{"id":1952,"date":"2023-11-29T01:07:57","date_gmt":"2023-11-28T17:07:57","guid":{"rendered":"http:\/\/zhang.mba\/?p=1952"},"modified":"2023-11-29T01:20:45","modified_gmt":"2023-11-28T17:20:45","slug":"cudf-tong-guo-gpu-jia-su-shu-ju-ke-xue-he-fen-xi-y","status":"publish","type":"post","link":"https:\/\/zhang.mba\/index.php\/2023\/11\/29\/01\/07\/57\/1952\/cudf-tong-guo-gpu-jia-su-shu-ju-ke-xue-he-fen-xi-y\/python\/zhangzhiqi\/","title":{"rendered":"cudf\uff0c\u901a\u8fc7 GPU \u52a0\u901f\u6570\u636e\u79d1\u5b66\u548c\u5206\u6790\u5e94\u7528\u7a0b\u5e8f"},"content":{"rendered":"<p><strong><a href=\"https:\/\/github.com\/rapidsai\/cudf\">https:\/\/github.com\/rapidsai\/cudf<\/a><\/strong><\/p>\n<p>cuDF \u662f\u4e00\u4e2a\u7531 NVIDIA \u5f00\u53d1\u7684 Python \u5e93\uff0c\u5b83\u662f RAPIDS \u6570\u636e\u79d1\u5b66\u6846\u67b6\u7684\u4e00\u90e8\u5206\u3002RAPIDS \u65e8\u5728\u5229\u7528 NVIDIA \u7684 CUDA \u6280\u672f\uff0c<strong>\u901a\u8fc7 GPU \u52a0\u901f\u6570\u636e\u79d1\u5b66\u548c\u5206\u6790\u5e94\u7528\u7a0b\u5e8f\u3002<\/strong><\/p>\n<p>\u200b\u00a0 \u00a0 \u00a0 \u00a0<\/p>\n<p>cuDF \u63d0\u4f9b\u4e86\u4e00\u4e2a\u7c7b\u4f3c\u4e8e Pandas \u7684 DataFrame \u63a5\u53e3\uff0c<strong>\u4f7f\u5f97\u5728 GPU \u4e0a\u8fdb\u884c\u6570\u636e\u5904\u7406\u548c\u5206\u6790\u53d8\u5f97\u66f4\u52a0\u9ad8\u6548\u548c\u5feb\u901f<\/strong>\u3002\u5b83\u4e0e Pandas API \u975e\u5e38\u5339\u914d\uff0c\u4f46\u5b83\u5e76\u6ca1\u6709\u5b8c\u5168\u53d6\u4ee3 Pandas\u3002cuDF \u548c Pandas \u4e4b\u95f4\u6709\u4e00\u4e9b\u76f8\u4f3c\u4e4b\u5904\u548c\u4e0d\u540c\u4e4b\u5904\u3002cuDF \u652f\u6301\u4e0e Pandas \u7c7b\u4f3c\u7684\u6570\u636e\u7ed3\u6784\u548c\u64cd\u4f5c\uff0c\u4f8b\u5982\u7d22\u5f15\u3001\u8fc7\u6ee4\u3001\u8fde\u63a5\u3001\u8fde\u63a5\u3001groupby \u7b49\u3002<\/p>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOZiaic3ZENeQKs7ffRLibP5AEeNRvj2n65qQ0DalLe7Jibt3U5Uic4qJSoyw\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOZiaic3ZENeQKs7ffRLibP5AEeNRvj2n65qQ0DalLe7Jibt3U5Uic4qJSoyw\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<h3><a id=\"%E4%BC%98%E7%82%B9\" class=\"anchor\" aria-hidden=\"true\"><span class=\"octicon octicon-link\"><\/span><\/a>\u4f18\u70b9<\/h3>\n<p>cuDF \u5229\u7528\u4e86 libcudf\uff08\u4e00\u4e2a\u8d85\u5feb\u7684\u00a0C++\/CUDA \u6570\u636e\u5e27\u5e93\uff09\u548c Apache Arrow \u5217\u683c\u5f0f\u6765\u63d0\u4f9b GPU \u52a0\u901f\u7684 pandas API\u3002<\/p>\n<p>\u5b83\u5177\u6709\u5982\u4e0b\u4f18\u70b9\u3002<\/p>\n<ul>\n<li>\n<p>**\u9ad8\u6027\u80fd\u8ba1\u7b97\uff1a**cuDF \u53ef\u4ee5\u663e\u8457\u52a0\u5feb\u6570\u636e\u5904\u7406\u548c\u5206\u6790\u4efb\u52a1\uff0c\u5c24\u5176\u662f\u5728\u6570\u636e\u6e05\u6d17\u3001\u8f6c\u6362\u548c\u805a\u5408\u7b49\u65b9\u9762\u3002<\/p>\n<\/li>\n<li>\n<p>**\u4e0e Pandas \u7c7b\u4f3c\u7684 API\uff1a**cuDF \u63d0\u4f9b\u4e86\u4e0e Pandas \u975e\u5e38\u76f8\u4f3c\u7684 API\uff0c\u8fd9\u964d\u4f4e\u4e86\u5b66\u4e60\u66f2\u7ebf\u3002<\/p>\n<\/li>\n<li>\n<p>**\u5185\u5b58\u6548\u7387\uff1a**cuDF \u901a\u8fc7\u5176\u9ad8\u6548\u7684\u5185\u5b58\u7ba1\u7406\u5728 GPU \u4e0a\u5b9e\u73b0\u4e86\u66f4\u9ad8\u7684\u6570\u636e\u5904\u7406\u6548\u7387\u3002<\/p>\n<\/li>\n<li>\n<p>**\u6613\u4e8e\u96c6\u6210\uff1a**cuDF \u53ef\u4ee5\u4e0e RAPIDS \u751f\u6001\u7cfb\u7edf\u4e2d\u7684\u5176\u4ed6\u5de5\u5177\uff08\u5982 cuML\u3001cuGraph\uff09\u65e0\u7f1d\u96c6\u6210\uff0c\u4e3a\u590d\u6742\u7684\u6570\u636e\u79d1\u5b66\u548c\u673a\u5668\u5b66\u4e60\u5de5\u4f5c\u6d41\u7a0b\u63d0\u4f9b\u652f\u6301\u3002<\/p>\n<\/li>\n<li>\n<p>**\u652f\u6301\u591a\u79cd\u6587\u4ef6\u683c\u5f0f\uff1a**cuDF \u652f\u6301\u591a\u79cd\u6d41\u884c\u7684\u6570\u636e\u683c\u5f0f\uff0c\u5982 CSV\u3001JSON\u3001Parquet \u7b49\uff0c\u65b9\u4fbf\u4e0e\u73b0\u6709\u6570\u636e\u5904\u7406\u6d41\u7a0b\u96c6\u6210\u3002<\/p>\n<\/li>\n<li>\n<p>**\u53ef\u6269\u5c55\u6027\uff1a**\u901a\u8fc7\u4e0e Dask \u7684\u96c6\u6210\uff0ccuDF \u652f\u6301\u5206\u5e03\u5f0f\u6570\u636e\u5904\u7406\uff0c\u53ef\u4ee5\u5904\u7406\u8d85\u51fa\u5355\u4e2a GPU \u5185\u5b58\u9650\u5236\u7684\u5927\u578b\u6570\u636e\u96c6\u3002<\/p>\n<\/li>\n<\/ul>\n<h3><a id=\"%E5%88%9D%E4%BD%93%E9%AA%8C\" class=\"anchor\" aria-hidden=\"true\"><span class=\"octicon octicon-link\"><\/span><\/a>\u521d\u4f53\u9a8c<\/h3>\n<h4><a id=\"%E5%BA%93%E7%9A%84%E5%AE%89%E8%A3%85\" class=\"anchor\" aria-hidden=\"true\"><span class=\"octicon octicon-link\"><\/span><\/a>\u5e93\u7684\u5b89\u88c5<\/h4>\n<h5><a id=\"cudagpu%E8%A6%81%E6%B1%82\" class=\"anchor\" aria-hidden=\"true\"><span class=\"octicon octicon-link\"><\/span><\/a>CUDA\/GPU \u8981\u6c42<\/h5>\n<ul>\n<li>\n<p>CUDA 11.2+<\/p>\n<\/li>\n<li>\n<p>NVIDIA \u9a71\u52a8\u7a0b\u5e8f 450.80.02+<\/p>\n<\/li>\n<li>\n<p>Pascal \u67b6\u6784\u6216\u66f4\u597d\uff08\u8ba1\u7b97\u80fd\u529b &gt;=6.0\uff09<\/p>\n<\/li>\n<\/ul>\n<h5><a id=\"%E5%AE%89%E8%A3%85\" class=\"anchor\" aria-hidden=\"true\"><span class=\"octicon octicon-link\"><\/span><\/a>\u5b89\u88c5<\/h5>\n<p>\u9996\u5148\uff0c\u9700\u8981\u9a8c\u8bc1 NVIDIA GPU \u662f\u5426\u8fd0\u884c\u6b63\u5e38\u3002<\/p>\n<pre class=\"line-numbers\"><code class=\"language-python\">!nvidia-smi\n<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOho8XlM3Gx5icR85VeWKl8slWmrsnvJxlWTLlyCbAX2Pl9dV9Y0mJy1g\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOho8XlM3Gx5icR85VeWKl8slWmrsnvJxlWTLlyCbAX2Pl9dV9Y0mJy1g\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<p>\u7136\u540e\u6211\u4eec\u76f4\u63a5\u4f7f\u7528 pip \u6765\u5b89\u88c5<\/p>\n<pre class=\"line-numbers\"><code class=\"language-python\">!pip\u00a0install\u00a0cudf-cu11\u00a0--extra-index-url=https:\/\/pypi.nvidia.com\n<\/code><\/pre>\n<p>\u6267\u884c\u5982\u4e0b\u547d\u4ee4\uff0c\u6765\u67e5\u770b\u5b89\u88c5\u662f\u5426\u6b63\u786e\u3002<\/p>\n<pre class=\"line-numbers\"><code class=\"language-python\">import\u00a0cudf\u00a0#\u00a0\u8fd9\u5e94\u8be5\u53ef\u4ee5\u6b63\u5e38\u5de5\u4f5c\uff0c\u6ca1\u6709\u4efb\u4f55\u9519\u8bef\ncudf.__version__\n<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOtibibRQR5thR3wUcWBKJicpLcTFyWMbbdqU4jJdr4J9iaV33NtUXqDvbibA\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOtibibRQR5thR3wUcWBKJicpLcTFyWMbbdqU4jJdr4J9iaV33NtUXqDvbibA\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<h4><a id=\"%E5%8A%A0%E8%BD%BD%E6%95%B0%E6%8D%AE%E9%9B%86\" class=\"anchor\" aria-hidden=\"true\"><span class=\"octicon octicon-link\"><\/span><\/a>\u52a0\u8f7d\u6570\u636e\u96c6<\/h4>\n<pre class=\"line-numbers\"><code class=\"language-python\">import pandas as pd\n\n# read 5 columns data:\ndf = pd.read_parquet(\n    &quot;https:\/\/data.rapids.ai\/datasets\/nyc_parking\/nyc_parking_violations_2022.parquet&quot;,\n    columns=[&quot;Registration State&quot;, &quot;Violation Description&quot;, &quot;Vehicle Body Type&quot;, &quot;Issue Date&quot;, &quot;Summons Number&quot;]\n)\n\n# view a random sample of 10 rows:\ndf.sample(10)\n<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOB8Bg6VVRpGH0wOaCHNLN68BM5unnDicC3YM1wVXf0iaPCOdqHzawpX3A\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOB8Bg6VVRpGH0wOaCHNLN68BM5unnDicC3YM1wVXf0iaPCOdqHzawpX3A\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<h4><a id=\"%E4%BD%BF%E7%94%A8%E6%A0%87%E5%87%86pandas%E8%BF%9B%E8%A1%8C%E5%88%86%E6%9E%90\" class=\"anchor\" aria-hidden=\"true\"><span class=\"octicon octicon-link\"><\/span><\/a>\u4f7f\u7528\u6807\u51c6 Pandas \u8fdb\u884c\u5206\u6790<\/h4>\n<p>\u8ba9\u6211\u4eec\u770b\u770b\u4f7f\u7528\u6807\u51c6 Pandas \u6267\u884c\u4ee3\u7801\u6240\u82b1\u8d39\u7684\u65f6\u95f4\u3002<\/p>\n<pre class=\"line-numbers\"><code class=\"language-python\">%%time\n(df[[&quot;Registration State&quot;, &quot;Violation Description&quot;]]\n .value_counts()\n .groupby(&quot;Registration State&quot;)\n .head(1)\n .sort_index()\n .reset_index()\n)\n<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOCriaicy85phYzNhUGgee4RdnHC0ppib5kR4aG2nB6UF0kpleDO5iaBtiaww\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOCriaicy85phYzNhUGgee4RdnHC0ppib5kR4aG2nB6UF0kpleDO5iaBtiaww\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<pre class=\"line-numbers\"><code class=\"language-python\">%%time\n\n(df\n .groupby([&quot;Vehicle Body Type&quot;])\n .agg({&quot;Summons Number&quot;: &quot;count&quot;})\n .rename(columns={&quot;Summons Number&quot;: &quot;Count&quot;})\n .sort_values([&quot;Count&quot;], ascending=False)\n)\n<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOibdsbNCibmGQ6QZZhMJ2bC0dLwKSSdG3l9eicYkRv4a8IuRaJib2yjzP5A\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOibdsbNCibmGQ6QZZhMJ2bC0dLwKSSdG3l9eicYkRv4a8IuRaJib2yjzP5A\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<p><strong>\u53ef\u4ee5\u770b\u5230\u4f7f\u7528 pandas \u4e0a\u7684\u8fd0\u884c\u65f6\u95f4\u5206\u522b\u662f 4.36 s \u548c 4.92 s\u3002<\/strong><\/p>\n<h4><a id=\"%E4%BD%BF%E7%94%A8cudf%E8%BF%9B%E8%A1%8C%E5%88%86%E6%9E%90\" class=\"anchor\" aria-hidden=\"true\"><span class=\"octicon octicon-link\"><\/span><\/a>\u4f7f\u7528 cudf \u8fdb\u884c\u5206\u6790<\/h4>\n<p>%load_ext\u00a0cudf.pandas<\/p>\n<pre class=\"line-numbers\"><code class=\"language-python\">%%time\n(df[[&quot;Registration State&quot;, &quot;Violation Description&quot;]]\n .value_counts()\n .groupby(&quot;Registration State&quot;)\n .head(1)\n .sort_index()\n .reset_index()\n)\n<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOaQN9Ejb6TibFtAhia0Gedk299XZ9X38lo7UzOjRnDVHWCy9ic4uu2cHpA\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOaQN9Ejb6TibFtAhia0Gedk299XZ9X38lo7UzOjRnDVHWCy9ic4uu2cHpA\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<pre class=\"line-numbers\"><code class=\"language-python\">%%time\n\n(df\n .groupby([&quot;Vehicle Body Type&quot;])\n .agg({&quot;Summons Number&quot;: &quot;count&quot;})\n .rename(columns={&quot;Summons Number&quot;: &quot;Count&quot;})\n .sort_values([&quot;Count&quot;], ascending=False)\n)\n<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOWdib8RldribibC4jG8VU1kaAHJibcuuqxibxNeeRAWEQS9KOullNn05b6sw\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOWdib8RldribibC4jG8VU1kaAHJibcuuqxibxNeeRAWEQS9KOullNn05b6sw\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<p>**\u53ef\u4ee5\u770b\u5230\u4f7f\u7528\u65f6\u95f4\u5206\u522b\u4e3a 196ms \u548c 27.9 ms\uff0c**<strong>\u548c\u4f7f\u7528 pandas \u76f8\u6bd4\uff0c\u6027\u80fd\u63d0\u9ad8\u4e86\u51e0\u767e\u500d\u3002<\/strong><\/p>\n<h4><a id=\"%E4%BA%86%E8%A7%A3%E6%80%A7%E8%83%BD\" class=\"anchor\" aria-hidden=\"true\"><span class=\"octicon octicon-link\"><\/span><\/a>\u4e86\u89e3\u6027\u80fd<\/h4>\n<p>cudf \u63d0\u4f9b\u5206\u6790\u5b9e\u7528\u7a0b\u5e8f\uff0c\u901a\u8fc7\u8bc6\u522b\u4ee3\u7801\u7684\u54ea\u4e9b\u90e8\u5206\u5728 GPU \u548c CPU \u4e0a\u8fd0\u884c\u6765\u5e2e\u52a9\u6211\u4eec\u66f4\u597d\u5730\u4e86\u89e3\u6027\u80fd\u3002<\/p>\n<pre class=\"line-numbers\"><code class=\"language-python\">%%cudf.pandas.profile\n\nsmall_df = pd.DataFrame({'a': [0, 1, 2], 'b': [&quot;x&quot;, &quot;y&quot;, &quot;z&quot;]})\nsmall_df = pd.concat([small_df, small_df])\n\naxis = 0\nfor i in range(0, 2):\n    small_df.min(axis=axis)\n    axis = 1\n\ncounts = small_df.groupby(&quot;a&quot;).b.count()\n<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dODVJPib4eeOuX0lR89JucYnP6phGeVxgTfuCZzDjiciaLWQ1mqQfdibLNLg\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dODVJPib4eeOuX0lR89JucYnP6phGeVxgTfuCZzDjiciaLWQ1mqQfdibLNLg\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<pre class=\"line-numbers\"><code class=\"language-python\">%%cudf.pandas.line_profile\n\nsmall_df = pd.DataFrame({'a': [0, 1, 2], 'b': [&quot;x&quot;, &quot;y&quot;, &quot;z&quot;]})\nsmall_df = pd.concat([small_df, small_df])\n\naxis = 0\nfor i in range(0, 2):\n    small_df.min(axis=axis)\n    axis = 1\n\ncounts = small_df.groupby(&quot;a&quot;).b.count()\n<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dO3MDY6tkziaFlOUz1p9WHicHwL9TNica43U1eE8mTzZcWZhC5PpyEIrkzQ\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dO3MDY6tkziaFlOUz1p9WHicHwL9TNica43U1eE8mTzZcWZhC5PpyEIrkzQ\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<p><strong>cuDF \u5e93\u4e0d\u652f\u6301\u7684\u64cd\u4f5c\u5c06\u81ea\u52a8\u56de\u9000\u5230\u6807\u51c6 Pandas\uff08\u5728 CPU \u4e0a\uff09\u3002<\/strong><\/p>\n<p>\u56e0\u6b64\uff0c\u7531\u4e8e\u6570\u636e\u5728 GPU \u548c CPU \u4e4b\u95f4\u590d\u5236\uff0c\u4f60\u53ef\u80fd\u4f1a\u9047\u5230\u6027\u80fd\u4f4e\u4e0b\u7684\u60c5\u51b5\u3002\u4f8b\u5982\uff0ccuDF \u76ee\u524d\u4e0d\u652f\u6301 count() \u51fd\u6570 axis \u7684\u53c2\u6570\u3002\u56e0\u6b64\uff0c\u6b64\u64cd\u4f5c\u5728 CPU \u4e0a\u6267\u884c\uff0c\u5e76\u4e14\u53ef\u80fd\u6bd4\u524d\u4e00\u4e2a\u64cd\u4f5c\u660e\u663e\u6162\u3002<\/p>\n<pre class=\"line-numbers\"><code class=\"language-python\">%%time\ndf.count()\n<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOE9aE406J2qnhdprROIDV3oic7aAUbdcRe3UmI3C6OgwUMTEkByp4T3Q\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOE9aE406J2qnhdprROIDV3oic7aAUbdcRe3UmI3C6OgwUMTEkByp4T3Q\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<pre class=\"line-numbers\"><code class=\"language-python\">%%time\ndf.count(axis=1)\n<\/code><\/pre>\n<p><div class='fancybox-wrapper lazyload-container-unload' data-fancybox='post-images' href='https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOHgiaRv0MGEiaTOHHsWaI0pib0khhjnic1ctP63kZrrBBRIC8NqKl17RicCw\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1'><img class=\"lazyload lazyload-style-2\" src=\"data:image\/svg+xml;base64,PCEtLUFyZ29uTG9hZGluZy0tPgo8c3ZnIHdpZHRoPSIxIiBoZWlnaHQ9IjEiIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyIgc3Ryb2tlPSIjZmZmZmZmMDAiPjxnPjwvZz4KPC9zdmc+\"  data-original=\"https:\/\/mmbiz.qpic.cn\/mmbiz_png\/ibPqnScajrbE9C1qXvPaCjWrDbSBLs0dOHgiaRv0MGEiaTOHHsWaI0pib0khhjnic1ctP63kZrrBBRIC8NqKl17RicCw\/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1\" src=\"data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsQAAA7EAZUrDhsAAAANSURBVBhXYzh8+PB\/AAffA0nNPuCLAAAAAElFTkSuQmCC\" alt=\"\u56fe\u7247\" \/><\/div><\/p>\n<p>\u6765\u6e90&#8212;&#8211;\u7a0b\u5e8f\u5458\u5b66\u957f<\/p>\n<!--CusAds0-->\n<div style=\"font-size: 0px; height: 0px; line-height: 0px; margin: 0; padding: 0; clear: both;\"><\/div>","protected":false},"excerpt":{"rendered":"<p>https:\/\/github.com\/rapidsai\/cudf cuDF \u662f\u4e00\u4e2a\u7531 NVIDIA \u5f00\u53d1\u7684 Python \u5e93\uff0c\u5b83\u662f RAPIDS \u6570\u636e\u79d1\u5b66\u6846\u67b6\u7684\u4e00\u90e8\u5206\u3002RAPIDS \u65e8\u5728\u5229\u7528 NVIDIA \u7684 CUDA \u6280\u672f\uff0c\u901a\u8fc7 GPU \u52a0\u901f\u6570\u636e\u79d1\u5b66\u548c\u5206\u6790\u5e94\u7528\u7a0b\u5e8f\u3002 \u200b\u00a0 \u00a0 \u00a0 \u00a0 cuDF \u63d0\u4f9b\u4e86\u4e00\u4e2a\u7c7b\u4f3c\u4e8e Pandas \u7684 DataFrame \u63a5\u53e3 &#8230;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_bbp_topic_count":0,"_bbp_reply_count":0,"_bbp_total_topic_count":0,"_bbp_total_reply_count":0,"_bbp_voice_count":0,"_bbp_anonymous_reply_count":0,"_bbp_topic_count_hidden":0,"_bbp_reply_count_hidden":0,"_bbp_forum_subforum_count":0},"categories":[12,71],"tags":[],"_links":{"self":[{"href":"https:\/\/zhang.mba\/index.php\/wp-json\/wp\/v2\/posts\/1952"}],"collection":[{"href":"https:\/\/zhang.mba\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/zhang.mba\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/zhang.mba\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/zhang.mba\/index.php\/wp-json\/wp\/v2\/comments?post=1952"}],"version-history":[{"count":0,"href":"https:\/\/zhang.mba\/index.php\/wp-json\/wp\/v2\/posts\/1952\/revisions"}],"wp:attachment":[{"href":"https:\/\/zhang.mba\/index.php\/wp-json\/wp\/v2\/media?parent=1952"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/zhang.mba\/index.php\/wp-json\/wp\/v2\/categories?post=1952"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/zhang.mba\/index.php\/wp-json\/wp\/v2\/tags?post=1952"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}