{"id":76066,"date":"2026-02-13T23:25:44","date_gmt":"2026-02-13T15:25:44","guid":{"rendered":"https:\/\/www.wsisp.com\/helps\/76066.html"},"modified":"2026-02-13T23:25:44","modified_gmt":"2026-02-13T15:25:44","slug":"linux-%e6%80%a7%e8%83%bd%e5%ae%9e%e6%88%98-%e7%ac%ac-11-%e7%af%87-%e6%80%a7%e8%83%bd%e8%ae%a1%e6%95%b0%e5%99%a8%e4%b8%8e-perf-%e5%b7%a5%e5%85%b7%e5%85%a5%e9%97%a8","status":"publish","type":"post","link":"https:\/\/www.wsisp.com\/helps\/76066.html","title":{"rendered":"Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8"},"content":{"rendered":"<h2>Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8 &#x1f50d;<\/h2>\n<h3>&#x1f517; \u4ece&#034;\u731c\u6d4b&#034;\u5230&#034;\u6d4b\u91cf&#034;&#xff1a;\u6027\u80fd\u4f18\u5316\u7684\u8303\u5f0f\u8f6c\u53d8<\/h3>\n<p>\u5728\u524d\u9762\u7684\u7ae0\u8282\u4e2d&#xff0c;\u6211\u4eec\u6df1\u5165\u63a2\u8ba8\u4e86 CPU \u7f13\u5b58\u3001\u5185\u5b58\u5ef6\u8fdf\u3001\u8f6f\u4e2d\u65ad\u7b49\u6027\u80fd\u95ee\u9898\u3002\u5728\u7b2c\u5341\u7ae0\u4e2d&#xff0c;\u6211\u4eec\u53d1\u73b0\u70b9\u4e91\u5904\u7406\u7b97\u6cd5\u5b58\u5728\u9ad8\u8fbe 15% \u7684\u7f13\u5b58\u672a\u547d\u4e2d\u7387&#xff0c;\u5e76\u901a\u8fc7\u4f18\u5316\u6570\u636e\u7ed3\u6784\u5c06\u5904\u7406\u5ef6\u8fdf\u4ece 45ms \u964d\u4f4e\u5230 9.8ms\u3002<\/p>\n<p>\u4f46\u4f60\u53ef\u80fd\u4f1a\u6709\u7591\u95ee&#xff1a;\u6211\u4eec\u662f\u5982\u4f55\u7cbe\u786e\u77e5\u9053\u7f13\u5b58\u672a\u547d\u4e2d\u7387\u7684&#xff1f;\u5982\u4f55\u5b9a\u4f4d\u5230\u5177\u4f53\u662f\u54ea\u4e2a\u51fd\u6570\u3001\u751a\u81f3\u54ea\u4e00\u884c\u4ee3\u7801\u5bfc\u81f4\u7684\u6027\u80fd\u95ee\u9898&#xff1f;<\/p>\n<p>\u7b54\u6848\u5c31\u662f CPU \u6027\u80fd\u8ba1\u6570\u5668&#xff08;Performance Monitoring Unit, PMU&#xff09; \u548c\u57fa\u4e8e\u5b83\u7684\u5f3a\u5927\u5de5\u5177 perf\u3002<\/p>\n<p>\u5728\u6ca1\u6709\u6027\u80fd\u8ba1\u6570\u5668\u7684\u65f6\u4ee3&#xff0c;\u7a0b\u5e8f\u5458\u53ea\u80fd\u901a\u8fc7\u5728\u4ee3\u7801\u4e2d\u63d2\u5165\u8ba1\u65f6\u5668\u3001\u8f93\u51fa\u65e5\u5fd7\u6765&#034;\u731c\u6d4b&#034;\u6027\u80fd\u74f6\u9888\u3002\u8fd9\u79cd\u65b9\u6cd5\u4e0d\u4ec5\u7e41\u7410\u3001\u4fb5\u5165\u6027\u5f3a&#xff0c;\u800c\u4e14\u4f1a\u4e25\u91cd\u5e72\u6270\u7a0b\u5e8f\u7684\u6b63\u5e38\u6267\u884c\u3002\u6027\u80fd\u8ba1\u6570\u5668\u7684\u51fa\u73b0&#xff0c;\u8ba9\u6211\u4eec\u80fd\u591f\u4ee5\u51e0\u4e4e\u96f6\u5f00\u9500\u7684\u65b9\u5f0f&#xff0c;\u7cbe\u786e\u6d4b\u91cf\u7a0b\u5e8f\u8fd0\u884c\u65f6\u7684\u5404\u79cd\u786c\u4ef6\u4e8b\u4ef6\u2014\u2014CPU \u5468\u671f\u6570\u3001\u6307\u4ee4\u6570\u3001\u7f13\u5b58\u547d\u4e2d\/\u672a\u547d\u4e2d\u3001\u5206\u652f\u9884\u6d4b\u6210\u529f\/\u5931\u8d25\u7b49\u3002<\/p>\n<p>\u6027\u80fd\u4f18\u5316\u7684\u9ec4\u91d1\u6cd5\u5219&#xff1a;\u6ca1\u6709\u6d4b\u91cf&#xff0c;\u5c31\u6ca1\u6709\u4f18\u5316\u3002\u4f60\u65e0\u6cd5\u4f18\u5316\u4f60\u65e0\u6cd5\u6d4b\u91cf\u7684\u4e1c\u897f\u3002<\/p>\n<hr \/>\n<h3>&#x1f914; CPU \u6027\u80fd\u8ba1\u6570\u5668&#xff08;PMU&#xff09;&#xff1a;\u786c\u4ef6\u7ea7\u7684&#034;\u9ed1\u5323\u5b50&#034;<\/h3>\n<h4>1. PMU \u7684\u5de5\u4f5c\u539f\u7406<\/h4>\n<p>\u73b0\u4ee3 CPU \u5185\u90e8\u96c6\u6210\u4e86\u4e00\u7ec4\u4e13\u95e8\u7684\u786c\u4ef6\u8ba1\u6570\u5668&#xff0c;\u5b83\u4eec\u80fd\u591f\u5728\u7a0b\u5e8f\u6267\u884c\u671f\u95f4&#xff0c;\u900f\u660e\u5730\u8bb0\u5f55\u5404\u79cd\u786c\u4ef6\u4e8b\u4ef6\u7684\u53d1\u751f\u6b21\u6570\u3002\u8fd9\u4e9b\u8ba1\u6570\u5668\u901a\u5e38\u662f 48 \u4f4d\u6216 64 \u4f4d\u7684\u5bc4\u5b58\u5668&#xff0c;\u6bcf\u5f53\u7279\u5b9a\u4e8b\u4ef6\u53d1\u751f\u65f6&#xff08;\u5982\u6307\u4ee4\u9000\u4f11\u3001\u7f13\u5b58\u672a\u547d\u4e2d\u3001\u5206\u652f\u9884\u6d4b\u5931\u8d25&#xff09;&#xff0c;\u5bf9\u5e94\u7684\u8ba1\u6570\u5668\u5c31\u4f1a\u81ea\u52a8\u52a0 1\u3002<\/p>\n<p>#mermaid-svg-VusfpqB0ZLF4taBs{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;fill:#333;}@keyframes edge-animation-frame{from{stroke-dashoffset:0;}}@keyframes dash{to{stroke-dashoffset:0;}}#mermaid-svg-VusfpqB0ZLF4taBs .edge-animation-slow{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 50s linear infinite;stroke-linecap:round;}#mermaid-svg-VusfpqB0ZLF4taBs .edge-animation-fast{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 20s linear infinite;stroke-linecap:round;}#mermaid-svg-VusfpqB0ZLF4taBs .error-icon{fill:#552222;}#mermaid-svg-VusfpqB0ZLF4taBs .error-text{fill:#552222;stroke:#552222;}#mermaid-svg-VusfpqB0ZLF4taBs .edge-thickness-normal{stroke-width:1px;}#mermaid-svg-VusfpqB0ZLF4taBs .edge-thickness-thick{stroke-width:3.5px;}#mermaid-svg-VusfpqB0ZLF4taBs .edge-pattern-solid{stroke-dasharray:0;}#mermaid-svg-VusfpqB0ZLF4taBs .edge-thickness-invisible{stroke-width:0;fill:none;}#mermaid-svg-VusfpqB0ZLF4taBs .edge-pattern-dashed{stroke-dasharray:3;}#mermaid-svg-VusfpqB0ZLF4taBs .edge-pattern-dotted{stroke-dasharray:2;}#mermaid-svg-VusfpqB0ZLF4taBs .marker{fill:#333333;stroke:#333333;}#mermaid-svg-VusfpqB0ZLF4taBs .marker.cross{stroke:#333333;}#mermaid-svg-VusfpqB0ZLF4taBs svg{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;}#mermaid-svg-VusfpqB0ZLF4taBs p{margin:0;}#mermaid-svg-VusfpqB0ZLF4taBs .label{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;color:#333;}#mermaid-svg-VusfpqB0ZLF4taBs .cluster-label text{fill:#333;}#mermaid-svg-VusfpqB0ZLF4taBs .cluster-label span{color:#333;}#mermaid-svg-VusfpqB0ZLF4taBs .cluster-label span p{background-color:transparent;}#mermaid-svg-VusfpqB0ZLF4taBs .label text,#mermaid-svg-VusfpqB0ZLF4taBs span{fill:#333;color:#333;}#mermaid-svg-VusfpqB0ZLF4taBs .node rect,#mermaid-svg-VusfpqB0ZLF4taBs .node circle,#mermaid-svg-VusfpqB0ZLF4taBs .node ellipse,#mermaid-svg-VusfpqB0ZLF4taBs .node polygon,#mermaid-svg-VusfpqB0ZLF4taBs .node path{fill:#ECECFF;stroke:#9370DB;stroke-width:1px;}#mermaid-svg-VusfpqB0ZLF4taBs .rough-node .label text,#mermaid-svg-VusfpqB0ZLF4taBs .node .label text,#mermaid-svg-VusfpqB0ZLF4taBs .image-shape .label,#mermaid-svg-VusfpqB0ZLF4taBs .icon-shape .label{text-anchor:middle;}#mermaid-svg-VusfpqB0ZLF4taBs .node .katex path{fill:#000;stroke:#000;stroke-width:1px;}#mermaid-svg-VusfpqB0ZLF4taBs .rough-node .label,#mermaid-svg-VusfpqB0ZLF4taBs .node .label,#mermaid-svg-VusfpqB0ZLF4taBs .image-shape .label,#mermaid-svg-VusfpqB0ZLF4taBs .icon-shape .label{text-align:center;}#mermaid-svg-VusfpqB0ZLF4taBs .node.clickable{cursor:pointer;}#mermaid-svg-VusfpqB0ZLF4taBs .root .anchor path{fill:#333333!important;stroke-width:0;stroke:#333333;}#mermaid-svg-VusfpqB0ZLF4taBs .arrowheadPath{fill:#333333;}#mermaid-svg-VusfpqB0ZLF4taBs .edgePath .path{stroke:#333333;stroke-width:2.0px;}#mermaid-svg-VusfpqB0ZLF4taBs .flowchart-link{stroke:#333333;fill:none;}#mermaid-svg-VusfpqB0ZLF4taBs .edgeLabel{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-VusfpqB0ZLF4taBs .edgeLabel p{background-color:rgba(232,232,232, 0.8);}#mermaid-svg-VusfpqB0ZLF4taBs .edgeLabel rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-VusfpqB0ZLF4taBs .labelBkg{background-color:rgba(232, 232, 232, 0.5);}#mermaid-svg-VusfpqB0ZLF4taBs .cluster rect{fill:#ffffde;stroke:#aaaa33;stroke-width:1px;}#mermaid-svg-VusfpqB0ZLF4taBs .cluster text{fill:#333;}#mermaid-svg-VusfpqB0ZLF4taBs .cluster span{color:#333;}#mermaid-svg-VusfpqB0ZLF4taBs div.mermaidTooltip{position:absolute;text-align:center;max-width:200px;padding:2px;font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:12px;background:hsl(80, 100%, 96.2745098039%);border:1px solid #aaaa33;border-radius:2px;pointer-events:none;z-index:100;}#mermaid-svg-VusfpqB0ZLF4taBs .flowchartTitleText{text-anchor:middle;font-size:18px;fill:#333;}#mermaid-svg-VusfpqB0ZLF4taBs rect.text{fill:none;stroke-width:0;}#mermaid-svg-VusfpqB0ZLF4taBs .icon-shape,#mermaid-svg-VusfpqB0ZLF4taBs .image-shape{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-VusfpqB0ZLF4taBs .icon-shape p,#mermaid-svg-VusfpqB0ZLF4taBs .image-shape p{background-color:rgba(232,232,232, 0.8);padding:2px;}#mermaid-svg-VusfpqB0ZLF4taBs .icon-shape rect,#mermaid-svg-VusfpqB0ZLF4taBs .image-shape rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-VusfpqB0ZLF4taBs .label-icon{display:inline-block;height:1em;overflow:visible;vertical-align:-0.125em;}#mermaid-svg-VusfpqB0ZLF4taBs .node .label-icon path{fill:currentColor;stroke:revert;stroke-width:revert;}#mermaid-svg-VusfpqB0ZLF4taBs :root{&#8211;mermaid-font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;}<span class=\"nodeLabel\"><\/p>\n<p>CPU \u6838\u5fc3<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u6027\u80fd\u76d1\u63a7\u5355\u5143 PMU<\/p>\n<p><\/span><span class=\"edgeLabel\"><\/p>\n<p>\u6bcf\u6761\u6307\u4ee4\u6267\u884c<\/p>\n<p><\/span><span class=\"edgeLabel\"><\/p>\n<p>\u6bcf\u4e2a\u65f6\u949f\u5468\u671f<\/p>\n<p><\/span><span class=\"edgeLabel\"><\/p>\n<p>\u672a\u547d\u4e2d\u4e8b\u4ef6<\/p>\n<p><\/span><span class=\"edgeLabel\"><\/p>\n<p>\u9884\u6d4b\u5931\u8d25<\/p>\n<p><\/span><span class=\"edgeLabel\"><\/p>\n<p>\u4e2d\u65ad\/\u91c7\u6837<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u7b97\u672f\u903b\u8f91\u5355\u5143ALU<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>L1\/L2 \u7f13\u5b58<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u5206\u652f\u9884\u6d4b\u5668<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u8ba1\u6570\u5668 0\u6307\u4ee4\u6570<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u8ba1\u6570\u5668 1CPU \u5468\u671f<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u8ba1\u6570\u5668 2\u7f13\u5b58\u672a\u547d\u4e2d<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u8ba1\u6570\u5668 3\u5206\u652f\u9884\u6d4b\u5931\u8d25<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u64cd\u4f5c\u7cfb\u7edfperf \u5de5\u5177<\/p>\n<p><\/span><\/p>\n<p>\u5173\u952e\u7279\u6027&#xff1a;<\/p>\n<ul>\n<li>\u786c\u4ef6\u7ea7\u522b&#xff1a;\u8ba1\u6570\u5668\u5185\u7f6e\u4e8e CPU&#xff0c;\u65e0\u9700\u8f6f\u4ef6\u5e72\u9884&#xff0c;\u5f00\u9500\u6781\u4f4e&#xff08;\u901a\u5e38 &lt; 1%&#xff09;<\/li>\n<li>\u5e76\u884c\u8fd0\u884c&#xff1a;\u591a\u4e2a\u8ba1\u6570\u5668\u53ef\u4ee5\u540c\u65f6\u5de5\u4f5c&#xff0c;\u540c\u65f6\u76d1\u63a7\u591a\u79cd\u4e8b\u4ef6<\/li>\n<li>\u91c7\u6837\u673a\u5236&#xff1a;\u53ef\u4ee5\u914d\u7f6e\u4e3a\u6bcf N \u4e2a\u4e8b\u4ef6\u89e6\u53d1\u4e00\u6b21\u4e2d\u65ad&#xff0c;\u8bb0\u5f55\u5f53\u65f6\u7684\u7a0b\u5e8f\u4e0a\u4e0b\u6587&#xff08;PC\u3001\u8c03\u7528\u6808&#xff09;<\/li>\n<\/ul>\n<h4>2. \u6027\u80fd\u8ba1\u6570\u5668\u7684\u7c7b\u578b<\/h4>\n<p>Intel\u3001AMD\u3001ARM \u7b49\u5904\u7406\u5668\u90fd\u63d0\u4f9b\u4e86\u6027\u80fd\u8ba1\u6570\u5668&#xff0c;\u4f46\u5177\u4f53\u652f\u6301\u7684\u4e8b\u4ef6\u7565\u6709\u4e0d\u540c\u3002\u4ee5 Intel \u5904\u7406\u5668\u4e3a\u4f8b&#xff1a;<\/p>\n<table>\n<tr>\u4e8b\u4ef6\u7c7b\u578b\u793a\u4f8b\u4e8b\u4ef6\u542b\u4e49<\/tr>\n<tbody>\n<tr>\n<td>\u57fa\u7840\u4e8b\u4ef6<\/td>\n<td>cpu-cycles<\/td>\n<td>CPU \u65f6\u949f\u5468\u671f\u6570<\/td>\n<\/tr>\n<tr>\n<td><\/td>\n<td>instructions<\/td>\n<td>\u6267\u884c\u7684\u6307\u4ee4\u6570<\/td>\n<\/tr>\n<tr>\n<td>\u7f13\u5b58\u4e8b\u4ef6<\/td>\n<td>cache-references<\/td>\n<td>\u7f13\u5b58\u8bbf\u95ee\u603b\u6b21\u6570<\/td>\n<\/tr>\n<tr>\n<td><\/td>\n<td>cache-misses<\/td>\n<td>\u7f13\u5b58\u672a\u547d\u4e2d\u6b21\u6570<\/td>\n<\/tr>\n<tr>\n<td><\/td>\n<td>LLC-loads<\/td>\n<td>\u6700\u540e\u4e00\u7ea7\u7f13\u5b58&#xff08;L3&#xff09;\u52a0\u8f7d\u6b21\u6570<\/td>\n<\/tr>\n<tr>\n<td><\/td>\n<td>LLC-load-misses<\/td>\n<td>L3 \u52a0\u8f7d\u672a\u547d\u4e2d\u6b21\u6570<\/td>\n<\/tr>\n<tr>\n<td>\u5206\u652f\u4e8b\u4ef6<\/td>\n<td>branches<\/td>\n<td>\u5206\u652f\u6307\u4ee4\u6570<\/td>\n<\/tr>\n<tr>\n<td><\/td>\n<td>branch-misses<\/td>\n<td>\u5206\u652f\u9884\u6d4b\u5931\u8d25\u6b21\u6570<\/td>\n<\/tr>\n<tr>\n<td>\u5185\u5b58\u4e8b\u4ef6<\/td>\n<td>dTLB-loads<\/td>\n<td>\u6570\u636e TLB \u8bbf\u95ee\u6b21\u6570<\/td>\n<\/tr>\n<tr>\n<td><\/td>\n<td>dTLB-load-misses<\/td>\n<td>\u6570\u636e TLB \u672a\u547d\u4e2d\u6b21\u6570<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u81ea\u52a8\u9a7e\u9a76\u573a\u666f&#xff1a;<\/p>\n<ul>\n<li>\u70b9\u4e91\u5904\u7406&#xff1a;\u5173\u6ce8 cache-misses\u3001LLC-load-misses&#xff08;\u5185\u5b58\u5bc6\u96c6\u578b&#xff09;<\/li>\n<li>\u76ee\u6807\u68c0\u6d4b&#xff1a;\u5173\u6ce8 instructions\u3001branch-misses&#xff08;\u8ba1\u7b97\u5bc6\u96c6\u578b\u3001\u5206\u652f\u591a&#xff09;<\/li>\n<li>\u8def\u5f84\u89c4\u5212&#xff1a;\u5173\u6ce8 cycles\u3001IPC&#xff08;\u7b97\u6cd5\u590d\u6742\u5ea6&#xff09;<\/li>\n<\/ul>\n<hr \/>\n<h3>&#x1f6e0;\ufe0f perf \u5de5\u5177\u4e09\u5251\u5ba2&#xff1a;top\u3001stat\u3001record<\/h3>\n<p>perf \u662f Linux \u5185\u6838\u81ea\u5e26\u7684\u6027\u80fd\u5206\u6790\u5de5\u5177&#xff0c;\u5b83\u662f PMU \u7684&#034;\u7528\u6237\u754c\u9762&#034;\u3002\u638c\u63e1 perf \u7684\u4e09\u4e2a\u6838\u5fc3\u547d\u4ee4&#xff0c;\u5c31\u80fd\u89e3\u51b3 90% \u7684\u6027\u80fd\u95ee\u9898\u3002<\/p>\n<h4>1. perf top&#xff1a;\u5b9e\u65f6\u70ed\u70b9\u96f7\u8fbe &#x1f3af;<\/h4>\n<p>perf top \u7c7b\u4f3c\u4e8e top \u547d\u4ee4&#xff0c;\u4f46\u5b83\u663e\u793a\u7684\u662f\u51fd\u6570\u7ea7\u522b\u7684 CPU \u5360\u7528&#xff0c;\u5b9e\u65f6\u66f4\u65b0\u3002<\/p>\n<p><span class=\"token comment\"># \u542f\u52a8\u5b9e\u65f6\u70ed\u70b9\u5206\u6790<\/span><br \/>\nperf <span class=\"token function\">top<\/span><\/p>\n<p><span class=\"token comment\"># \u6307\u5b9a\u91c7\u6837\u4e8b\u4ef6&#xff08;\u9ed8\u8ba4\u662f cpu-cycles&#xff09;<\/span><br \/>\nperf <span class=\"token function\">top<\/span> -e cache-misses<\/p>\n<p><span class=\"token comment\"># \u5206\u6790\u7279\u5b9a\u8fdb\u7a0b<\/span><br \/>\nperf <span class=\"token function\">top<\/span> -p <span class=\"token operator\">&lt;<\/span>PID<span class=\"token operator\">&gt;<\/span><\/p>\n<p><span class=\"token comment\"># \u663e\u793a\u8c03\u7528\u56fe&#xff08;\u9700\u8981\u7a0b\u5e8f\u5e26\u8c03\u8bd5\u7b26\u53f7&#xff09;<\/span><br \/>\nperf <span class=\"token function\">top<\/span> -g<\/p>\n<p>\u8f93\u51fa\u793a\u4f8b&#xff08;\u81ea\u52a8\u9a7e\u9a76\u76ee\u6807\u68c0\u6d4b\u7a0b\u5e8f&#xff09;&#xff1a;<\/p>\n<p>Samples: 58K of event &#039;cpu-cycles&#039;, 4000 Hz, Event count (approx.): 14528394672<br \/>\nOverhead  Shared Object       Symbol<br \/>\n  15.23%  object_detector     [.] NonMaxSuppression<br \/>\n  12.87%  libopencv_core.so   [.] cv::Mat::copyTo<br \/>\n   8.45%  libc-2.31.so        [.] __memcpy_avx_unaligned<br \/>\n   6.32%  object_detector     [.] ComputeIoU<br \/>\n   5.91%  libopencv_dnn.so    [.] cv::dnn::Net::forward<br \/>\n   &#8230;<\/p>\n<p>\u89e3\u8bfb&#xff1a;<\/p>\n<ul>\n<li>NonMaxSuppression \u51fd\u6570\u5360\u7528\u4e86 15.23% \u7684 CPU \u5468\u671f&#xff0c;\u662f\u6700\u70ed\u7684\u70ed\u70b9<\/li>\n<li>cv::Mat::copyTo \u5360\u7528 12.87%&#xff0c;\u8bf4\u660e\u5b58\u5728\u5927\u91cf\u7684\u77e9\u9635\u62f7\u8d1d\u64cd\u4f5c<\/li>\n<li>__memcpy_avx_unaligned \u5360\u7528 8.45%&#xff0c;\u8fd9\u662f libc \u7684\u5185\u5b58\u62f7\u8d1d\u51fd\u6570&#xff0c;\u8fdb\u4e00\u6b65\u8bc1\u5b9e\u4e86\u62f7\u8d1d\u5f00\u9500<\/li>\n<\/ul>\n<p>\u4f18\u5316\u601d\u8def&#xff1a;<\/p>\n<li>\u51cf\u5c11\u4e0d\u5fc5\u8981\u7684\u77e9\u9635\u62f7\u8d1d&#xff08;\u4f7f\u7528 ROI\u3001\u5171\u4eab\u5185\u5b58&#xff09;<\/li>\n<li>\u4f18\u5316 NMS \u7b97\u6cd5&#xff08;\u4f7f\u7528\u7a7a\u95f4\u7d22\u5f15\u3001GPU \u52a0\u901f&#xff09;<\/li>\n<h4>2. perf stat&#xff1a;\u6027\u80fd\u4e8b\u4ef6\u7edf\u8ba1\u4eea\u8868\u76d8 &#x1f4ca;<\/h4>\n<p>perf stat \u8fd0\u884c\u7a0b\u5e8f\u5e76\u6536\u96c6\u6027\u80fd\u8ba1\u6570\u5668\u7684\u7edf\u8ba1\u6570\u636e&#xff0c;\u5728\u7a0b\u5e8f\u7ed3\u675f\u540e\u8f93\u51fa\u6c47\u603b\u62a5\u544a\u3002<\/p>\n<p><span class=\"token comment\"># \u57fa\u7840\u7edf\u8ba1&#xff08;\u9ed8\u8ba4\u4e8b\u4ef6\u96c6&#xff09;<\/span><br \/>\nperf <span class=\"token function\">stat<\/span> .\/my_program<\/p>\n<p><span class=\"token comment\"># \u6307\u5b9a\u4e8b\u4ef6<\/span><br \/>\nperf <span class=\"token function\">stat<\/span> -e cycles,instructions,cache-misses,branch-misses .\/my_program<\/p>\n<p><span class=\"token comment\"># \u5206\u6790\u8fd0\u884c\u4e2d\u7684\u8fdb\u7a0b&#xff08;\u6301\u7eed 10 \u79d2&#xff09;<\/span><br \/>\nperf <span class=\"token function\">stat<\/span> -p <span class=\"token operator\">&lt;<\/span>PID<span class=\"token operator\">&gt;<\/span> <span class=\"token function\">sleep<\/span> <span class=\"token number\">10<\/span><\/p>\n<p><span class=\"token comment\"># \u591a\u6838\u7edf\u8ba1&#xff08;\u663e\u793a\u6bcf\u4e2a CPU \u6838\u5fc3\u7684\u6570\u636e&#xff09;<\/span><br \/>\nperf <span class=\"token function\">stat<\/span> -a -A <span class=\"token function\">sleep<\/span> <span class=\"token number\">5<\/span><\/p>\n<p>\u8f93\u51fa\u793a\u4f8b&#xff08;\u70b9\u4e91\u6ee4\u6ce2\u7a0b\u5e8f&#xff09;&#xff1a;<\/p>\n<p>perf <span class=\"token function\">stat<\/span> -e cycles,instructions,cache-references,cache-misses,branches,branch-misses <span class=\"token punctuation\">\\\\<\/span><br \/>\n    .\/voxel_filter &#8211;input lidar_frame.pcd<\/p>\n<p> Performance counter stats for &#039;.\/voxel_filter &#8211;input lidar_frame.pcd&#039;:<\/p>\n<p>     3,456,789,012      cycles                    #    3.457 GHz<br \/>\n     5,678,901,234      instructions              #    1.64  insn per cycle<br \/>\n       456,789,123      cache-references          #  132.123 M\/sec<br \/>\n        68,513,867      cache-misses              #   15.00 % of all cache refs<br \/>\n       890,123,456      branches                  #  257.456 M\/sec<br \/>\n        44,506,172      branch-misses             #    5.00 % of all branches<\/p>\n<p>       1.001234567 seconds time elapsed<br \/>\n       0.987654321 seconds user<br \/>\n       0.013580246 seconds sys<\/p>\n<p>\u5173\u952e\u6307\u6807\u89e3\u8bfb&#xff1a;<\/p>\n<li>\n<p>IPC (Instructions Per Cycle) &#061; 1.64<\/p>\n<p>IPC &#061; instructions \/ cycles &#061; 5,678,901,234 \/ 3,456,789,012 \u2248 1.64<\/p>\n<ul>\n<li>\u7406\u60f3\u503c&#xff1a;\u73b0\u4ee3 CPU \u7684\u7406\u8bba IPC \u53ef\u8fbe 4-6&#xff08;\u8d85\u6807\u91cf\u3001\u4e71\u5e8f\u6267\u884c&#xff09;<\/li>\n<li>\u5b9e\u9645\u503c 1.64&#xff1a;\u8bf4\u660e CPU \u6709\u5927\u91cf\u65f6\u95f4\u5728\u7b49\u5f85&#xff08;\u5185\u5b58\u3001\u7f13\u5b58\u672a\u547d\u4e2d&#xff09;<\/li>\n<li>\u4f18\u5316\u76ee\u6807&#xff1a;\u63d0\u5347 IPC \u5230 2.5&#043;<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u7f13\u5b58\u672a\u547d\u4e2d\u7387 &#061; 15%<\/p>\n<p>Cache Miss Rate &#061; cache-misses \/ cache-references &#061; 15%<\/p>\n<ul>\n<li>\u6b63\u5e38\u503c&#xff1a;&lt; 5%<\/li>\n<li>15% \u8bf4\u660e&#xff1a;\u6570\u636e\u8bbf\u95ee\u6a21\u5f0f\u975e\u5e38\u4e0d\u53cb\u597d&#xff0c;\u5927\u91cf\u968f\u673a\u8bbf\u95ee<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u5206\u652f\u9884\u6d4b\u5931\u8d25\u7387 &#061; 5%<\/p>\n<p>Branch Miss Rate &#061; branch-misses \/ branches &#061; 5%<\/p>\n<ul>\n<li>\u6b63\u5e38\u503c&#xff1a;&lt; 2%<\/li>\n<li>5% \u8bf4\u660e&#xff1a;\u5b58\u5728\u5927\u91cf\u96be\u4ee5\u9884\u6d4b\u7684\u6761\u4ef6\u5206\u652f<\/li>\n<\/ul>\n<\/li>\n<h4>3. perf record&#xff1a;\u6027\u80fd\u4e8b\u4ef6\u91c7\u6837\u8bb0\u5f55\u4eea &#x1f3ac;<\/h4>\n<p>perf record \u4ee5\u91c7\u6837\u7684\u65b9\u5f0f\u8bb0\u5f55\u6027\u80fd\u4e8b\u4ef6&#xff0c;\u5e76\u4fdd\u5b58\u5230 perf.data \u6587\u4ef6\u4e2d&#xff0c;\u4e4b\u540e\u53ef\u4ee5\u7528 perf report \u8fdb\u884c\u79bb\u7ebf\u5206\u6790\u3002<\/p>\n<p><span class=\"token comment\"># \u8bb0\u5f55 CPU \u5468\u671f\u4e8b\u4ef6&#xff08;\u9ed8\u8ba4&#xff09;<\/span><br \/>\nperf record .\/my_program<\/p>\n<p><span class=\"token comment\"># \u8bb0\u5f55\u7f13\u5b58\u672a\u547d\u4e2d\u4e8b\u4ef6<\/span><br \/>\nperf record -e LLC-load-misses -c <span class=\"token number\">10000<\/span> .\/my_program<br \/>\n<span class=\"token comment\"># -c 10000: \u6bcf 10000 \u6b21\u4e8b\u4ef6\u89e6\u53d1\u4e00\u6b21\u91c7\u6837<\/span><\/p>\n<p><span class=\"token comment\"># \u8bb0\u5f55\u8c03\u7528\u6808&#xff08;-g \u4f7f\u7528\u5e27\u6307\u9488&#xff0c;-F \u8bbe\u7f6e\u91c7\u6837\u9891\u7387&#xff09;<\/span><br \/>\nperf record -g -F <span class=\"token number\">99<\/span> .\/my_program<br \/>\n<span class=\"token comment\"># -F 99: \u6bcf\u79d2\u91c7\u6837 99 \u6b21&#xff08;\u907f\u514d\u4e0e\u5b9a\u65f6\u5668\u51b2\u7a81&#xff09;<\/span><\/p>\n<p><span class=\"token comment\"># \u8bb0\u5f55\u8fd0\u884c\u4e2d\u7684\u8fdb\u7a0b<\/span><br \/>\nperf record -p <span class=\"token operator\">&lt;<\/span>PID<span class=\"token operator\">&gt;<\/span> -g <span class=\"token function\">sleep<\/span> <span class=\"token number\">30<\/span><\/p>\n<p><span class=\"token comment\"># \u751f\u6210\u62a5\u544a<\/span><br \/>\nperf report<\/p>\n<p>\u5de5\u4f5c\u539f\u7406&#xff1a;<\/p>\n<p>perf record\u5185\u6838\u6027\u80fd\u8ba1\u6570\u5668\u76ee\u6807\u68c0\u6d4b\u7a0b\u5e8fperf record\u5185\u6838\u6027\u80fd\u8ba1\u6570\u5668\u76ee\u6807\u68c0\u6d4b\u7a0b\u5e8f#mermaid-svg-OqDviGaRcaOR9jZW{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;fill:#333;}@keyframes edge-animation-frame{from{stroke-dashoffset:0;}}@keyframes dash{to{stroke-dashoffset:0;}}#mermaid-svg-OqDviGaRcaOR9jZW .edge-animation-slow{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 50s linear infinite;stroke-linecap:round;}#mermaid-svg-OqDviGaRcaOR9jZW .edge-animation-fast{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 20s linear infinite;stroke-linecap:round;}#mermaid-svg-OqDviGaRcaOR9jZW .error-icon{fill:#552222;}#mermaid-svg-OqDviGaRcaOR9jZW .error-text{fill:#552222;stroke:#552222;}#mermaid-svg-OqDviGaRcaOR9jZW .edge-thickness-normal{stroke-width:1px;}#mermaid-svg-OqDviGaRcaOR9jZW .edge-thickness-thick{stroke-width:3.5px;}#mermaid-svg-OqDviGaRcaOR9jZW .edge-pattern-solid{stroke-dasharray:0;}#mermaid-svg-OqDviGaRcaOR9jZW .edge-thickness-invisible{stroke-width:0;fill:none;}#mermaid-svg-OqDviGaRcaOR9jZW .edge-pattern-dashed{stroke-dasharray:3;}#mermaid-svg-OqDviGaRcaOR9jZW .edge-pattern-dotted{stroke-dasharray:2;}#mermaid-svg-OqDviGaRcaOR9jZW .marker{fill:#333333;stroke:#333333;}#mermaid-svg-OqDviGaRcaOR9jZW .marker.cross{stroke:#333333;}#mermaid-svg-OqDviGaRcaOR9jZW svg{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;}#mermaid-svg-OqDviGaRcaOR9jZW p{margin:0;}#mermaid-svg-OqDviGaRcaOR9jZW .actor{stroke:hsl(259.6261682243, 59.7765363128%, 87.9019607843%);fill:#ECECFF;}#mermaid-svg-OqDviGaRcaOR9jZW text.actor&gt;tspan{fill:black;stroke:none;}#mermaid-svg-OqDviGaRcaOR9jZW .actor-line{stroke:hsl(259.6261682243, 59.7765363128%, 87.9019607843%);}#mermaid-svg-OqDviGaRcaOR9jZW .innerArc{stroke-width:1.5;stroke-dasharray:none;}#mermaid-svg-OqDviGaRcaOR9jZW .messageLine0{stroke-width:1.5;stroke-dasharray:none;stroke:#333;}#mermaid-svg-OqDviGaRcaOR9jZW .messageLine1{stroke-width:1.5;stroke-dasharray:2,2;stroke:#333;}#mermaid-svg-OqDviGaRcaOR9jZW #arrowhead path{fill:#333;stroke:#333;}#mermaid-svg-OqDviGaRcaOR9jZW .sequenceNumber{fill:white;}#mermaid-svg-OqDviGaRcaOR9jZW #sequencenumber{fill:#333;}#mermaid-svg-OqDviGaRcaOR9jZW #crosshead path{fill:#333;stroke:#333;}#mermaid-svg-OqDviGaRcaOR9jZW .messageText{fill:#333;stroke:none;}#mermaid-svg-OqDviGaRcaOR9jZW .labelBox{stroke:hsl(259.6261682243, 59.7765363128%, 87.9019607843%);fill:#ECECFF;}#mermaid-svg-OqDviGaRcaOR9jZW .labelText,#mermaid-svg-OqDviGaRcaOR9jZW .labelText&gt;tspan{fill:black;stroke:none;}#mermaid-svg-OqDviGaRcaOR9jZW .loopText,#mermaid-svg-OqDviGaRcaOR9jZW .loopText&gt;tspan{fill:black;stroke:none;}#mermaid-svg-OqDviGaRcaOR9jZW .loopLine{stroke-width:2px;stroke-dasharray:2,2;stroke:hsl(259.6261682243, 59.7765363128%, 87.9019607843%);fill:hsl(259.6261682243, 59.7765363128%, 87.9019607843%);}#mermaid-svg-OqDviGaRcaOR9jZW .note{stroke:#aaaa33;fill:#fff5ad;}#mermaid-svg-OqDviGaRcaOR9jZW .noteText,#mermaid-svg-OqDviGaRcaOR9jZW .noteText&gt;tspan{fill:black;stroke:none;}#mermaid-svg-OqDviGaRcaOR9jZW .activation0{fill:#f4f4f4;stroke:#666;}#mermaid-svg-OqDviGaRcaOR9jZW .activation1{fill:#f4f4f4;stroke:#666;}#mermaid-svg-OqDviGaRcaOR9jZW .activation2{fill:#f4f4f4;stroke:#666;}#mermaid-svg-OqDviGaRcaOR9jZW .actorPopupMenu{position:absolute;}#mermaid-svg-OqDviGaRcaOR9jZW .actorPopupMenuPanel{position:absolute;fill:#ECECFF;box-shadow:0px 8px 16px 0px rgba(0,0,0,0.2);filter:drop-shadow(3px 5px 2px rgb(0 0 0 \/ 0.4));}#mermaid-svg-OqDviGaRcaOR9jZW .actor-man line{stroke:hsl(259.6261682243, 59.7765363128%, 87.9019607843%);fill:#ECECFF;}#mermaid-svg-OqDviGaRcaOR9jZW .actor-man circle,#mermaid-svg-OqDviGaRcaOR9jZW line{stroke:hsl(259.6261682243, 59.7765363128%, 87.9019607843%);fill:#ECECFF;stroke-width:2px;}#mermaid-svg-OqDviGaRcaOR9jZW :root{&#8211;mermaid-font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;}alt[\u8ba1\u6570\u5668\u8fbe\u5230\u9608\u503c]loop[\u7a0b\u5e8f\u6267\u884c]\u914d\u7f6e\u91c7\u6837\u4e8b\u4ef6&#xff08;cycles, \u6bcf 4000 \u6b21\u91c7\u6837&#xff09;\u542f\u7528\u8ba1\u6570\u5668\u6267\u884c\u6307\u4ee4\u8ba1\u6570\u5668\u7d2f\u52a0\u89e6\u53d1\u4e2d\u65ad\u8bb0\u5f55 PC\u3001\u8c03\u7528\u6808\u3001\u65f6\u95f4\u6233\u5199\u5165 perf.data\u91cd\u7f6e\u8ba1\u6570\u5668\u7a0b\u5e8f\u7ed3\u675f\u751f\u6210 perf.data<\/p>\n<p>perf report \u8f93\u51fa\u793a\u4f8b&#xff1a;<\/p>\n<p>Samples: 45K of event &#039;cycles&#039;, Event count (approx.): 11250000000<br \/>\nOverhead  Command         Shared Object       Symbol<br \/>\n  18.56%  object_detect   object_detect       [.] NonMaxSuppression<br \/>\n            |<br \/>\n            &#8212;NonMaxSuppression<br \/>\n               |<br \/>\n               |&#8211;12.34% std::sort<br \/>\n               |<br \/>\n               |&#8211;4.23% ComputeIoU<br \/>\n               |<br \/>\n                &#8211;1.99% std::vector::resize<\/p>\n<p>  14.23%  object_detect   libopencv_dnn.so    [.] cv::dnn::LayerNorm::forward<br \/>\n  &#8230;<\/p>\n<p>\u89e3\u8bfb&#xff1a;<\/p>\n<ul>\n<li>NonMaxSuppression \u5360\u7528 18.56% \u7684\u5468\u671f<\/li>\n<li>\u5176\u4e2d 12.34% \u82b1\u5728\u4e86 std::sort \u4e0a&#xff08;\u6392\u5e8f\u68c0\u6d4b\u6846&#xff09;<\/li>\n<li>4.23% \u82b1\u5728\u4e86 ComputeIoU \u4e0a&#xff08;\u8ba1\u7b97 IoU&#xff09;<\/li>\n<\/ul>\n<hr \/>\n<h3>&#x1f52c; \u5b9e\u6218\u6848\u4f8b&#xff1a;\u4f18\u5316\u81ea\u52a8\u9a7e\u9a76\u76ee\u6807\u68c0\u6d4b\u7684 NMS \u7b97\u6cd5<\/h3>\n<h4>1. \u95ee\u9898\u73b0\u8c61<\/h4>\n<p>\u4e00\u4e2a\u57fa\u4e8e YOLO \u7684\u76ee\u6807\u68c0\u6d4b\u6a21\u5757&#xff0c;\u5728\u5904\u7406\u4e00\u5e27 1920\u00d71080 \u7684\u56fe\u50cf\u65f6&#xff0c;NMS&#xff08;\u975e\u6781\u5927\u503c\u6291\u5236&#xff09;\u9636\u6bb5\u8017\u65f6\u8fbe\u5230 35ms&#xff0c;\u8fdc\u8d85\u7cfb\u7edf\u8981\u6c42\u7684 10ms\u3002<\/p>\n<h4>2. \u6027\u80fd\u5256\u6790<\/h4>\n<p>\u6b65\u9aa4 1&#xff1a;\u5feb\u901f\u5b9a\u4f4d\u70ed\u70b9<\/p>\n<p><span class=\"token comment\"># \u8fd0\u884c\u7a0b\u5e8f\u5e76\u5b9e\u65f6\u67e5\u770b\u70ed\u70b9<\/span><br \/>\nperf <span class=\"token function\">top<\/span> -p <span class=\"token variable\"><span class=\"token variable\">$(<\/span>pgrep object_detector<span class=\"token variable\">)<\/span><\/span><\/p>\n<p>\u53d1\u73b0 NonMaxSuppression \u51fd\u6570 CPU \u5360\u7528\u6700\u9ad8&#xff08;18%&#xff09;\u3002<\/p>\n<p>\u6b65\u9aa4 2&#xff1a;\u7edf\u8ba1\u6027\u80fd\u6307\u6807<\/p>\n<p>perf <span class=\"token function\">stat<\/span> -e cycles,instructions,cache-misses,branch-misses <span class=\"token punctuation\">\\\\<\/span><br \/>\n    -p <span class=\"token variable\"><span class=\"token variable\">$(<\/span>pgrep object_detector<span class=\"token variable\">)<\/span><\/span> <span class=\"token function\">sleep<\/span> <span class=\"token number\">10<\/span><\/p>\n<p>\u8f93\u51fa&#xff1a;<\/p>\n<p>     28,456,789,012      cycles<br \/>\n     34,567,890,123      instructions              #    1.21  insn per cycle  \u2b05\ufe0f \u4f4e IPC<br \/>\n      1,234,567,890      cache-misses              #   22.00 % of all cache refs  \u2b05\ufe0f \u9ad8\u7f13\u5b58\u672a\u547d\u4e2d<br \/>\n        890,123,456      branch-misses             #    8.50 % of all branches  \u2b05\ufe0f \u9ad8\u5206\u652f\u9884\u6d4b\u5931\u8d25<\/p>\n<p>\u95ee\u9898\u5b9a\u4f4d&#xff1a;<\/p>\n<ul>\n<li>IPC \u53ea\u6709 1.21&#xff1a;CPU \u5927\u91cf\u65f6\u95f4\u5728\u7b49\u5f85<\/li>\n<li>\u7f13\u5b58\u672a\u547d\u4e2d\u7387 22%&#xff1a;\u5185\u5b58\u8bbf\u95ee\u6a21\u5f0f\u6781\u5dee<\/li>\n<li>\u5206\u652f\u9884\u6d4b\u5931\u8d25\u7387 8.5%&#xff1a;\u6761\u4ef6\u5206\u652f\u96be\u4ee5\u9884\u6d4b<\/li>\n<\/ul>\n<p>\u6b65\u9aa4 3&#xff1a;\u91c7\u6837\u5206\u6790\u8c03\u7528\u6808<\/p>\n<p>perf record -e cache-misses -g -p <span class=\"token variable\"><span class=\"token variable\">$(<\/span>pgrep object_detector<span class=\"token variable\">)<\/span><\/span> <span class=\"token function\">sleep<\/span> <span class=\"token number\">10<\/span><br \/>\nperf report<\/p>\n<p>\u8f93\u51fa\u663e\u793a\u7f13\u5b58\u672a\u547d\u4e2d\u4e3b\u8981\u53d1\u751f\u5728&#xff1a;<\/p>\n<p>  55.23%  object_detect  object_detect  [.] NonMaxSuppression<br \/>\n          |<br \/>\n          &#8212;NonMaxSuppression<br \/>\n             |<br \/>\n             |&#8211;45.67% std::vector::operator[]  \u2b05\ufe0f \u968f\u673a\u8bbf\u95ee\u68c0\u6d4b\u6846\u6570\u7ec4<br \/>\n             |<br \/>\n              &#8211;9.56% ComputeIoU<\/p>\n<p>\u6839\u56e0&#xff1a;\u539f\u59cb NMS \u7b97\u6cd5\u4f7f\u7528\u5d4c\u5957\u5faa\u73af&#xff0c;\u5bf9\u68c0\u6d4b\u6846\u8fdb\u884c\u4e24\u4e24\u6bd4\u8f83&#xff0c;\u6d89\u53ca\u5927\u91cf\u968f\u673a\u8bbf\u95ee&#xff1a;<\/p>\n<p><span class=\"token comment\">\/\/ \u539f\u59cb\u4ee3\u7801&#xff08;\u7f13\u5b58\u4e0d\u53cb\u597d\u3001\u5206\u652f\u591a&#xff09;<\/span><br \/>\nstd<span class=\"token double-colon punctuation\">::<\/span>vector<span class=\"token operator\">&lt;<\/span>BBox<span class=\"token operator\">&gt;<\/span> <span class=\"token function\">NonMaxSuppression<\/span><span class=\"token punctuation\">(<\/span>std<span class=\"token double-colon punctuation\">::<\/span>vector<span class=\"token operator\">&lt;<\/span>BBox<span class=\"token operator\">&gt;<\/span><span class=\"token operator\">&amp;<\/span> boxes<span class=\"token punctuation\">,<\/span> <span class=\"token keyword\">float<\/span> threshold<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n    std<span class=\"token double-colon punctuation\">::<\/span>vector<span class=\"token operator\">&lt;<\/span>BBox<span class=\"token operator\">&gt;<\/span> result<span class=\"token punctuation\">;<\/span><br \/>\n    std<span class=\"token double-colon punctuation\">::<\/span>vector<span class=\"token operator\">&lt;<\/span><span class=\"token keyword\">bool<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token function\">suppressed<\/span><span class=\"token punctuation\">(<\/span>boxes<span class=\"token punctuation\">.<\/span><span class=\"token function\">size<\/span><span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token boolean\">false<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span><\/p>\n<p>    <span class=\"token keyword\">for<\/span> <span class=\"token punctuation\">(<\/span>size_t i <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">0<\/span><span class=\"token punctuation\">;<\/span> i <span class=\"token operator\">&lt;<\/span> boxes<span class=\"token punctuation\">.<\/span><span class=\"token function\">size<\/span><span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span> <span class=\"token operator\">&#043;&#043;<\/span>i<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n        <span class=\"token keyword\">if<\/span> <span class=\"token punctuation\">(<\/span>suppressed<span class=\"token punctuation\">[<\/span>i<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token keyword\">continue<\/span><span class=\"token punctuation\">;<\/span>  <span class=\"token comment\">\/\/ \u2b05\ufe0f \u5206\u652f 1&#xff08;\u96be\u9884\u6d4b&#xff09;<\/span><\/p>\n<p>        result<span class=\"token punctuation\">.<\/span><span class=\"token function\">push_back<\/span><span class=\"token punctuation\">(<\/span>boxes<span class=\"token punctuation\">[<\/span>i<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span><\/p>\n<p>        <span class=\"token keyword\">for<\/span> <span class=\"token punctuation\">(<\/span>size_t j <span class=\"token operator\">&#061;<\/span> i <span class=\"token operator\">&#043;<\/span> <span class=\"token number\">1<\/span><span class=\"token punctuation\">;<\/span> j <span class=\"token operator\">&lt;<\/span> boxes<span class=\"token punctuation\">.<\/span><span class=\"token function\">size<\/span><span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span> <span class=\"token operator\">&#043;&#043;<\/span>j<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n            <span class=\"token keyword\">if<\/span> <span class=\"token punctuation\">(<\/span>suppressed<span class=\"token punctuation\">[<\/span>j<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token keyword\">continue<\/span><span class=\"token punctuation\">;<\/span>  <span class=\"token comment\">\/\/ \u2b05\ufe0f \u5206\u652f 2&#xff08;\u96be\u9884\u6d4b&#xff09;<\/span><\/p>\n<p>            <span class=\"token keyword\">if<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token function\">ComputeIoU<\/span><span class=\"token punctuation\">(<\/span>boxes<span class=\"token punctuation\">[<\/span>i<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">,<\/span> boxes<span class=\"token punctuation\">[<\/span>j<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&gt;<\/span> threshold<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span>  <span class=\"token comment\">\/\/ \u2b05\ufe0f \u968f\u673a\u8bbf\u95ee<\/span><br \/>\n                suppressed<span class=\"token punctuation\">[<\/span>j<span class=\"token punctuation\">]<\/span> <span class=\"token operator\">&#061;<\/span> <span class=\"token boolean\">true<\/span><span class=\"token punctuation\">;<\/span><br \/>\n            <span class=\"token punctuation\">}<\/span><br \/>\n        <span class=\"token punctuation\">}<\/span><br \/>\n    <span class=\"token punctuation\">}<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> result<span class=\"token punctuation\">;<\/span><br \/>\n<span class=\"token punctuation\">}<\/span><\/p>\n<h4>3. \u4f18\u5316\u65b9\u6848<\/h4>\n<p>\u4f18\u5316 1&#xff1a;\u7a7a\u95f4\u5206\u533a\u51cf\u5c11\u6bd4\u8f83\u6b21\u6570&#xff08;\u964d\u4f4e\u5206\u652f\u9884\u6d4b\u5931\u8d25&#xff09;<\/p>\n<p><span class=\"token comment\">\/\/ \u4f7f\u7528\u7f51\u683c\u5206\u533a&#xff0c;\u53ea\u6bd4\u8f83\u76f8\u90bb\u7f51\u683c\u4e2d\u7684\u6846<\/span><br \/>\n<span class=\"token keyword\">struct<\/span> <span class=\"token class-name\">Grid<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n    std<span class=\"token double-colon punctuation\">::<\/span>unordered_map<span class=\"token operator\">&lt;<\/span><span class=\"token keyword\">int<\/span><span class=\"token punctuation\">,<\/span> std<span class=\"token double-colon punctuation\">::<\/span>vector<span class=\"token operator\">&lt;<\/span><span class=\"token keyword\">int<\/span><span class=\"token operator\">&gt;&gt;<\/span> cells<span class=\"token punctuation\">;<\/span><\/p>\n<p>    <span class=\"token keyword\">int<\/span> <span class=\"token function\">GetCellID<\/span><span class=\"token punctuation\">(<\/span><span class=\"token keyword\">const<\/span> BBox<span class=\"token operator\">&amp;<\/span> box<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n        <span class=\"token keyword\">int<\/span> x <span class=\"token operator\">&#061;<\/span> <span class=\"token generic-function\"><span class=\"token function\">static_cast<\/span><span class=\"token generic class-name\"><span class=\"token operator\">&lt;<\/span><span class=\"token keyword\">int<\/span><span class=\"token operator\">&gt;<\/span><\/span><\/span><span class=\"token punctuation\">(<\/span>box<span class=\"token punctuation\">.<\/span>x <span class=\"token operator\">\/<\/span> cell_size<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span><br \/>\n        <span class=\"token keyword\">int<\/span> y <span class=\"token operator\">&#061;<\/span> <span class=\"token generic-function\"><span class=\"token function\">static_cast<\/span><span class=\"token generic class-name\"><span class=\"token operator\">&lt;<\/span><span class=\"token keyword\">int<\/span><span class=\"token operator\">&gt;<\/span><\/span><\/span><span class=\"token punctuation\">(<\/span>box<span class=\"token punctuation\">.<\/span>y <span class=\"token operator\">\/<\/span> cell_size<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span><br \/>\n        <span class=\"token keyword\">return<\/span> y <span class=\"token operator\">*<\/span> grid_width <span class=\"token operator\">&#043;<\/span> x<span class=\"token punctuation\">;<\/span><br \/>\n    <span class=\"token punctuation\">}<\/span><\/p>\n<p>    <span class=\"token keyword\">void<\/span> <span class=\"token function\">Insert<\/span><span class=\"token punctuation\">(<\/span><span class=\"token keyword\">int<\/span> idx<span class=\"token punctuation\">,<\/span> <span class=\"token keyword\">const<\/span> BBox<span class=\"token operator\">&amp;<\/span> box<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n        cells<span class=\"token punctuation\">[<\/span><span class=\"token function\">GetCellID<\/span><span class=\"token punctuation\">(<\/span>box<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">.<\/span><span class=\"token function\">push_back<\/span><span class=\"token punctuation\">(<\/span>idx<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span><br \/>\n    <span class=\"token punctuation\">}<\/span><br \/>\n<span class=\"token punctuation\">}<\/span><span class=\"token punctuation\">;<\/span><\/p>\n<p>std<span class=\"token double-colon punctuation\">::<\/span>vector<span class=\"token operator\">&lt;<\/span>BBox<span class=\"token operator\">&gt;<\/span> <span class=\"token function\">OptimizedNMS<\/span><span class=\"token punctuation\">(<\/span>std<span class=\"token double-colon punctuation\">::<\/span>vector<span class=\"token operator\">&lt;<\/span>BBox<span class=\"token operator\">&gt;<\/span><span class=\"token operator\">&amp;<\/span> boxes<span class=\"token punctuation\">,<\/span> <span class=\"token keyword\">float<\/span> threshold<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n    Grid grid<span class=\"token punctuation\">;<\/span><br \/>\n    <span class=\"token keyword\">for<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token keyword\">int<\/span> i <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">0<\/span><span class=\"token punctuation\">;<\/span> i <span class=\"token operator\">&lt;<\/span> boxes<span class=\"token punctuation\">.<\/span><span class=\"token function\">size<\/span><span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span> <span class=\"token operator\">&#043;&#043;<\/span>i<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n        grid<span class=\"token punctuation\">.<\/span><span class=\"token function\">Insert<\/span><span class=\"token punctuation\">(<\/span>i<span class=\"token punctuation\">,<\/span> boxes<span class=\"token punctuation\">[<\/span>i<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span><br \/>\n    <span class=\"token punctuation\">}<\/span><\/p>\n<p>    std<span class=\"token double-colon punctuation\">::<\/span>vector<span class=\"token operator\">&lt;<\/span>BBox<span class=\"token operator\">&gt;<\/span> result<span class=\"token punctuation\">;<\/span><br \/>\n    std<span class=\"token double-colon punctuation\">::<\/span>vector<span class=\"token operator\">&lt;<\/span><span class=\"token keyword\">bool<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token function\">suppressed<\/span><span class=\"token punctuation\">(<\/span>boxes<span class=\"token punctuation\">.<\/span><span class=\"token function\">size<\/span><span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token boolean\">false<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span><\/p>\n<p>    <span class=\"token keyword\">for<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token keyword\">auto<\/span><span class=\"token operator\">&amp;<\/span> <span class=\"token punctuation\">[<\/span>cell_id<span class=\"token punctuation\">,<\/span> indices<span class=\"token punctuation\">]<\/span> <span class=\"token operator\">:<\/span> grid<span class=\"token punctuation\">.<\/span>cells<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n        <span class=\"token comment\">\/\/ \u53ea\u6bd4\u8f83\u540c\u4e00\u4e2a\u548c\u76f8\u90bb\u7f51\u683c\u7684\u6846<\/span><br \/>\n        <span class=\"token keyword\">for<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token keyword\">int<\/span> i <span class=\"token operator\">:<\/span> indices<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n            <span class=\"token keyword\">if<\/span> <span class=\"token punctuation\">(<\/span>suppressed<span class=\"token punctuation\">[<\/span>i<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token keyword\">continue<\/span><span class=\"token punctuation\">;<\/span><\/p>\n<p>            result<span class=\"token punctuation\">.<\/span><span class=\"token function\">push_back<\/span><span class=\"token punctuation\">(<\/span>boxes<span class=\"token punctuation\">[<\/span>i<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span><\/p>\n<p>            <span class=\"token comment\">\/\/ \u83b7\u53d6\u76f8\u90bb\u7f51\u683c<\/span><br \/>\n            <span class=\"token keyword\">for<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token keyword\">int<\/span> neighbor_cell <span class=\"token operator\">:<\/span> <span class=\"token function\">GetNeighborCells<\/span><span class=\"token punctuation\">(<\/span>cell_id<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n                <span class=\"token keyword\">for<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token keyword\">int<\/span> j <span class=\"token operator\">:<\/span> grid<span class=\"token punctuation\">.<\/span>cells<span class=\"token punctuation\">[<\/span>neighbor_cell<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n                    <span class=\"token keyword\">if<\/span> <span class=\"token punctuation\">(<\/span>i <span class=\"token operator\">&gt;&#061;<\/span> j <span class=\"token operator\">||<\/span> suppressed<span class=\"token punctuation\">[<\/span>j<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token keyword\">continue<\/span><span class=\"token punctuation\">;<\/span><\/p>\n<p>                    <span class=\"token keyword\">if<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token function\">ComputeIoU<\/span><span class=\"token punctuation\">(<\/span>boxes<span class=\"token punctuation\">[<\/span>i<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">,<\/span> boxes<span class=\"token punctuation\">[<\/span>j<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&gt;<\/span> threshold<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n                        suppressed<span class=\"token punctuation\">[<\/span>j<span class=\"token punctuation\">]<\/span> <span class=\"token operator\">&#061;<\/span> <span class=\"token boolean\">true<\/span><span class=\"token punctuation\">;<\/span><br \/>\n                    <span class=\"token punctuation\">}<\/span><br \/>\n                <span class=\"token punctuation\">}<\/span><br \/>\n            <span class=\"token punctuation\">}<\/span><br \/>\n        <span class=\"token punctuation\">}<\/span><br \/>\n    <span class=\"token punctuation\">}<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> result<span class=\"token punctuation\">;<\/span><br \/>\n<span class=\"token punctuation\">}<\/span><\/p>\n<p>\u4f18\u5316 2&#xff1a;SIMD \u52a0\u901f IoU \u8ba1\u7b97&#xff08;\u63d0\u5347 IPC&#xff09;<\/p>\n<p><span class=\"token comment\">\/\/ \u4f7f\u7528 AVX2 \u6307\u4ee4\u96c6\u5e76\u884c\u8ba1\u7b97 IoU<\/span><br \/>\n__m256 <span class=\"token function\">ComputeIoU_AVX2<\/span><span class=\"token punctuation\">(<\/span>__m256 x1<span class=\"token punctuation\">,<\/span> __m256 y1<span class=\"token punctuation\">,<\/span> __m256 w1<span class=\"token punctuation\">,<\/span> __m256 h1<span class=\"token punctuation\">,<\/span><br \/>\n                       __m256 x2<span class=\"token punctuation\">,<\/span> __m256 y2<span class=\"token punctuation\">,<\/span> __m256 w2<span class=\"token punctuation\">,<\/span> __m256 h2<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n    <span class=\"token comment\">\/\/ \u4e00\u6b21\u8ba1\u7b97 8 \u4e2a IoU&#xff08;256 \u4f4d \/ 32 \u4f4d &#061; 8&#xff09;<\/span><br \/>\n    <span class=\"token comment\">\/\/ &#8230; SIMD \u5b9e\u73b0 &#8230;<\/span><br \/>\n<span class=\"token punctuation\">}<\/span><\/p>\n<h4>4. \u4f18\u5316\u6548\u679c<\/h4>\n<p>\u518d\u6b21\u8fd0\u884c perf stat&#xff1a;<\/p>\n<p>perf <span class=\"token function\">stat<\/span> -e cycles,instructions,cache-misses,branch-misses <span class=\"token punctuation\">\\\\<\/span><br \/>\n    -p <span class=\"token variable\"><span class=\"token variable\">$(<\/span>pgrep object_detector_optimized<span class=\"token variable\">)<\/span><\/span> <span class=\"token function\">sleep<\/span> <span class=\"token number\">10<\/span><\/p>\n<p>\u4f18\u5316\u540e&#xff1a;<\/p>\n<p>     12,345,678,901      cycles                    #  \u2b07\ufe0f \u4ece 28B \u964d\u5230 12B&#xff08;\u964d\u4f4e 56%&#xff09;<br \/>\n     30,123,456,789      instructions              #    2.44  insn per cycle  \u2b06\ufe0f \u4ece 1.21 \u63d0\u5347\u5230 2.44<br \/>\n        456,789,012      cache-misses              #    8.00 % of all cache refs  \u2b07\ufe0f \u4ece 22% \u964d\u5230 8%<br \/>\n        234,567,890      branch-misses             #    2.50 % of all branches  \u2b07\ufe0f \u4ece 8.5% \u964d\u5230 2.5%<\/p>\n<p>\u6027\u80fd\u63d0\u5347&#xff1a;<\/p>\n<table>\n<tr>\u6307\u6807\u4f18\u5316\u524d\u4f18\u5316\u540e\u6539\u5584<\/tr>\n<tbody>\n<tr>\n<td>NMS \u5ef6\u8fdf<\/td>\n<td>35ms<\/td>\n<td>8ms<\/td>\n<td>\u2193 77%<\/td>\n<\/tr>\n<tr>\n<td>IPC<\/td>\n<td>1.21<\/td>\n<td>2.44<\/td>\n<td>\u2191 102%<\/td>\n<\/tr>\n<tr>\n<td>\u7f13\u5b58\u672a\u547d\u4e2d\u7387<\/td>\n<td>22%<\/td>\n<td>8%<\/td>\n<td>\u2193 64%<\/td>\n<\/tr>\n<tr>\n<td>\u5206\u652f\u9884\u6d4b\u5931\u8d25\u7387<\/td>\n<td>8.5%<\/td>\n<td>2.5%<\/td>\n<td>\u2193 71%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<hr \/>\n<h3>&#x1f4da; \u6df1\u5165\u7406\u89e3&#xff1a;\u786c\u4ef6\u4e8b\u4ef6 vs \u8f6f\u4ef6\u4e8b\u4ef6 vs \u8ddf\u8e2a\u70b9\u4e8b\u4ef6<\/h3>\n<p>perf \u652f\u6301\u4e09\u7c7b\u4e8b\u4ef6&#xff0c;\u5b83\u4eec\u7684\u6570\u636e\u6765\u6e90\u548c\u5f00\u9500\u5404\u4e0d\u76f8\u540c&#xff1a;<\/p>\n<h4>1. \u786c\u4ef6\u4e8b\u4ef6&#xff08;Hardware Events&#xff09;<\/h4>\n<p>\u6570\u636e\u6765\u6e90&#xff1a;CPU \u6027\u80fd\u8ba1\u6570\u5668&#xff08;PMU&#xff09;<br \/>\n\u5f00\u9500&#xff1a;\u6781\u4f4e&#xff08;&lt; 1%&#xff09;<br \/>\n\u793a\u4f8b&#xff1a;<\/p>\n<p>perf list <span class=\"token operator\">|<\/span> <span class=\"token function\">grep<\/span> Hardware<\/p>\n<p><span class=\"token comment\"># \u8f93\u51fa<\/span><br \/>\n  cpu-cycles OR cycles                <span class=\"token punctuation\">[<\/span>Hardware event<span class=\"token punctuation\">]<\/span><br \/>\n  instructions                        <span class=\"token punctuation\">[<\/span>Hardware event<span class=\"token punctuation\">]<\/span><br \/>\n  cache-references                    <span class=\"token punctuation\">[<\/span>Hardware event<span class=\"token punctuation\">]<\/span><br \/>\n  cache-misses                        <span class=\"token punctuation\">[<\/span>Hardware event<span class=\"token punctuation\">]<\/span><br \/>\n  branch-instructions OR branches     <span class=\"token punctuation\">[<\/span>Hardware event<span class=\"token punctuation\">]<\/span><br \/>\n  branch-misses                       <span class=\"token punctuation\">[<\/span>Hardware event<span class=\"token punctuation\">]<\/span><br \/>\n  bus-cycles                          <span class=\"token punctuation\">[<\/span>Hardware event<span class=\"token punctuation\">]<\/span><\/p>\n<p>\u9002\u7528\u573a\u666f&#xff1a;<\/p>\n<ul>\n<li>CPU \u6548\u7387\u5206\u6790&#xff08;IPC&#xff09;<\/li>\n<li>\u7f13\u5b58\u6027\u80fd\u5206\u6790<\/li>\n<li>\u5206\u652f\u9884\u6d4b\u5206\u6790<\/li>\n<\/ul>\n<h4>2. \u8f6f\u4ef6\u4e8b\u4ef6&#xff08;Software Events&#xff09;<\/h4>\n<p>\u6570\u636e\u6765\u6e90&#xff1a;\u64cd\u4f5c\u7cfb\u7edf\u5185\u6838\u7edf\u8ba1<br \/>\n\u5f00\u9500&#xff1a;\u4f4e&#xff08;1-3%&#xff09;<br \/>\n\u793a\u4f8b&#xff1a;<\/p>\n<p>perf list <span class=\"token operator\">|<\/span> <span class=\"token function\">grep<\/span> Software<\/p>\n<p><span class=\"token comment\"># \u8f93\u51fa<\/span><br \/>\n  cpu-clock                           <span class=\"token punctuation\">[<\/span>Software event<span class=\"token punctuation\">]<\/span><br \/>\n  task-clock                          <span class=\"token punctuation\">[<\/span>Software event<span class=\"token punctuation\">]<\/span><br \/>\n  page-faults OR faults               <span class=\"token punctuation\">[<\/span>Software event<span class=\"token punctuation\">]<\/span><br \/>\n  context-switches OR cs              <span class=\"token punctuation\">[<\/span>Software event<span class=\"token punctuation\">]<\/span><br \/>\n  cpu-migrations                      <span class=\"token punctuation\">[<\/span>Software event<span class=\"token punctuation\">]<\/span><br \/>\n  minor-faults                        <span class=\"token punctuation\">[<\/span>Software event<span class=\"token punctuation\">]<\/span><br \/>\n  major-faults                        <span class=\"token punctuation\">[<\/span>Software event<span class=\"token punctuation\">]<\/span><\/p>\n<p>\u9002\u7528\u573a\u666f&#xff1a;<\/p>\n<ul>\n<li>\u4e0a\u4e0b\u6587\u5207\u6362\u5206\u6790<\/li>\n<li>\u9875\u9519\u8bef\u5206\u6790<\/li>\n<li>CPU \u8fc1\u79fb\u5206\u6790<\/li>\n<\/ul>\n<h4>3. \u8ddf\u8e2a\u70b9\u4e8b\u4ef6&#xff08;Tracepoint Events&#xff09;<\/h4>\n<p>\u6570\u636e\u6765\u6e90&#xff1a;\u5185\u6838 ftrace \u6846\u67b6<br \/>\n\u5f00\u9500&#xff1a;\u4e2d\u7b49&#xff08;3-10%&#xff0c;\u53d6\u51b3\u4e8e\u4e8b\u4ef6\u9891\u7387&#xff09;<br \/>\n\u793a\u4f8b&#xff1a;<\/p>\n<p>perf list <span class=\"token operator\">|<\/span> <span class=\"token function\">grep<\/span> sched:<\/p>\n<p><span class=\"token comment\"># \u8f93\u51fa<\/span><br \/>\n  sched:sched_switch              <span class=\"token punctuation\">[<\/span>Tracepoint event<span class=\"token punctuation\">]<\/span><br \/>\n  sched:sched_wakeup              <span class=\"token punctuation\">[<\/span>Tracepoint event<span class=\"token punctuation\">]<\/span><br \/>\n  sched:sched_process_fork        <span class=\"token punctuation\">[<\/span>Tracepoint event<span class=\"token punctuation\">]<\/span><br \/>\n  sched:sched_process_exit        <span class=\"token punctuation\">[<\/span>Tracepoint event<span class=\"token punctuation\">]<\/span><\/p>\n<p>\u9002\u7528\u573a\u666f&#xff1a;<\/p>\n<ul>\n<li>\u5185\u6838\u8def\u5f84\u5206\u6790<\/li>\n<li>\u7cfb\u7edf\u8c03\u7528\u5206\u6790<\/li>\n<li>\u8c03\u5ea6\u884c\u4e3a\u5206\u6790<\/li>\n<\/ul>\n<p>\u6df7\u5408\u4f7f\u7528\u793a\u4f8b&#xff1a;<\/p>\n<p><span class=\"token comment\"># \u540c\u65f6\u8bb0\u5f55\u786c\u4ef6\u4e8b\u4ef6\u548c\u8ddf\u8e2a\u70b9\u4e8b\u4ef6<\/span><br \/>\nperf record -e cycles,cache-misses,sched:sched_switch -g .\/my_program<\/p>\n<hr \/>\n<h3>&#x1f3af; IPC \u5206\u6790&#xff1a;CPU \u6548\u7387\u7684\u7ec8\u6781\u6307\u6807<\/h3>\n<h4>\u4ec0\u4e48\u662f IPC&#xff1f;<\/h4>\n<p>IPC (Instructions Per Cycle) \u8868\u793a CPU \u6bcf\u4e2a\u65f6\u949f\u5468\u671f\u5e73\u5747\u6267\u884c\u7684\u6307\u4ee4\u6570\u3002\u5b83\u662f\u8861\u91cf CPU \u6548\u7387\u7684\u6838\u5fc3\u6307\u6807\u3002<\/p>\n<p>IPC &#061; \u6267\u884c\u7684\u6307\u4ee4\u603b\u6570 \/ \u6d88\u8017\u7684 CPU \u5468\u671f\u603b\u6570<\/p>\n<p>\u7406\u8bba\u4e0a\u9650&#xff1a;<\/p>\n<ul>\n<li>\u6807\u91cf\u5904\u7406\u5668&#xff08;\u65e9\u671f CPU&#xff09;&#xff1a;IPC \u2264 1&#xff08;\u6bcf\u5468\u671f\u6700\u591a\u6267\u884c 1 \u6761\u6307\u4ee4&#xff09;<\/li>\n<li>\u8d85\u6807\u91cf\u5904\u7406\u5668&#xff08;\u73b0\u4ee3 CPU&#xff09;&#xff1a;IPC \u2264 4-6&#xff08;\u6bcf\u5468\u671f\u53ef\u540c\u65f6\u6267\u884c\u591a\u6761\u6307\u4ee4&#xff09;<\/li>\n<\/ul>\n<p>\u5b9e\u9645\u60c5\u51b5&#xff1a;<\/p>\n<p>#mermaid-svg-w33gCnSzcar3Mnvq{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;fill:#333;}@keyframes edge-animation-frame{from{stroke-dashoffset:0;}}@keyframes dash{to{stroke-dashoffset:0;}}#mermaid-svg-w33gCnSzcar3Mnvq .edge-animation-slow{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 50s linear infinite;stroke-linecap:round;}#mermaid-svg-w33gCnSzcar3Mnvq .edge-animation-fast{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 20s linear infinite;stroke-linecap:round;}#mermaid-svg-w33gCnSzcar3Mnvq .error-icon{fill:#552222;}#mermaid-svg-w33gCnSzcar3Mnvq .error-text{fill:#552222;stroke:#552222;}#mermaid-svg-w33gCnSzcar3Mnvq .edge-thickness-normal{stroke-width:1px;}#mermaid-svg-w33gCnSzcar3Mnvq .edge-thickness-thick{stroke-width:3.5px;}#mermaid-svg-w33gCnSzcar3Mnvq .edge-pattern-solid{stroke-dasharray:0;}#mermaid-svg-w33gCnSzcar3Mnvq .edge-thickness-invisible{stroke-width:0;fill:none;}#mermaid-svg-w33gCnSzcar3Mnvq .edge-pattern-dashed{stroke-dasharray:3;}#mermaid-svg-w33gCnSzcar3Mnvq .edge-pattern-dotted{stroke-dasharray:2;}#mermaid-svg-w33gCnSzcar3Mnvq .marker{fill:#333333;stroke:#333333;}#mermaid-svg-w33gCnSzcar3Mnvq .marker.cross{stroke:#333333;}#mermaid-svg-w33gCnSzcar3Mnvq svg{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;}#mermaid-svg-w33gCnSzcar3Mnvq p{margin:0;}#mermaid-svg-w33gCnSzcar3Mnvq .label{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;color:#333;}#mermaid-svg-w33gCnSzcar3Mnvq .cluster-label text{fill:#333;}#mermaid-svg-w33gCnSzcar3Mnvq .cluster-label span{color:#333;}#mermaid-svg-w33gCnSzcar3Mnvq .cluster-label span p{background-color:transparent;}#mermaid-svg-w33gCnSzcar3Mnvq .label text,#mermaid-svg-w33gCnSzcar3Mnvq span{fill:#333;color:#333;}#mermaid-svg-w33gCnSzcar3Mnvq .node rect,#mermaid-svg-w33gCnSzcar3Mnvq .node circle,#mermaid-svg-w33gCnSzcar3Mnvq .node ellipse,#mermaid-svg-w33gCnSzcar3Mnvq .node polygon,#mermaid-svg-w33gCnSzcar3Mnvq .node path{fill:#ECECFF;stroke:#9370DB;stroke-width:1px;}#mermaid-svg-w33gCnSzcar3Mnvq .rough-node .label text,#mermaid-svg-w33gCnSzcar3Mnvq .node .label text,#mermaid-svg-w33gCnSzcar3Mnvq .image-shape .label,#mermaid-svg-w33gCnSzcar3Mnvq .icon-shape .label{text-anchor:middle;}#mermaid-svg-w33gCnSzcar3Mnvq .node .katex path{fill:#000;stroke:#000;stroke-width:1px;}#mermaid-svg-w33gCnSzcar3Mnvq .rough-node .label,#mermaid-svg-w33gCnSzcar3Mnvq .node .label,#mermaid-svg-w33gCnSzcar3Mnvq .image-shape .label,#mermaid-svg-w33gCnSzcar3Mnvq .icon-shape .label{text-align:center;}#mermaid-svg-w33gCnSzcar3Mnvq .node.clickable{cursor:pointer;}#mermaid-svg-w33gCnSzcar3Mnvq .root .anchor path{fill:#333333!important;stroke-width:0;stroke:#333333;}#mermaid-svg-w33gCnSzcar3Mnvq .arrowheadPath{fill:#333333;}#mermaid-svg-w33gCnSzcar3Mnvq .edgePath .path{stroke:#333333;stroke-width:2.0px;}#mermaid-svg-w33gCnSzcar3Mnvq .flowchart-link{stroke:#333333;fill:none;}#mermaid-svg-w33gCnSzcar3Mnvq .edgeLabel{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-w33gCnSzcar3Mnvq .edgeLabel p{background-color:rgba(232,232,232, 0.8);}#mermaid-svg-w33gCnSzcar3Mnvq .edgeLabel rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-w33gCnSzcar3Mnvq .labelBkg{background-color:rgba(232, 232, 232, 0.5);}#mermaid-svg-w33gCnSzcar3Mnvq .cluster rect{fill:#ffffde;stroke:#aaaa33;stroke-width:1px;}#mermaid-svg-w33gCnSzcar3Mnvq .cluster text{fill:#333;}#mermaid-svg-w33gCnSzcar3Mnvq .cluster span{color:#333;}#mermaid-svg-w33gCnSzcar3Mnvq div.mermaidTooltip{position:absolute;text-align:center;max-width:200px;padding:2px;font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:12px;background:hsl(80, 100%, 96.2745098039%);border:1px solid #aaaa33;border-radius:2px;pointer-events:none;z-index:100;}#mermaid-svg-w33gCnSzcar3Mnvq .flowchartTitleText{text-anchor:middle;font-size:18px;fill:#333;}#mermaid-svg-w33gCnSzcar3Mnvq rect.text{fill:none;stroke-width:0;}#mermaid-svg-w33gCnSzcar3Mnvq .icon-shape,#mermaid-svg-w33gCnSzcar3Mnvq .image-shape{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-w33gCnSzcar3Mnvq .icon-shape p,#mermaid-svg-w33gCnSzcar3Mnvq .image-shape p{background-color:rgba(232,232,232, 0.8);padding:2px;}#mermaid-svg-w33gCnSzcar3Mnvq .icon-shape rect,#mermaid-svg-w33gCnSzcar3Mnvq .image-shape rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-w33gCnSzcar3Mnvq .label-icon{display:inline-block;height:1em;overflow:visible;vertical-align:-0.125em;}#mermaid-svg-w33gCnSzcar3Mnvq .node .label-icon path{fill:currentColor;stroke:revert;stroke-width:revert;}#mermaid-svg-w33gCnSzcar3Mnvq :root{&#8211;mermaid-font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;}<span class=\"nodeLabel\"><\/p>\n<p>IPC \u503c\u7684\u5de5\u7a0b\u610f\u4e49<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>IPC &lt; 1\u26a0\ufe0f CPU \u4e25\u91cd\u963b\u585e\u7b49\u5f85\u5185\u5b58\/IO<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>IPC &#061; 1-2\u26a1 \u6709\u4f18\u5316\u7a7a\u95f4\u5b58\u5728\u7f13\u5b58\u672a\u547d\u4e2d<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>IPC &#061; 2-3\u2705 \u8f83\u597d\u6027\u80fd\u7f13\u5b58\u53cb\u597d<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>IPC &gt; 3&#x1f680; \u4f18\u79c0\u6027\u80fd\u9ad8\u6548\u6d41\u6c34\u7ebf<\/p>\n<p><\/span><\/p>\n<h4>IPC \u4f4e\u7684\u5e38\u89c1\u539f\u56e0<\/h4>\n<li>\u7f13\u5b58\u672a\u547d\u4e2d&#xff1a;CPU \u7b49\u5f85\u6570\u636e\u4ece\u4e3b\u5b58\u52a0\u8f7d<\/li>\n<li>TLB \u672a\u547d\u4e2d&#xff1a;CPU \u7b49\u5f85\u9875\u8868\u67e5\u8be2<\/li>\n<li>\u5206\u652f\u9884\u6d4b\u5931\u8d25&#xff1a;CPU \u9700\u8981\u5237\u65b0\u6d41\u6c34\u7ebf<\/li>\n<li>\u6570\u636e\u4f9d\u8d56&#xff1a;\u540e\u7eed\u6307\u4ee4\u4f9d\u8d56\u524d\u9762\u6307\u4ee4\u7684\u7ed3\u679c<\/li>\n<li>\u8d44\u6e90\u7ade\u4e89&#xff1a;\u591a\u4e2a\u6307\u4ee4\u7ade\u4e89\u540c\u4e00\u4e2a\u6267\u884c\u5355\u5143<\/li>\n<h4>\u81ea\u52a8\u9a7e\u9a76\u573a\u666f\u4e2d\u7684 IPC \u5206\u6790<\/h4>\n<p>\u573a\u666f 1&#xff1a;\u70b9\u4e91\u5904\u7406<\/p>\n<p>perf <span class=\"token function\">stat<\/span> .\/voxel_filter<\/p>\n<p><span class=\"token comment\"># \u7ed3\u679c<\/span><br \/>\n  instructions:  <span class=\"token number\">8,456<\/span>,789,012<br \/>\n  cycles:        <span class=\"token number\">6,234<\/span>,567,890<br \/>\n  IPC:           <span class=\"token number\">1.36<\/span>  \u2b05\ufe0f \u4f4e\u4e8e <span class=\"token number\">2<\/span>&#xff0c;\u8bf4\u660e\u5b58\u5728\u7b49\u5f85<br \/>\n  cache-misses:  <span class=\"token number\">18<\/span>%   \u2b05\ufe0f \u6839\u56e0&#xff1a;\u7f13\u5b58\u672a\u547d\u4e2d<\/p>\n<p>\u4f18\u5316\u65b9\u5411&#xff1a;\u6539\u8fdb\u6570\u636e\u8bbf\u95ee\u6a21\u5f0f&#xff0c;\u63d0\u9ad8\u7f13\u5b58\u547d\u4e2d\u7387<\/p>\n<p>\u573a\u666f 2&#xff1a;\u76ee\u6807\u68c0\u6d4b&#xff08;\u5377\u79ef\u8fd0\u7b97&#xff09;<\/p>\n<p>perf <span class=\"token function\">stat<\/span> .\/cnn_inference<\/p>\n<p><span class=\"token comment\"># \u7ed3\u679c<\/span><br \/>\n  instructions:  <span class=\"token number\">45,678<\/span>,901,234<br \/>\n  cycles:        <span class=\"token number\">15,234<\/span>,567,890<br \/>\n  IPC:           <span class=\"token number\">3.00<\/span>  \u2b05\ufe0f \u63a5\u8fd1 <span class=\"token number\">3<\/span>&#xff0c;\u6027\u80fd\u8f83\u597d<br \/>\n  cache-misses:  <span class=\"token number\">2<\/span>%    \u2b05\ufe0f \u7f13\u5b58\u53cb\u597d<\/p>\n<p>\u7ed3\u8bba&#xff1a;CNN \u7684\u5377\u79ef\u8fd0\u7b97\u5177\u6709\u826f\u597d\u7684\u5c40\u90e8\u6027&#xff0c;\u7f13\u5b58\u5229\u7528\u7387\u9ad8<\/p>\n<hr \/>\n<h3>\u26a1 \u5206\u652f\u9884\u6d4b\u5931\u8d25\u7684\u6027\u80fd\u5f71\u54cd<\/h3>\n<p>\u73b0\u4ee3 CPU \u4f7f\u7528\u6d41\u6c34\u7ebf\u6280\u672f\u548c\u5206\u652f\u9884\u6d4b\u6765\u63d0\u5347\u6027\u80fd\u3002\u5f53\u9047\u5230\u6761\u4ef6\u5206\u652f&#xff08;if\u3001for\u3001while&#xff09;\u65f6&#xff0c;CPU \u4f1a\u9884\u6d4b\u5206\u652f\u7684\u8d70\u5411&#xff08;\u8df3\u8f6c\u6216\u4e0d\u8df3\u8f6c&#xff09;&#xff0c;\u5e76\u63d0\u524d\u6267\u884c\u9884\u6d4b\u8def\u5f84\u4e0a\u7684\u6307\u4ee4\u3002<\/p>\n<p>#mermaid-svg-aDBZ0vdCv7QBeo0l{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;fill:#333;}@keyframes edge-animation-frame{from{stroke-dashoffset:0;}}@keyframes dash{to{stroke-dashoffset:0;}}#mermaid-svg-aDBZ0vdCv7QBeo0l .edge-animation-slow{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 50s linear infinite;stroke-linecap:round;}#mermaid-svg-aDBZ0vdCv7QBeo0l .edge-animation-fast{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 20s linear infinite;stroke-linecap:round;}#mermaid-svg-aDBZ0vdCv7QBeo0l .error-icon{fill:#552222;}#mermaid-svg-aDBZ0vdCv7QBeo0l .error-text{fill:#552222;stroke:#552222;}#mermaid-svg-aDBZ0vdCv7QBeo0l .edge-thickness-normal{stroke-width:1px;}#mermaid-svg-aDBZ0vdCv7QBeo0l .edge-thickness-thick{stroke-width:3.5px;}#mermaid-svg-aDBZ0vdCv7QBeo0l .edge-pattern-solid{stroke-dasharray:0;}#mermaid-svg-aDBZ0vdCv7QBeo0l .edge-thickness-invisible{stroke-width:0;fill:none;}#mermaid-svg-aDBZ0vdCv7QBeo0l .edge-pattern-dashed{stroke-dasharray:3;}#mermaid-svg-aDBZ0vdCv7QBeo0l .edge-pattern-dotted{stroke-dasharray:2;}#mermaid-svg-aDBZ0vdCv7QBeo0l .marker{fill:#333333;stroke:#333333;}#mermaid-svg-aDBZ0vdCv7QBeo0l .marker.cross{stroke:#333333;}#mermaid-svg-aDBZ0vdCv7QBeo0l svg{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;}#mermaid-svg-aDBZ0vdCv7QBeo0l p{margin:0;}#mermaid-svg-aDBZ0vdCv7QBeo0l .label{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;color:#333;}#mermaid-svg-aDBZ0vdCv7QBeo0l .cluster-label text{fill:#333;}#mermaid-svg-aDBZ0vdCv7QBeo0l .cluster-label span{color:#333;}#mermaid-svg-aDBZ0vdCv7QBeo0l .cluster-label span p{background-color:transparent;}#mermaid-svg-aDBZ0vdCv7QBeo0l .label text,#mermaid-svg-aDBZ0vdCv7QBeo0l span{fill:#333;color:#333;}#mermaid-svg-aDBZ0vdCv7QBeo0l .node rect,#mermaid-svg-aDBZ0vdCv7QBeo0l .node circle,#mermaid-svg-aDBZ0vdCv7QBeo0l .node ellipse,#mermaid-svg-aDBZ0vdCv7QBeo0l .node polygon,#mermaid-svg-aDBZ0vdCv7QBeo0l .node path{fill:#ECECFF;stroke:#9370DB;stroke-width:1px;}#mermaid-svg-aDBZ0vdCv7QBeo0l .rough-node .label text,#mermaid-svg-aDBZ0vdCv7QBeo0l .node .label text,#mermaid-svg-aDBZ0vdCv7QBeo0l .image-shape .label,#mermaid-svg-aDBZ0vdCv7QBeo0l .icon-shape .label{text-anchor:middle;}#mermaid-svg-aDBZ0vdCv7QBeo0l .node .katex path{fill:#000;stroke:#000;stroke-width:1px;}#mermaid-svg-aDBZ0vdCv7QBeo0l .rough-node .label,#mermaid-svg-aDBZ0vdCv7QBeo0l .node .label,#mermaid-svg-aDBZ0vdCv7QBeo0l .image-shape .label,#mermaid-svg-aDBZ0vdCv7QBeo0l .icon-shape .label{text-align:center;}#mermaid-svg-aDBZ0vdCv7QBeo0l .node.clickable{cursor:pointer;}#mermaid-svg-aDBZ0vdCv7QBeo0l .root .anchor path{fill:#333333!important;stroke-width:0;stroke:#333333;}#mermaid-svg-aDBZ0vdCv7QBeo0l .arrowheadPath{fill:#333333;}#mermaid-svg-aDBZ0vdCv7QBeo0l .edgePath .path{stroke:#333333;stroke-width:2.0px;}#mermaid-svg-aDBZ0vdCv7QBeo0l .flowchart-link{stroke:#333333;fill:none;}#mermaid-svg-aDBZ0vdCv7QBeo0l .edgeLabel{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-aDBZ0vdCv7QBeo0l .edgeLabel p{background-color:rgba(232,232,232, 0.8);}#mermaid-svg-aDBZ0vdCv7QBeo0l .edgeLabel rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-aDBZ0vdCv7QBeo0l .labelBkg{background-color:rgba(232, 232, 232, 0.5);}#mermaid-svg-aDBZ0vdCv7QBeo0l .cluster rect{fill:#ffffde;stroke:#aaaa33;stroke-width:1px;}#mermaid-svg-aDBZ0vdCv7QBeo0l .cluster text{fill:#333;}#mermaid-svg-aDBZ0vdCv7QBeo0l .cluster span{color:#333;}#mermaid-svg-aDBZ0vdCv7QBeo0l div.mermaidTooltip{position:absolute;text-align:center;max-width:200px;padding:2px;font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:12px;background:hsl(80, 100%, 96.2745098039%);border:1px solid #aaaa33;border-radius:2px;pointer-events:none;z-index:100;}#mermaid-svg-aDBZ0vdCv7QBeo0l .flowchartTitleText{text-anchor:middle;font-size:18px;fill:#333;}#mermaid-svg-aDBZ0vdCv7QBeo0l rect.text{fill:none;stroke-width:0;}#mermaid-svg-aDBZ0vdCv7QBeo0l .icon-shape,#mermaid-svg-aDBZ0vdCv7QBeo0l .image-shape{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-aDBZ0vdCv7QBeo0l .icon-shape p,#mermaid-svg-aDBZ0vdCv7QBeo0l .image-shape p{background-color:rgba(232,232,232, 0.8);padding:2px;}#mermaid-svg-aDBZ0vdCv7QBeo0l .icon-shape rect,#mermaid-svg-aDBZ0vdCv7QBeo0l .image-shape rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-aDBZ0vdCv7QBeo0l .label-icon{display:inline-block;height:1em;overflow:visible;vertical-align:-0.125em;}#mermaid-svg-aDBZ0vdCv7QBeo0l .node .label-icon path{fill:currentColor;stroke:revert;stroke-width:revert;}#mermaid-svg-aDBZ0vdCv7QBeo0l :root{&#8211;mermaid-font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;}<span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u53d6\u6307 Fetch<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u8bd1\u7801 Decode<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u6267\u884c Execute<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u8bbf\u5b58 Memory<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u5199\u56de Write-back<\/p>\n<p><\/span><\/p>\n<p>\u5206\u652f\u9884\u6d4b\u6210\u529f&#xff1a;\u6d41\u6c34\u7ebf\u987a\u7545&#xff0c;\u6307\u4ee4\u8fde\u7eed\u6267\u884c<br \/>\n\u5206\u652f\u9884\u6d4b\u5931\u8d25&#xff1a;\u6d41\u6c34\u7ebf\u88ab\u6e05\u7a7a&#xff0c;\u5df2\u6267\u884c\u7684\u6307\u4ee4\u4f5c\u5e9f&#xff0c;\u91cd\u65b0\u4ece\u6b63\u786e\u8def\u5f84\u5f00\u59cb&#xff0c;\u6d6a\u8d39 10-20 \u4e2a CPU \u5468\u671f<\/p>\n<h4>\u6d4b\u91cf\u5206\u652f\u9884\u6d4b\u6027\u80fd<\/h4>\n<p>perf <span class=\"token function\">stat<\/span> -e branches,branch-misses .\/my_program<\/p>\n<p><span class=\"token comment\"># \u8f93\u51fa<\/span><br \/>\n  <span class=\"token number\">890,123<\/span>,456      branches<br \/>\n   <span class=\"token number\">44,506<\/span>,172      branch-misses             <span class=\"token comment\">#    5.00 % of all branches<\/span><\/p>\n<p>\u5206\u652f\u9884\u6d4b\u5931\u8d25\u7387&#xff1a;<\/p>\n<ul>\n<li>&lt; 2%&#xff1a;\u4f18\u79c0<\/li>\n<li>2-5%&#xff1a;\u6b63\u5e38<\/li>\n<li>&gt; 5%&#xff1a;\u9700\u8981\u4f18\u5316<\/li>\n<\/ul>\n<h4>\u81ea\u52a8\u9a7e\u9a76\u573a\u666f&#xff1a;\u4f18\u5316\u5206\u652f\u5bc6\u96c6\u7684\u4ee3\u7801<\/h4>\n<p>\u6848\u4f8b&#xff1a;\u969c\u788d\u7269\u8fc7\u6ee4\u51fd\u6570<\/p>\n<p><span class=\"token comment\">\/\/ \u539f\u59cb\u4ee3\u7801&#xff08;\u5206\u652f\u591a&#xff09;<\/span><br \/>\n<span class=\"token keyword\">int<\/span> <span class=\"token function\">FilterObstacles<\/span><span class=\"token punctuation\">(<\/span><span class=\"token keyword\">const<\/span> std<span class=\"token double-colon punctuation\">::<\/span>vector<span class=\"token operator\">&lt;<\/span>Point<span class=\"token operator\">&gt;<\/span><span class=\"token operator\">&amp;<\/span> points<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n    <span class=\"token keyword\">int<\/span> count <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">0<\/span><span class=\"token punctuation\">;<\/span><br \/>\n    <span class=\"token keyword\">for<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token keyword\">const<\/span> <span class=\"token keyword\">auto<\/span><span class=\"token operator\">&amp;<\/span> p <span class=\"token operator\">:<\/span> points<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n        <span class=\"token keyword\">if<\/span> <span class=\"token punctuation\">(<\/span>p<span class=\"token punctuation\">.<\/span>x <span class=\"token operator\">&gt;<\/span> min_x <span class=\"token operator\">&amp;&amp;<\/span> p<span class=\"token punctuation\">.<\/span>x <span class=\"token operator\">&lt;<\/span> max_x <span class=\"token operator\">&amp;&amp;<\/span>  <span class=\"token comment\">\/\/ \u2b05\ufe0f \u5206\u652f 1<\/span><br \/>\n            p<span class=\"token punctuation\">.<\/span>y <span class=\"token operator\">&gt;<\/span> min_y <span class=\"token operator\">&amp;&amp;<\/span> p<span class=\"token punctuation\">.<\/span>y <span class=\"token operator\">&lt;<\/span> max_y <span class=\"token operator\">&amp;&amp;<\/span>  <span class=\"token comment\">\/\/ \u2b05\ufe0f \u5206\u652f 2<\/span><br \/>\n            p<span class=\"token punctuation\">.<\/span>z <span class=\"token operator\">&gt;<\/span> min_z <span class=\"token operator\">&amp;&amp;<\/span> p<span class=\"token punctuation\">.<\/span>z <span class=\"token operator\">&lt;<\/span> max_z<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span>  <span class=\"token comment\">\/\/ \u2b05\ufe0f \u5206\u652f 3<\/span><br \/>\n            count<span class=\"token operator\">&#043;&#043;<\/span><span class=\"token punctuation\">;<\/span><br \/>\n        <span class=\"token punctuation\">}<\/span><br \/>\n    <span class=\"token punctuation\">}<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> count<span class=\"token punctuation\">;<\/span><br \/>\n<span class=\"token punctuation\">}<\/span><\/p>\n<p>\u6d4b\u91cf&#xff1a;<\/p>\n<p>perf <span class=\"token function\">stat<\/span> -e branches,branch-misses .\/obstacle_filter<\/p>\n<p><span class=\"token comment\"># \u7ed3\u679c<\/span><br \/>\n  branch-misses: <span class=\"token number\">12.3<\/span>%  \u2b05\ufe0f \u8fdc\u9ad8\u4e8e <span class=\"token number\">5<\/span>%<\/p>\n<p>\u4f18\u5316&#xff1a;\u51cf\u5c11\u5206\u652f<\/p>\n<p><span class=\"token comment\">\/\/ \u4f18\u5316\u540e&#xff08;\u4f7f\u7528\u4f4d\u8fd0\u7b97\u51cf\u5c11\u5206\u652f&#xff09;<\/span><br \/>\n<span class=\"token keyword\">int<\/span> <span class=\"token function\">FilterObstacles_Optimized<\/span><span class=\"token punctuation\">(<\/span><span class=\"token keyword\">const<\/span> std<span class=\"token double-colon punctuation\">::<\/span>vector<span class=\"token operator\">&lt;<\/span>Point<span class=\"token operator\">&gt;<\/span><span class=\"token operator\">&amp;<\/span> points<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n    <span class=\"token keyword\">int<\/span> count <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">0<\/span><span class=\"token punctuation\">;<\/span><br \/>\n    <span class=\"token keyword\">for<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token keyword\">const<\/span> <span class=\"token keyword\">auto<\/span><span class=\"token operator\">&amp;<\/span> p <span class=\"token operator\">:<\/span> points<span class=\"token punctuation\">)<\/span> <span class=\"token punctuation\">{<\/span><br \/>\n        <span class=\"token comment\">\/\/ \u4f7f\u7528\u4f4d\u8fd0\u7b97\u5408\u5e76\u6761\u4ef6&#xff0c;\u51cf\u5c11\u5206\u652f<\/span><br \/>\n        <span class=\"token keyword\">bool<\/span> in_range <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">(<\/span>p<span class=\"token punctuation\">.<\/span>x <span class=\"token operator\">&gt;<\/span> min_x<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&amp;<\/span> <span class=\"token punctuation\">(<\/span>p<span class=\"token punctuation\">.<\/span>x <span class=\"token operator\">&lt;<\/span> max_x<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&amp;<\/span><br \/>\n                        <span class=\"token punctuation\">(<\/span>p<span class=\"token punctuation\">.<\/span>y <span class=\"token operator\">&gt;<\/span> min_y<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&amp;<\/span> <span class=\"token punctuation\">(<\/span>p<span class=\"token punctuation\">.<\/span>y <span class=\"token operator\">&lt;<\/span> max_y<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&amp;<\/span><br \/>\n                        <span class=\"token punctuation\">(<\/span>p<span class=\"token punctuation\">.<\/span>z <span class=\"token operator\">&gt;<\/span> min_z<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&amp;<\/span> <span class=\"token punctuation\">(<\/span>p<span class=\"token punctuation\">.<\/span>z <span class=\"token operator\">&lt;<\/span> max_z<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">;<\/span><br \/>\n        count <span class=\"token operator\">&#043;&#061;<\/span> in_range<span class=\"token punctuation\">;<\/span>  <span class=\"token comment\">\/\/ \u2b05\ufe0f \u65e0\u5206\u652f<\/span><br \/>\n    <span class=\"token punctuation\">}<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> count<span class=\"token punctuation\">;<\/span><br \/>\n<span class=\"token punctuation\">}<\/span><\/p>\n<p>\u4f18\u5316\u540e&#xff1a;<\/p>\n<p>perf <span class=\"token function\">stat<\/span> -e branches,branch-misses .\/obstacle_filter_optimized<\/p>\n<p><span class=\"token comment\"># \u7ed3\u679c<\/span><br \/>\n  branch-misses: <span class=\"token number\">2.1<\/span>%  \u2b05\ufe0f \u4ece <span class=\"token number\">12.3<\/span>% \u964d\u5230 <span class=\"token number\">2.1<\/span>%<\/p>\n<hr \/>\n<h3>&#x1f4dd; \u603b\u7ed3\u4e0e\u6700\u4f73\u5b9e\u8df5<\/h3>\n<h4>\u6838\u5fc3\u8981\u70b9<\/h4>\n<ul>\n<li>\u6027\u80fd\u8ba1\u6570\u5668\u662f\u6027\u80fd\u5206\u6790\u7684\u57fa\u77f3&#xff1a;\u63d0\u4f9b\u786c\u4ef6\u7ea7\u7684\u3001\u51e0\u4e4e\u96f6\u5f00\u9500\u7684\u7cbe\u786e\u6d4b\u91cf<\/li>\n<li>perf \u5de5\u5177\u4e09\u5251\u5ba2&#xff1a;\n<ul>\n<li>perf top&#xff1a;\u5b9e\u65f6\u70ed\u70b9\u96f7\u8fbe&#xff0c;\u5feb\u901f\u5b9a\u4f4d\u70ed\u70b9\u51fd\u6570<\/li>\n<li>perf stat&#xff1a;\u6027\u80fd\u4e8b\u4ef6\u7edf\u8ba1\u4eea\u8868\u76d8&#xff0c;\u91cf\u5316\u6027\u80fd\u6307\u6807<\/li>\n<li>perf record\/report&#xff1a;\u6027\u80fd\u4e8b\u4ef6\u91c7\u6837\u5f55\u50cf\u673a&#xff0c;\u6df1\u5ea6\u5206\u6790\u8c03\u7528\u6808<\/li>\n<\/ul>\n<\/li>\n<li>IPC \u662f CPU \u6548\u7387\u7684\u7ec8\u6781\u6307\u6807&#xff1a;IPC &lt; 2 \u8bf4\u660e CPU \u5728\u7b49\u5f85&#xff0c;\u9700\u8981\u4f18\u5316\u6570\u636e\u8bbf\u95ee\u6216\u51cf\u5c11\u5206\u652f<\/li>\n<li>\u7f13\u5b58\u672a\u547d\u4e2d\u548c\u5206\u652f\u9884\u6d4b\u5931\u8d25\u662f\u4e24\u5927\u6027\u80fd\u6740\u624b&#xff1a;\u901a\u8fc7\u6539\u8fdb\u6570\u636e\u7ed3\u6784\u548c\u7b97\u6cd5\u53ef\u4ee5\u5927\u5e45\u4f18\u5316<\/li>\n<\/ul>\n<h4>perf \u4f7f\u7528\u6e05\u5355<\/h4>\n<p>\u2705 \u65e5\u5e38\u6027\u80fd\u5de1\u68c0<\/p>\n<p><span class=\"token comment\"># \u5feb\u901f\u67e5\u770b\u7cfb\u7edf\u70ed\u70b9<\/span><br \/>\nperf <span class=\"token function\">top<\/span><\/p>\n<p><span class=\"token comment\"># \u7edf\u8ba1\u7a0b\u5e8f\u6027\u80fd\u6307\u6807<\/span><br \/>\nperf <span class=\"token function\">stat<\/span> -d .\/my_program<\/p>\n<p>\u2705 \u6df1\u5ea6\u6027\u80fd\u5206\u6790<\/p>\n<p><span class=\"token comment\"># \u8bb0\u5f55 CPU \u70ed\u70b9<\/span><br \/>\nperf record -g -F <span class=\"token number\">99<\/span> .\/my_program<br \/>\nperf report<\/p>\n<p><span class=\"token comment\"># \u5206\u6790\u7f13\u5b58\u6027\u80fd<\/span><br \/>\nperf <span class=\"token function\">stat<\/span> -e cache-references,cache-misses,LLC-loads,LLC-load-misses .\/my_program<\/p>\n<p><span class=\"token comment\"># \u5206\u6790\u5206\u652f\u9884\u6d4b<\/span><br \/>\nperf <span class=\"token function\">stat<\/span> -e branches,branch-misses .\/my_program<\/p>\n<p>\u2705 \u81ea\u52a8\u9a7e\u9a76\u7279\u5b9a\u573a\u666f<\/p>\n<p><span class=\"token comment\"># \u70b9\u4e91\u5904\u7406&#xff08;\u5173\u6ce8\u7f13\u5b58&#xff09;<\/span><br \/>\nperf record -e LLC-load-misses -g .\/point_cloud_processor<\/p>\n<p><span class=\"token comment\"># \u76ee\u6807\u68c0\u6d4b&#xff08;\u5173\u6ce8 IPC&#xff09;<\/span><br \/>\nperf <span class=\"token function\">stat<\/span> -e cycles,instructions,cache-misses,branch-misses .\/object_detector<\/p>\n<p><span class=\"token comment\"># \u8def\u5f84\u89c4\u5212&#xff08;\u5173\u6ce8\u8c03\u5ea6&#xff09;<\/span><br \/>\nperf record -e sched:sched_switch -g .\/path_planner<\/p>\n<h4>\u6027\u80fd\u4f18\u5316\u7684\u7cfb\u7edf\u65b9\u6cd5<\/h4>\n<p>#mermaid-svg-OiqUBYv7HQ2AEC0y{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;fill:#333;}@keyframes edge-animation-frame{from{stroke-dashoffset:0;}}@keyframes dash{to{stroke-dashoffset:0;}}#mermaid-svg-OiqUBYv7HQ2AEC0y .edge-animation-slow{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 50s linear infinite;stroke-linecap:round;}#mermaid-svg-OiqUBYv7HQ2AEC0y .edge-animation-fast{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 20s linear infinite;stroke-linecap:round;}#mermaid-svg-OiqUBYv7HQ2AEC0y .error-icon{fill:#552222;}#mermaid-svg-OiqUBYv7HQ2AEC0y .error-text{fill:#552222;stroke:#552222;}#mermaid-svg-OiqUBYv7HQ2AEC0y .edge-thickness-normal{stroke-width:1px;}#mermaid-svg-OiqUBYv7HQ2AEC0y .edge-thickness-thick{stroke-width:3.5px;}#mermaid-svg-OiqUBYv7HQ2AEC0y .edge-pattern-solid{stroke-dasharray:0;}#mermaid-svg-OiqUBYv7HQ2AEC0y .edge-thickness-invisible{stroke-width:0;fill:none;}#mermaid-svg-OiqUBYv7HQ2AEC0y .edge-pattern-dashed{stroke-dasharray:3;}#mermaid-svg-OiqUBYv7HQ2AEC0y .edge-pattern-dotted{stroke-dasharray:2;}#mermaid-svg-OiqUBYv7HQ2AEC0y .marker{fill:#333333;stroke:#333333;}#mermaid-svg-OiqUBYv7HQ2AEC0y .marker.cross{stroke:#333333;}#mermaid-svg-OiqUBYv7HQ2AEC0y svg{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;}#mermaid-svg-OiqUBYv7HQ2AEC0y p{margin:0;}#mermaid-svg-OiqUBYv7HQ2AEC0y .label{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;color:#333;}#mermaid-svg-OiqUBYv7HQ2AEC0y .cluster-label text{fill:#333;}#mermaid-svg-OiqUBYv7HQ2AEC0y .cluster-label span{color:#333;}#mermaid-svg-OiqUBYv7HQ2AEC0y .cluster-label span p{background-color:transparent;}#mermaid-svg-OiqUBYv7HQ2AEC0y .label text,#mermaid-svg-OiqUBYv7HQ2AEC0y span{fill:#333;color:#333;}#mermaid-svg-OiqUBYv7HQ2AEC0y .node rect,#mermaid-svg-OiqUBYv7HQ2AEC0y .node circle,#mermaid-svg-OiqUBYv7HQ2AEC0y .node ellipse,#mermaid-svg-OiqUBYv7HQ2AEC0y .node polygon,#mermaid-svg-OiqUBYv7HQ2AEC0y .node path{fill:#ECECFF;stroke:#9370DB;stroke-width:1px;}#mermaid-svg-OiqUBYv7HQ2AEC0y .rough-node .label text,#mermaid-svg-OiqUBYv7HQ2AEC0y .node .label text,#mermaid-svg-OiqUBYv7HQ2AEC0y .image-shape .label,#mermaid-svg-OiqUBYv7HQ2AEC0y .icon-shape .label{text-anchor:middle;}#mermaid-svg-OiqUBYv7HQ2AEC0y .node .katex path{fill:#000;stroke:#000;stroke-width:1px;}#mermaid-svg-OiqUBYv7HQ2AEC0y .rough-node .label,#mermaid-svg-OiqUBYv7HQ2AEC0y .node .label,#mermaid-svg-OiqUBYv7HQ2AEC0y .image-shape .label,#mermaid-svg-OiqUBYv7HQ2AEC0y .icon-shape .label{text-align:center;}#mermaid-svg-OiqUBYv7HQ2AEC0y .node.clickable{cursor:pointer;}#mermaid-svg-OiqUBYv7HQ2AEC0y .root .anchor path{fill:#333333!important;stroke-width:0;stroke:#333333;}#mermaid-svg-OiqUBYv7HQ2AEC0y .arrowheadPath{fill:#333333;}#mermaid-svg-OiqUBYv7HQ2AEC0y .edgePath .path{stroke:#333333;stroke-width:2.0px;}#mermaid-svg-OiqUBYv7HQ2AEC0y .flowchart-link{stroke:#333333;fill:none;}#mermaid-svg-OiqUBYv7HQ2AEC0y .edgeLabel{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-OiqUBYv7HQ2AEC0y .edgeLabel p{background-color:rgba(232,232,232, 0.8);}#mermaid-svg-OiqUBYv7HQ2AEC0y .edgeLabel rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-OiqUBYv7HQ2AEC0y .labelBkg{background-color:rgba(232, 232, 232, 0.5);}#mermaid-svg-OiqUBYv7HQ2AEC0y .cluster rect{fill:#ffffde;stroke:#aaaa33;stroke-width:1px;}#mermaid-svg-OiqUBYv7HQ2AEC0y .cluster text{fill:#333;}#mermaid-svg-OiqUBYv7HQ2AEC0y .cluster span{color:#333;}#mermaid-svg-OiqUBYv7HQ2AEC0y div.mermaidTooltip{position:absolute;text-align:center;max-width:200px;padding:2px;font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:12px;background:hsl(80, 100%, 96.2745098039%);border:1px solid #aaaa33;border-radius:2px;pointer-events:none;z-index:100;}#mermaid-svg-OiqUBYv7HQ2AEC0y .flowchartTitleText{text-anchor:middle;font-size:18px;fill:#333;}#mermaid-svg-OiqUBYv7HQ2AEC0y rect.text{fill:none;stroke-width:0;}#mermaid-svg-OiqUBYv7HQ2AEC0y .icon-shape,#mermaid-svg-OiqUBYv7HQ2AEC0y .image-shape{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-OiqUBYv7HQ2AEC0y .icon-shape p,#mermaid-svg-OiqUBYv7HQ2AEC0y .image-shape p{background-color:rgba(232,232,232, 0.8);padding:2px;}#mermaid-svg-OiqUBYv7HQ2AEC0y .icon-shape rect,#mermaid-svg-OiqUBYv7HQ2AEC0y .image-shape rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-OiqUBYv7HQ2AEC0y .label-icon{display:inline-block;height:1em;overflow:visible;vertical-align:-0.125em;}#mermaid-svg-OiqUBYv7HQ2AEC0y .node .label-icon path{fill:currentColor;stroke:revert;stroke-width:revert;}#mermaid-svg-OiqUBYv7HQ2AEC0y :root{&#8211;mermaid-font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;}<span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/p>\n<p>IPC &lt; 2<\/p>\n<p><\/span><span class=\"edgeLabel\"><\/p>\n<p>\u5206\u652f\u9884\u6d4b\u5931\u8d25 &gt; 5%<\/p>\n<p><\/span><span class=\"edgeLabel\"><\/p>\n<p>\u7f13\u5b58\u672a\u547d\u4e2d &gt; 10%<\/p>\n<p><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/span><span class=\"edgeLabel\"><\/p>\n<p>\u5426<\/p>\n<p><\/span><span class=\"edgeLabel\"><\/p>\n<p>\u662f<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u53d1\u73b0\u6027\u80fd\u95ee\u9898<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u4f7f\u7528 perf top<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u5b9a\u4f4d\u70ed\u70b9\u51fd\u6570<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u4f7f\u7528 perf stat<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u91cf\u5316\u6027\u80fd\u6307\u6807<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u5224\u65ad\u74f6\u9888\u7c7b\u578b<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u5185\u5b58\/\u7f13\u5b58\u95ee\u9898<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u5206\u652f\u5bc6\u96c6\u4ee3\u7801<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u6570\u636e\u8bbf\u95ee\u6a21\u5f0f\u5dee<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u4f7f\u7528 perf record\u5b9a\u4f4d\u5177\u4f53\u4ee3\u7801<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u4ee3\u7801\u4f18\u5316<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u91cd\u65b0\u6d4b\u91cf<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u6027\u80fd\u8fbe\u6807?<\/p>\n<p><\/span><span class=\"nodeLabel\"><\/p>\n<p>\u4f18\u5316\u5b8c\u6210<\/p>\n<p><\/span><\/p>\n<hr \/>\n<h3>&#x1f3af; \u4e0b\u4e00\u7ae0\u9884\u544a<\/h3>\n<p>\u5728\u672c\u7ae0\u4e2d&#xff0c;\u6211\u4eec\u638c\u63e1\u4e86 perf \u5de5\u5177\u7684\u6838\u5fc3\u7528\u6cd5&#xff0c;\u5b66\u4f1a\u4e86\u5982\u4f55\u4f7f\u7528\u6027\u80fd\u8ba1\u6570\u5668\u7cbe\u786e\u6d4b\u91cf\u7a0b\u5e8f\u7684 CPU \u6548\u7387\u3001\u7f13\u5b58\u6027\u80fd\u548c\u5206\u652f\u9884\u6d4b\u6027\u80fd\u3002\u6211\u4eec\u901a\u8fc7\u4e00\u4e2a\u771f\u5b9e\u7684\u76ee\u6807\u68c0\u6d4b NMS \u7b97\u6cd5\u4f18\u5316\u6848\u4f8b&#xff0c;\u5c55\u793a\u4e86\u5982\u4f55\u5c06 IPC \u4ece 1.21 \u63d0\u5347\u5230 2.44&#xff0c;\u5c06\u5904\u7406\u5ef6\u8fdf\u4ece 35ms \u964d\u4f4e\u5230 8ms\u3002<\/p>\n<p>\u4f46 CPU \u53ea\u662f\u6027\u80fd\u5206\u6790\u7684\u4e00\u4e2a\u7ef4\u5ea6\u3002\u5728\u4e0b\u4e00\u7ae0\u300a\u5185\u6838\u5185\u5b58\u5206\u914d\u539f\u7406\u4e0e Slab \u5206\u6790\u300b\u4e2d&#xff0c;\u6211\u4eec\u5c06\u6df1\u5165\u5185\u6838\u5185\u5b58\u7ba1\u7406\u673a\u5236&#xff1a;<\/p>\n<ul>\n<li>Slab \u5206\u914d\u5668\u5982\u4f55\u9ad8\u6548\u7ba1\u7406\u5185\u6838\u5bf9\u8c61\u7f13\u5b58<\/li>\n<li>kmalloc\u3001vmalloc\u3001kzalloc \u7684\u533a\u522b\u4e0e\u9009\u62e9<\/li>\n<li>\u5982\u4f55\u4f7f\u7528 \/proc\/slabinfo \u548c slabtop \u8bca\u65ad\u5185\u6838\u5185\u5b58\u6cc4\u6f0f<\/li>\n<li>dentry cache\u3001inode cache \u5bf9\u6587\u4ef6\u7cfb\u7edf\u6027\u80fd\u7684\u5f71\u54cd<\/li>\n<\/ul>\n<p>\u656c\u8bf7\u671f\u5f85&#xff01;&#x1f680;<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/02\/20260213152542-698f4276ba559.png\" alt=\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\" \/><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8 &#x1f50d;<br \/>\n&#x1f517; \u4ece\\&#8221;\u731c\u6d4b\\&#8221;\u5230\\&#8221;\u6d4b\u91cf\\&#8221;&#xff1a;\u6027\u80fd\u4f18\u5316\u7684\u8303\u5f0f\u8f6c\u53d8<br \/>\n\u5728\u524d\u9762\u7684\u7ae0\u8282\u4e2d&#xff0c;\u6211\u4eec\u6df1\u5165\u63a2\u8ba8\u4e86 CPU \u7f13\u5b58\u3001\u5185\u5b58\u5ef6\u8fdf\u3001\u8f6f\u4e2d\u65ad\u7b49\u6027\u80fd\u95ee\u9898\u3002\u5728\u7b2c\u5341\u7ae0\u4e2d&#xff0c;\u6211\u4eec\u53d1\u73b0\u70b9\u4e91\u5904\u7406\u7b97\u6cd5\u5b58\u5728\u9ad8\u8fbe 15% \u7684\u7f13\u5b58\u672a\u547d\u4e2d\u7387&#xff0c;\u5e76\u901a\u8fc7\u4f18\u5316\u6570\u636e\u7ed3\u6784\u5c06\u5904\u7406\u5ef6\u8fdf\u4ece 45ms \u964d\u4f4e\u5230 9.8ms\u3002<br \/>\n\u4f46\u4f60\u53ef\u80fd\u4f1a\u6709\u7591\u95ee&#xff1a;\u6211\u4eec\u662f\u5982\u4f55\u7cbe\u786e\u77e5\u9053\u7f13\u5b58\u672a\u547d\u4e2d\u7387\u7684<\/p>\n","protected":false},"author":2,"featured_media":76065,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[58,56,43,2051,44],"topic":[],"class_list":["post-76066","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-server","tag-linux","tag-56","tag-43","tag-2051","tag-44"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v20.3 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/76066.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8 &#x1f50d; &#x1f517; \u4ece&quot;\u731c\u6d4b&quot;\u5230&quot;\u6d4b\u91cf&quot;&#xff1a;\u6027\u80fd\u4f18\u5316\u7684\u8303\u5f0f\u8f6c\u53d8 \u5728\u524d\u9762\u7684\u7ae0\u8282\u4e2d&#xff0c;\u6211\u4eec\u6df1\u5165\u63a2\u8ba8\u4e86 CPU \u7f13\u5b58\u3001\u5185\u5b58\u5ef6\u8fdf\u3001\u8f6f\u4e2d\u65ad\u7b49\u6027\u80fd\u95ee\u9898\u3002\u5728\u7b2c\u5341\u7ae0\u4e2d&#xff0c;\u6211\u4eec\u53d1\u73b0\u70b9\u4e91\u5904\u7406\u7b97\u6cd5\u5b58\u5728\u9ad8\u8fbe 15% \u7684\u7f13\u5b58\u672a\u547d\u4e2d\u7387&#xff0c;\u5e76\u901a\u8fc7\u4f18\u5316\u6570\u636e\u7ed3\u6784\u5c06\u5904\u7406\u5ef6\u8fdf\u4ece 45ms \u964d\u4f4e\u5230 9.8ms\u3002 \u4f46\u4f60\u53ef\u80fd\u4f1a\u6709\u7591\u95ee&#xff1a;\u6211\u4eec\u662f\u5982\u4f55\u7cbe\u786e\u77e5\u9053\u7f13\u5b58\u672a\u547d\u4e2d\u7387\u7684\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/76066.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2026-02-13T15:25:44+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/02\/20260213152542-698f4276ba559.png\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"19 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/76066.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/76066.html\",\"name\":\"Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2026-02-13T15:25:44+00:00\",\"dateModified\":\"2026-02-13T15:25:44+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/76066.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/76066.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/76066.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/76066.html","og_locale":"zh_CN","og_type":"article","og_title":"Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8 &#x1f50d; &#x1f517; \u4ece\"\u731c\u6d4b\"\u5230\"\u6d4b\u91cf\"&#xff1a;\u6027\u80fd\u4f18\u5316\u7684\u8303\u5f0f\u8f6c\u53d8 \u5728\u524d\u9762\u7684\u7ae0\u8282\u4e2d&#xff0c;\u6211\u4eec\u6df1\u5165\u63a2\u8ba8\u4e86 CPU \u7f13\u5b58\u3001\u5185\u5b58\u5ef6\u8fdf\u3001\u8f6f\u4e2d\u65ad\u7b49\u6027\u80fd\u95ee\u9898\u3002\u5728\u7b2c\u5341\u7ae0\u4e2d&#xff0c;\u6211\u4eec\u53d1\u73b0\u70b9\u4e91\u5904\u7406\u7b97\u6cd5\u5b58\u5728\u9ad8\u8fbe 15% \u7684\u7f13\u5b58\u672a\u547d\u4e2d\u7387&#xff0c;\u5e76\u901a\u8fc7\u4f18\u5316\u6570\u636e\u7ed3\u6784\u5c06\u5904\u7406\u5ef6\u8fdf\u4ece 45ms \u964d\u4f4e\u5230 9.8ms\u3002 \u4f46\u4f60\u53ef\u80fd\u4f1a\u6709\u7591\u95ee&#xff1a;\u6211\u4eec\u662f\u5982\u4f55\u7cbe\u786e\u77e5\u9053\u7f13\u5b58\u672a\u547d\u4e2d\u7387\u7684","og_url":"https:\/\/www.wsisp.com\/helps\/76066.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2026-02-13T15:25:44+00:00","og_image":[{"url":"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/02\/20260213152542-698f4276ba559.png"}],"author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"19 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/76066.html","url":"https:\/\/www.wsisp.com\/helps\/76066.html","name":"Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2026-02-13T15:25:44+00:00","dateModified":"2026-02-13T15:25:44+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/76066.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/76066.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/76066.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"Linux \u6027\u80fd\u5b9e\u6218 | \u7b2c 11 \u7bc7 \u6027\u80fd\u8ba1\u6570\u5668\u4e0e perf \u5de5\u5177\u5165\u95e8"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/76066","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=76066"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/76066\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media\/76065"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=76066"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=76066"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=76066"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=76066"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}