其他分享
首页 > 其他分享> > 哪些性能事件可以使用PEBS?

哪些性能事件可以使用PEBS?

作者:互联网

我想了解哪些事件可以在我的事件上使用精确修饰符
CPU(桑迪桥).

英特尔软件开发人员手册(表18-32.PEBS性能
英特尔微体系结构代码名称Sandy Bridge的事件包含
仅以下事件:INST_RETIRED,UOPS_RETIRED,
BR_INST_RETIRED,BR_MISP_RETIRED,MEM_UOPS_RETIRED,
MEM_LOAD_UOPS_RETIRED,MEM_LOAD_UOPS_LLC_HIT_RETIRED.并且SandyBridge_core_V15.json用PEBS>列出了相同的事件. 0.

但是,有some examples个使用perf,它们将:p添加到cycles事件中.而且我可以在我的机器上成功运行perf record -e cycle:p.

同样,perf record -e cycles:p -vv-sleep 1打印精确的IP1.那么这是否意味着CPU_CLK_UNHALTED事件实际上使用了PEBS?

是否有可能获得支持:p的事件的完整列表?

解决方法:

在SandyBridge上有一些技巧可以支持cycles:p,而CPU_CLK_UNHALTED.*没有PEBS. hack是在intel_pebs_aliases_snb()中的perf内核部分中实现的.当用户使用非零精确修饰符请求-e周期,即PERF_COUNT_HW_CPU_CYCLES(转换为CPU_CLK_UNHALTED.CORE)时,此函数将使用PEBS将硬件事件更改为UOPS_RETIRED.ALL.

  29    [PERF_COUNT_HW_CPU_CYCLES]      = 0x003c,

2739 static void intel_pebs_aliases_snb(struct perf_event *event)
2740 {
2741    if ((event->hw.config & X86_RAW_EVENT_MASK) == 0x003c) {
2742        /*
2743         * Use an alternative encoding for CPU_CLK_UNHALTED.THREAD_P
2744         * (0x003c) so that we can use it with PEBS.
2745         *
2746         * The regular CPU_CLK_UNHALTED.THREAD_P event (0x003c) isn't
2747         * PEBS capable. However we can use UOPS_RETIRED.ALL
2748         * (0x01c2), which is a PEBS capable event, to get the same
2749         * count.
2750         *
2751         * UOPS_RETIRED.ALL counts the number of cycles that retires
2752         * CNTMASK micro-ops. By setting CNTMASK to a value (16)
2753         * larger than the maximum number of micro-ops that can be
2754         * retired per cycle (4) and then inverting the condition, we
2755         * count all cycles that retire 16 or less micro-ops, which
2756         * is every cycle.
2757         *
2758         * Thereby we gain a PEBS capable cycle counter.
2759         */
2760        u64 alt_config = X86_CONFIG(.event=0xc2, .umask=0x01, .inv=1, .cmask=16);
2761 
2762        alt_config |= (event->hw.config & ~X86_RAW_EVENT_MASK);
2763        event->hw.config = alt_config;
2764    }
2765 }

intel_pebs_aliases_snb hack已在案例案例INTEL_FAM6_SANDYBRIDGE:/案例案例INTEL_FAM6_SANDYBRIDGE_X:中注册为3557 __init int intel_pmu_init(void)

3772        x86_pmu.event_constraints = intel_snb_event_constraints;
3773        x86_pmu.pebs_constraints = intel_snb_pebs_event_constraints;
3774        x86_pmu.pebs_aliases = intel_pebs_aliases_snb;

当precision_ip设置为非零时,从intel_pmu_hw_config()调用pebs_aliases:

2814 static int intel_pmu_hw_config(struct perf_event *event)
2815 {

2821    if (event->attr.precise_ip) {

2828        if (x86_pmu.pebs_aliases)
2829            x86_pmu.pebs_aliases(event);
2830    }

该黑客攻击于2012年实施,lkml线程“ [PATCH] perf,x86:使周期:p在SNB上工作”,“ [tip:perf / core] perf / x86:为SNB / IVB实施周期:p”,cccb9ba9e4ee0d750265f53de9258df69655c40b, http://git.kernel.org/cgit/linux/kernel/git/tip/tip.git/commit/?id=cccb9ba9e4ee0d750265f53de9258df69655c40b

perf/x86: Implement cycles:p for SNB/IVB

Now that there’s finally a chip with working PEBS (IvyBridge), we can
enable the hardware and implement cycles:p for SNB/IVB.

而且我认为,除了arch/x86/events/intel/core.c中的Linux源代码,用于静态void intel_pebs_aliases的grep(通常实现了cycle:p / CPU_CLK_UNHALTED 0x003c的grep)之外,没有完整的此类“精确”转换hack列表,并检查intel_pmu_init的实际模型和确切的x86_pmu.选择的pebs_aliases变体:

> intel_pebs_aliases_core2,INST_RETIRED.ANY_P(0x00c0)CNTMASK = 16而不是周期:p
> intel_pebs_aliases_snb,UOPS_RETIRED.ALL(0x01c2)CNTMASK = 16而不是cycles:p
> intel_pebs_aliases_precdist可获取Precision_ip,INST_RETIRED.PREC_DIST(0x01c0)的最大值,而不是周期:SKL,IVB,HSW,BDW上的ppp

标签:performance,intel,performancecounter,perf,linux
来源: https://codeday.me/bug/20191111/2021926.html