[PATCH] oprofile: Fixes for Skylake event lists
Andi Kleen <[email protected]>
| Newsgroups | gmane.linux.oprofile |
|---|---|
| Message-ID | <[email protected]> |
From: Andi Kleen <[email protected]> This fixes the review feedback for the Skylake event list. - Fix event codes for INST_RETIRED, CPU_CLK_UNHALTED. - Fix OFFCORE_REQUESTS_OUTSTANDING events - Add br_inst_retired.all_branches_pebs - Fill in correct default event --- events/i386/skylake/events | 4 ++-- events/i386/skylake/unit_masks | 25 +++++++++++++------------ libop/op_events.c | 5 ++++- 3 files changed, 19 insertions(+), 15 deletions(-) diff --git a/events/i386/skylake/events b/events/i386/skylake/events index 28d6654..9a04a86 100644 --- a/events/i386/skylake/events +++ b/events/i386/skylake/events @@ -6,8 +6,6 @@ # Note the minimum counts are not discovered experimentally and could be likely # lowered in many cases without ill effect. # -event:0x00 counters:1 um:inst_retired minimum:2000003 name:inst_retired : -event:0x00 counters:cpuid um:cpu_clk_unhalted minimum:2000003 name:cpu_clk_unhalted : event:0x03 counters:cpuid um:ld_blocks minimum:100003 name:ld_blocks : event:0x07 counters:cpuid um:ld_blocks_partial minimum:100003 name:ld_blocks_partial_address_alias : event:0x08 counters:cpuid um:dtlb_load_misses minimum:2000003 name:dtlb_load_misses : @@ -16,6 +14,7 @@ event:0x0e counters:cpuid um:uops_issued minimum:2000003 name:uops_issued : event:0x14 counters:cpuid um:arith minimum:2000003 name:arith_divider_active : event:0x24 counters:cpuid um:l2_rqsts minimum:200003 name:l2_rqsts : event:0x2e counters:cpuid um:longest_lat_cache minimum:100003 name:longest_lat_cache : +event:0x3c counters:cpuid um:cpu_clk_unhalted minimum:2000003 name:cpu_clk_unhalted : event:0x3c counters:cpuid um:cpu_clk_thread_unhalted minimum:2000003 name:cpu_clk_thread_unhalted : event:0x48 counters:cpuid um:l1d_pend_miss minimum:2000003 name:l1d_pend_miss : event:0x49 counters:cpuid um:dtlb_store_misses minimum:2000003 name:dtlb_store_misses : @@ -44,6 +43,7 @@ event:0xb0 counters:cpuid um:offcore_requests minimum:100003 name:offcore_reques event:0xb1 counters:cpuid um:uops_executed minimum:2000003 name:uops_executed : event:0xb2 counters:cpuid um:offcore_requests_buffer minimum:2000003 name:offcore_requests_buffer_sq_full : event:0xbd counters:cpuid um:tlb_flush minimum:100007 name:tlb_flush : +event:0xc0 counters:1 um:inst_retired minimum:2000003 name:inst_retired : event:0xc1 counters:cpuid um:other_assists minimum:100003 name:other_assists_any : event:0xc2 counters:cpuid um:uops_retired minimum:2000003 name:uops_retired : event:0xc3 counters:cpuid um:machine_clears minimum:100003 name:machine_clears : diff --git a/events/i386/skylake/unit_masks b/events/i386/skylake/unit_masks index 98ed65c..b505769 100644 --- a/events/i386/skylake/unit_masks +++ b/events/i386/skylake/unit_masks @@ -37,16 +37,6 @@ name:offcore_requests_buffer type:mandatory default:0x1 0x1 extra: sq_full Offcore requests buffer cannot take more entries for this thread core. name:other_assists type:mandatory default:0x3f 0x3f extra: any Number of times a microcode assist is invoked by HW other than FP-assist. Examples include AD (page Access Dirty) and AVX* related assists. -name:inst_retired type:exclusive default:any - 0x1 extra: any Instructions retired from execution.mem - 0x0 extra: any_p Number of instructions retired. General Counter - architectural event - 0x1 extra:pebs prec_dist Precise instruction retired event with HW to reduce effect of PEBS shadow in IP distribution -name:cpu_clk_unhalted type:exclusive default:thread - 0x2 extra: thread Core cycles when the thread is not in halt state - 0x3 extra: ref_tsc Reference cycles when the core is not in halt state. - 0x0 extra: thread_p Thread cycles when thread is not in halt state - 0x2 extra:any thread_any Core cycles when at least one thread on the physical core is not in halt state - 0x0 extra:any thread_p_any Core cycles when at least one thread on the physical core is not in halt state name:ld_blocks type:exclusive default:0x2 0x2 extra: store_forward loads blocked by overlapping with store buffer that cannot be forwarded . 0x8 extra: no_sr The number of times that split load operations are temporarily blocked because all resources for handling the split accesses are in use @@ -85,6 +75,12 @@ name:l2_rqsts type:exclusive default:0x21 name:longest_lat_cache type:exclusive default:0x41 0x41 extra: miss Core-originated cacheable demand requests missed L3 0x4f extra: reference Core-originated cacheable demand requests that refer to L3 +name:cpu_clk_unhalted type:exclusive default:thread + 0x2 extra: thread Core cycles when the thread is not in halt state + 0x3 extra: ref_tsc Reference cycles when the core is not in halt state. + 0x0 extra: thread_p Thread cycles when thread is not in halt state + 0x2 extra:any thread_any Core cycles when at least one thread on the physical core is not in halt state + 0x0 extra:any thread_p_any Core cycles when at least one thread on the physical core is not in halt state name:cpu_clk_thread_unhalted type:exclusive default:ref_xclk 0x1 extra: ref_xclk Reference cycles when the thread is unhalted (counts at 100 MHz rate) 0x2 extra: one_thread_active Count XClk pulses when this thread is unhalted and the other thread is halted. @@ -119,8 +115,8 @@ name:rs_events type:exclusive default:empty_cycles 0x1 extra:cmask=1,inv,edge empty_end Counts end of periods where the Reservation Station (RS) was empty. Could be useful to precisely locate Frontend Latency Bound issues. name:offcore_requests_outstanding type:exclusive default:demand_data_rd 0x1 extra: demand_data_rd Offcore outstanding Demand Data Read transactions in uncore queue. - 0x2 extra:cmask=1 demand_code_rd Cycles with offcore outstanding Code Reads transactions in the SuperQueue (SQ), queue to uncore, every cycle. - 0x4 extra:cmask=1 demand_rfo Offcore outstanding demand rfo reads transactions in SuperQueue (SQ), queue to uncore, every cycle + 0x2 extra: demand_code_rd Offcore outstanding Code Reads transactions in the SuperQueue (SQ), queue to uncore. + 0x4 extra: demand_rfo Offcore outstanding demand rfo reads transactions in SuperQueue (SQ), queue to uncore, every cycle 0x8 extra: all_data_rd Offcore outstanding cacheable Core Data Read transactions in SuperQueue (SQ), queue to uncore 0x10 extra: l3_miss_demand_data_rd Counts number of Offcore outstanding Demand Data Read requests who miss L3 cache in the superQ every cycle. 0x1 extra:cmask=1 cycles_with_demand_data_rd Cycles when offcore outstanding Demand Data Read transactions are present in SuperQueue (SQ), queue to uncore @@ -217,6 +213,10 @@ name:uops_executed type:exclusive default:thread name:tlb_flush type:exclusive default:0x1 0x1 extra: dtlb_thread DTLB flush attempts of the thread-specific entries 0x20 extra: stlb_any STLB flush attempts +name:inst_retired type:exclusive default:any + 0x1 extra: any Instructions retired from execution.mem + 0x0 extra: any_p Number of instructions retired. General Counter - architectural event + 0x1 extra:pebs prec_dist Precise instruction retired event with HW to reduce effect of PEBS shadow in IP distribution name:uops_retired type:exclusive default:retire_slots 0x2 extra: retire_slots Retirement slots used. 0x1 extra:cmask=1,inv stall_cycles Cycles without actually retired uops. @@ -231,6 +231,7 @@ name:br_inst_retired type:exclusive default:all_branches 0x1 extra:pebs conditional_pebs Conditional branch instructions retired. 0x2 extra: near_call Direct and indirect near call instructions retired. 0x2 extra:pebs near_call_pebs Direct and indirect near call instructions retired. + 0x0 extra:pebs all_branches_pebs All (macro) branch instructions retired. 0x8 extra: near_return Return instructions retired. 0x8 extra:pebs near_return_pebs Return instructions retired. 0x10 extra: not_taken Not taken branch instructions retired. diff --git a/libop/op_events.c b/libop/op_events.c index f58d243..25f010e 100644 --- a/libop/op_events.c +++ b/libop/op_events.c @@ -1200,7 +1200,6 @@ void op_default_event(op_cpu cpu_type, struct op_default_event_descr * descr) case CPU_NEHALEM: case CPU_HASWELL: case CPU_BROADWELL: - case CPU_SKYLAKE: case CPU_SILVERMONT: case CPU_WESTMERE: case CPU_SANDYBRIDGE: @@ -1213,6 +1212,10 @@ void op_default_event(op_cpu cpu_type, struct op_default_event_descr * descr) descr->name = "CPU_CLK_UNHALTED"; break; + case CPU_SKYLAKE: + descr->name = "cpu_clk_unhalted"; + break; + case CPU_P4: case CPU_P4_HT2: descr->name = "GLOBAL_POWER_EVENTS"; -- 2.4.3 ------------------------------------------------------------------------------ Don't Limit Your Business. Reach for the Cloud. GigeNET's Cloud Solutions provide you with the tools and support that you need to offload your IT needs and focus on growing your business. Configured For All Businesses. Start Your Cloud Today. https://www.gigenetcloud.com/