c: patch does not apply error: patch failed: tools/perf/util/expr.h:2 error: tools/perf/util/expr.h: patch does not apply error: patch failed: tools/perf/util/expr.y:73 error: tools/perf/util/expr.y: patch does not apply Patch failed at 0001 perf expr: migrate expr ids table to libbpf's hashmap thanks, jirka > > Thanks! > > Ian Rogers (7): > perf expr: migrate expr ids table to libbpf's hashmap > perf metricgroup: change evlist_used to a bitmap > perf metricgroup: free metric_events on error > perf metricgroup: always place duration_time last > perf metricgroup: delay events string creation > perf metricgroup: order event groups by size > perf metricgroup: remove duped metric group events > > tools/perf/tests/expr.c | 32 ++--- > tools/perf/tests/pmu-events.c | 22 ++-- > tools/perf/util/expr.c | 125 ++++++++++-------- > tools/perf/util/expr.h | 22 ++-- > tools/perf/util/expr.y | 22 +--- > tools/perf/util/metricgroup.c | 242 +++++++++++++++++++++------------- > tools/perf/util/stat-shadow.c | 46 ++++--- > 7 files changed, 280 insertions(+), 231 deletions(-) > > -- > 2.26.2.526.g744177e7f7-goog >
On Thu, May 7, 2020 at 6:49 AM Jiri Olsa <jolsa@redhat.com> wrote: > > On Thu, May 07, 2020 at 01:14:29AM -0700, Ian Rogers wrote: > > Metric groups contain metrics. Metrics create groups of events to > > ideally be scheduled together. Often metrics refer to the same events, > > for example, a cache hit and cache miss rate. Using separate event > > groups means these metrics are multiplexed at different times and the > > counts don't sum to 100%. More multiplexing also decreases the > > accuracy of the measurement. > > > > This change orders metrics from groups or the command line, so that > > the ones with the most events are set up first. Later metrics see if > > groups already provide their events, and reuse them if > > possible. Unnecessary events and groups are eliminated. > > > > RFC because: > > - without this change events within a metric may get scheduled > > together, after they may appear as part of a larger group and be > > multiplexed at different times, lowering accuracy - however, less > > multiplexing may compensate for this. > > - libbpf's hashmap is used, however, libbpf is an optional > > requirement for building perf. > > - other things I'm not thinking of. > > hi, > I can't apply this, what branch/commit is this based on? > > Applying: perf expr: migrate expr ids table to libbpf's hashmap > error: patch failed: tools/perf/tests/pmu-events.c:428 > error: tools/perf/tests/pmu-events.c: patch does not apply > error: patch failed: tools/perf/util/expr.h:2 > error: tools/perf/util/expr.h: patch does not apply > error: patch failed: tools/perf/util/expr.y:73 > error: tools/perf/util/expr.y: patch does not apply > Patch failed at 0001 perf expr: migrate expr ids table to libbpf's hashmap > > thanks, > jirka Thanks for trying! I have resent the entire patch series here: https://lore.kernel.org/lkml/20200507140819.126960-1-irogers@google.com/ It is acme's perf/core tree + metric fix/test CLs + some minor fixes. Details in the cover letter. Thanks, Ian > > > > Thanks! > > > > Ian Rogers (7): > > perf expr: migrate expr ids table to libbpf's hashmap > > perf metricgroup: change evlist_used to a bitmap > > perf metricgroup: free metric_events on error > > perf metricgroup: always place duration_time last > > perf metricgroup: delay events string creation > > perf metricgroup: order event groups by size > > perf metricgroup: remove duped metric group events > > > > tools/perf/tests/expr.c | 32 ++--- > > tools/perf/tests/pmu-events.c | 22 ++-- > > tools/perf/util/expr.c | 125 ++++++++++-------- > > tools/perf/util/expr.h | 22 ++-- > > tools/perf/util/expr.y | 22 +--- > > tools/perf/util/metricgroup.c | 242 +++++++++++++++++++++------------- > > tools/perf/util/stat-shadow.c | 46 ++++--- > > 7 files changed, 280 insertions(+), 231 deletions(-) > > > > -- > > 2.26.2.526.g744177e7f7-goog > > >
On Thu, May 07, 2020 at 01:14:29AM -0700, Ian Rogers wrote: > Metric groups contain metrics. Metrics create groups of events to > ideally be scheduled together. Often metrics refer to the same events, > for example, a cache hit and cache miss rate. Using separate event > groups means these metrics are multiplexed at different times and the > counts don't sum to 100%. More multiplexing also decreases the > accuracy of the measurement. > > This change orders metrics from groups or the command line, so that > the ones with the most events are set up first. Later metrics see if > groups already provide their events, and reuse them if > possible. Unnecessary events and groups are eliminated. Note this actually may make multiplexing errors worse. For metrics it is often important that all the input values to the metric run at the same time. So e.g. if you have two metrics and they each fit into a group, but not together, even though you have more multiplexing it will give more accurate results for each metric. I think you change can make sense for metrics that don't fit into single groups anyways. perf currently doesn't quite know this but some heuristic could be added. But I wouldn't do it for simple metrics that fit into groups. The result may well be worse. My toplev tool has some heuristics for this, also some more sophisticated ones that tracks subexpressions. That would be far too complicated for perf likely. -Andi
On Thu, May 7, 2020 at 10:48 AM Andi Kleen <ak@linux.intel.com> wrote: > > On Thu, May 07, 2020 at 01:14:29AM -0700, Ian Rogers wrote: > > Metric groups contain metrics. Metrics create groups of events to > > ideally be scheduled together. Often metrics refer to the same events, > > for example, a cache hit and cache miss rate. Using separate event > > groups means these metrics are multiplexed at different times and the > > counts don't sum to 100%. More multiplexing also decreases the > > accuracy of the measurement. > > > > This change orders metrics from groups or the command line, so that > > the ones with the most events are set up first. Later metrics see if > > groups already provide their events, and reuse them if > > possible. Unnecessary events and groups are eliminated. > > Note this actually may make multiplexing errors worse. > > For metrics it is often important that all the input values to > the metric run at the same time. > > So e.g. if you have two metrics and they each fit into a group, > but not together, even though you have more multiplexing it > will give more accurate results for each metric. > > I think you change can make sense for metrics that don't fit > into single groups anyways. perf currently doesn't quite know > this but some heuristic could be added. > > But I wouldn't do it for simple metrics that fit into groups. > The result may well be worse. > > My toplev tool has some heuristics for this, also some more > sophisticated ones that tracks subexpressions. That would > be far too complicated for perf likely. > > -Andi Thanks Andi! I was trying to be mindful of the multiplexing issue in the description: > - without this change events within a metric may get scheduled > together, after they may appear as part of a larger group and be > multiplexed at different times, lowering accuracy - however, less > multiplexing may compensate for this. I agree the heuristic in this patch set is naive and would welcome to improve it from your toplev experience. I think this change is progress on TopDownL1 - would you agree? I'm wondering if what is needed are flags to control behavior. For example, avoiding the use of groups altogether. For TopDownL1 I see. Currently: 27,294,614,172 idq_uops_not_delivered.core # 0.3 Frontend_Bound (49.96%) 24,498,363,923 cycles (49.96%) 21,449,143,854 uops_issued.any # 0.1 Bad_Speculation (16.68%) 16,450,676,961 uops_retired.retire_slots (16.68%) 880,423,103 int_misc.recovery_cycles (16.68%) 24,180,775,568 cycles (16.68%) 27,662,201,567 idq_uops_not_delivered.core # 0.5 Backend_Bound (16.67%) 25,354,331,290 cycles (16.67%) 22,642,218,398 uops_issued.any (16.67%) 17,564,211,383 uops_retired.retire_slots (16.67%) 896,938,527 int_misc.recovery_cycles (16.67%) 17,872,454,517 uops_retired.retire_slots # 0.2 Retiring (16.68%) 25,122,100,836 cycles (16.68%) 15,101,167,608 inst_retired.any # 0.6 IPC (33.34%) 24,977,816,793 cpu_clk_unhalted.thread (33.34%) 24,868,717,684 cycles # 99474870736.0 SLOTS (49.98%) With proposed (RFC) sharing of events over metrics: 22,780,823,620 cycles # 91123294480.0 SLOTS # 0.2 Retiring # 0.3 Frontend_Bound # 0.1 Bad_Speculation # 0.4 Backend_Bound (50.01%) 26,097,362,439 idq_uops_not_delivered.core (50.01%) 790,521,504 int_misc.recovery_cycles (50.01%) 21,025,308,329 uops_issued.any (50.01%) 17,041,506,149 uops_retired.retire_slots (50.01%) 22,964,891,526 cpu_clk_unhalted.thread # 0.6 IPC (49.99%) 14,531,724,741 inst_retired.any (49.99%) No groups: 1,517,455,258 cycles # 6069821032.0 SLOTS # 0.1 Retiring # 0.3 Frontend_Bound # 0.1 Bad_Speculation # 0.5 Backend_Bound (85.64%) 1,943,047,724 idq_uops_not_delivered.core (85.61%) 54,257,713 int_misc.recovery_cycles (85.63%) 1,050,787,137 uops_issued.any (85.63%) 881,310,530 uops_retired.retire_slots (85.68%) 1,553,561,836 cpu_clk_unhalted.thread # 0.5 IPC (71.81%) 742,087,439 inst_retired.any (85.85%) So with no groups there is a lot less multiplexing. So I'm thinking of two flags: - disable sharing of events between metrics - defaulted off - this keeps the current behavior in case there is a use-case where multiplexing is detrimental. I'm not sure how necessary this flag is, if we could quantify it based on experience elsewhere it'd be nice. Default off as without sharing metrics within a metric group fail to add to 100%. Fwiw, I can imagine phony metrics that exist just to cause sharing of events within a group. - disable grouping of events in metrics - defaulted off - this would change the behavior of groups like TopDownL1 as I show above for "no groups". I see in toplev: https://github.com/andikleen/pmu-tools/wiki/toplev-manual --no-group which is similar to the second flag. Do you have any pointers in toplev for better grouping heuristics? Thoughts and better ways to do this very much appreciated! Thanks, Ian
> > - without this change events within a metric may get scheduled > > together, after they may appear as part of a larger group and be > > multiplexed at different times, lowering accuracy - however, less > > multiplexing may compensate for this. > > I agree the heuristic in this patch set is naive and would welcome to > improve it from your toplev experience. I think this change is > progress on TopDownL1 - would you agree? TopdownL1 in non SMT mode should always fit. Inside a group deduping always makes sense. The problem is SMT mode where it doesn't fit. toplev tries to group each node and each level together. > > I'm wondering if what is needed are flags to control behavior. For > example, avoiding the use of groups altogether. For TopDownL1 I see. Yes the current situation isn't great. For Topdown your patch clearly is an improvement, I'm not sure it's for everything though. Probably the advanced heuristics are only useful for a few formulas, most are very simple. So maybe it's ok. I guess would need some testing over the existing formulas. -Andi
On Thu, May 7, 2020 at 2:47 PM Andi Kleen <ak@linux.intel.com> wrote: > > > > - without this change events within a metric may get scheduled > > > together, after they may appear as part of a larger group and be > > > multiplexed at different times, lowering accuracy - however, less > > > multiplexing may compensate for this. > > > > I agree the heuristic in this patch set is naive and would welcome to > > improve it from your toplev experience. I think this change is > > progress on TopDownL1 - would you agree? > > TopdownL1 in non SMT mode should always fit. Inside a group > deduping always makes sense. > > The problem is SMT mode where it doesn't fit. toplev tries > to group each node and each level together. Thanks Andi, I've provided some examples of TopDownL3_SMT in the cover letter of the v3 patch set: https://lore.kernel.org/lkml/20200508053629.210324-1-irogers@google.com/ I tested sandybridge and cascadelake and the results look similar to the non-SMT version. Let me know if there's a different variant to test. > > > > I'm wondering if what is needed are flags to control behavior. For > > example, avoiding the use of groups altogether. For TopDownL1 I see. > > Yes the current situation isn't great. > > For Topdown your patch clearly is an improvement, I'm not sure > it's for everything though. > > Probably the advanced heuristics are only useful for a few > formulas, most are very simple. So maybe it's ok. I guess > would need some testing over the existing formulas. Agreed, do you have a pointer on a metric group where things would obviously be worse? I started off with a cache miss and hit rate metric and similar to topdown this approach is a benefit. In v3 I've added a --metric-no-merge option to retain existing grouping behavior, I've also added a --metric-no-group that avoids groups for all metrics. This may be useful if the NMI watchdog can't be disabled. Thanks for the input! Ian > -Andi