{"id":8323,"date":"2022-09-13T00:20:36","date_gmt":"2022-09-12T18:50:36","guid":{"rendered":"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/"},"modified":"2022-09-13T00:20:36","modified_gmt":"2022-09-12T18:50:36","slug":"solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array","status":"publish","type":"post","link":"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/","title":{"rendered":"[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array"},"content":{"rendered":"<p> [ad_1]<br \/>\n<\/p>\n<div id=\"answer-37174977\" class=\"answer js-answer accepted-answer js-accepted-answer\" data-answerid=\"37174977\" data-parentid=\"37149330\" data-score=\"0\" data-position-on-page=\"1\" data-highest-scored=\"1\" data-question-has-accepted-highest-score=\"1\" itemprop=\"acceptedAnswer\" itemscope itemtype=\"https:\/\/schema.org\/Answer\">\n<div class=\"post-layout\">\n<div class=\"votecell post-layout--left\"><\/div>\n<div class=\"answercell post-layout--right\">\n<div class=\"s-prose js-post-body\" itemprop=\"text\">\n<p>I think you&#8217;re really trying to ask about branchless vs. branching ways to do <code>m = max(m, array[i])<\/code>.  C compilers will already compile the <code>if()<\/code> version to branchless code (using <code>cmov<\/code>) depending on optimization settings.  It can even auto-vectorize to a packed compare or packed-max function.<\/p>\n<p>Your <code>0.5 * abs()<\/code> version is obviously terrible (a lot slower than a conditional-move) because it converts to <code>double<\/code> and back.   rather than dividing by two with a right shift.<\/p>\n<p>See the asm on <a rel=\"nofollow noopener\" target=\"_blank\" href=\"http:\/\/gcc.godbolt.org\/#compilers:!((compiler:g530,options:'-xc+-O3++-Wall++-fno-verbose-asm+-march%3Dhaswell+-mno-avx',source:'%23include+%3Cstdlib.h%3E%0A%0A\/\/+conversion+to\/from+double,+and+an+FP+multiply,+in+the+loop-carried+dependency+chain%0Aint+maxarray_float(int+arr%5B%5D,+int+n)+%7B%0A++int+result+%3D+arr%5B0%5D%3B%0A++for+(int+i%3D0+%3B+i%3Cn%3B+%2B%2Bi)+%7B%0A++++int+tmp+%3D+arr%5Bi%5D%3B%0A++++result+%3D++0.5+*+(+result+%2B+tmp+%2B+abs(result-tmp))%3B%0A++%7D%0A++return+result%3B%0A%7D%0A%0A%0A\/\/+auto-vectorizes+to+PMAXSD,+or+without+SSE4.1,+to+pcmpgt+\/+pand\/por+emulation+of+the+same%0Aint+maxarray_if(int+arr%5B%5D,+int+n)+%7B%0A++int+result+%3D+arr%5B0%5D%3B%0A++for+(int+i%3D0+%3B+i%3Cn%3B+%2B%2Bi)+%7B%0A++++int+tmp+%3D+arr%5Bi%5D%3B%0A++++if+(result+%3C+tmp)%0A++++++result+%3D+tmp%3B%0A++%7D%0A++return+result%3B%0A%7D%0A')),filterAsm:(commentOnly:!t,directives:!t,labels:!t),version:3\">the Godbolt Compiler Explorer<\/a>:<\/p>\n<pre><code>\/\/ auto-vectorizes to PMAXSD, or without SSE4.1, to pcmpgt \/ pand\/por emulation of the same\nint maxarray_if(int arr[], int n) {\n  int result = arr[0];\n  for (int i=0 ; i&lt;n; ++i) {\n    int tmp = arr[i];\n    if (result &lt; tmp)\n      result = tmp;\n  }\n  return result;\n}\n<\/code><\/pre>\n<p>gcc 5.3 <code>-O3 -march=haswell -mno-avx<\/code> auto-vectorized inner loop:<\/p>\n<pre><code>.L13:\n    add     eax, 1\n    pmaxsd  xmm0, XMMWORD PTR [rdx]\n    add     rdx, 16\n    cmp     r8d, eax\n    ja      .L13\n<\/code><\/pre>\n<p>Vs. the FP version:<\/p>\n<pre><code>    ... whole bunch of integer crap\n    cvtsi2sd        xmm0, eax\n    mulsd   xmm0, xmm1\n    cvttsd2si       eax, xmm0\n<\/code><\/pre>\n<p>So the FP version is obviously complete garbage.<\/p>\n<p>You&#8217;ll get similar results for any target architecture.  The conversion to\/from <code>double<\/code> isn&#8217;t going to go away.  gcc keeps it even with <code>-ffast-math<\/code>.<\/p>\n<\/p><\/div>\n<div class=\"mt24\"><\/div>\n<\/div>\n<p>            <span class=\"d-none\" itemprop=\"commentCount\">2<\/span> <\/p><\/div>\n<\/div>\n<p>[ad_2]<\/p>\n<p>solved Comparing 2 different scenarios on 2 different architectures when finding the max element of an array <\/p>\n","protected":false},"excerpt":{"rendered":"<p>[ad_1] I think you&#8217;re really trying to ask about branchless vs. branching ways to do m = max(m, array[i]). C compilers will already compile the if() version to branchless code (using cmov) depending on optimization settings. It can even auto-vectorize to a packed compare or packed-max function. Your 0.5 * abs() version is obviously terrible &#8230; <a title=\"[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array\" class=\"read-more\" href=\"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/\" aria-label=\"More on [Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array\">Read more<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[320],"tags":[324,1564,2357,2356,1369],"class_list":["post-8323","post","type-post","status-publish","format-standard","hentry","category-solved","tag-c","tag-cuda","tag-gpgpu","tag-opencl","tag-optimization"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v26.5 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array - JassWeb<\/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:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array - JassWeb\" \/>\n<meta property=\"og:description\" content=\"[ad_1] I think you&#8217;re really trying to ask about branchless vs. branching ways to do m = max(m, array[i]). C compilers will already compile the if() version to branchless code (using cmov) depending on optimization settings. It can even auto-vectorize to a packed compare or packed-max function. Your 0.5 * abs() version is obviously terrible ... Read more\" \/>\n<meta property=\"og:url\" content=\"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/\" \/>\n<meta property=\"og:site_name\" content=\"JassWeb\" \/>\n<meta property=\"article:published_time\" content=\"2022-09-12T18:50:36+00:00\" \/>\n<meta name=\"author\" content=\"Kirat\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Kirat\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"1 minute\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/\"},\"author\":{\"name\":\"Kirat\",\"@id\":\"https:\/\/jassweb.com\/solved\/#\/schema\/person\/65c9c7b7958150c0dc8371fa35dd7c31\"},\"headline\":\"[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array\",\"datePublished\":\"2022-09-12T18:50:36+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/\"},\"wordCount\":147,\"publisher\":{\"@id\":\"https:\/\/jassweb.com\/solved\/#organization\"},\"keywords\":[\"c++\",\"cuda\",\"gpgpu\",\"opencl\",\"optimization\"],\"articleSection\":[\"Solved\"],\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/\",\"url\":\"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/\",\"name\":\"[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array - JassWeb\",\"isPartOf\":{\"@id\":\"https:\/\/jassweb.com\/solved\/#website\"},\"datePublished\":\"2022-09-12T18:50:36+00:00\",\"breadcrumb\":{\"@id\":\"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/jassweb.com\/solved\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/jassweb.com\/solved\/#website\",\"url\":\"https:\/\/jassweb.com\/solved\/\",\"name\":\"JassWeb\",\"description\":\"Build High-quality Websites\",\"publisher\":{\"@id\":\"https:\/\/jassweb.com\/solved\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/jassweb.com\/solved\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/jassweb.com\/solved\/#organization\",\"name\":\"Jass Web\",\"url\":\"https:\/\/jassweb.com\/solved\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/jassweb.com\/solved\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/jassweb.com\/wp-content\/uploads\/2021\/02\/jass-website-logo-1.png\",\"contentUrl\":\"https:\/\/jassweb.com\/wp-content\/uploads\/2021\/02\/jass-website-logo-1.png\",\"width\":693,\"height\":132,\"caption\":\"Jass Web\"},\"image\":{\"@id\":\"https:\/\/jassweb.com\/solved\/#\/schema\/logo\/image\/\"}},{\"@type\":\"Person\",\"@id\":\"https:\/\/jassweb.com\/solved\/#\/schema\/person\/65c9c7b7958150c0dc8371fa35dd7c31\",\"name\":\"Kirat\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/jassweb.com\/solved\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/jassweb.com\/solved\/wp-content\/litespeed\/avatar\/1261af3c9451399fa1336d28b98ea3bb.jpg?ver=1775798750\",\"contentUrl\":\"https:\/\/jassweb.com\/solved\/wp-content\/litespeed\/avatar\/1261af3c9451399fa1336d28b98ea3bb.jpg?ver=1775798750\",\"caption\":\"Kirat\"},\"sameAs\":[\"http:\/\/jassweb.com\"],\"url\":\"https:\/\/jassweb.com\/solved\/author\/jaspritsinghghumangmail-com\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array - JassWeb","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:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/","og_locale":"en_US","og_type":"article","og_title":"[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array - JassWeb","og_description":"[ad_1] I think you&#8217;re really trying to ask about branchless vs. branching ways to do m = max(m, array[i]). C compilers will already compile the if() version to branchless code (using cmov) depending on optimization settings. It can even auto-vectorize to a packed compare or packed-max function. Your 0.5 * abs() version is obviously terrible ... Read more","og_url":"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/","og_site_name":"JassWeb","article_published_time":"2022-09-12T18:50:36+00:00","author":"Kirat","twitter_card":"summary_large_image","twitter_misc":{"Written by":"Kirat","Est. reading time":"1 minute"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/#article","isPartOf":{"@id":"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/"},"author":{"name":"Kirat","@id":"https:\/\/jassweb.com\/solved\/#\/schema\/person\/65c9c7b7958150c0dc8371fa35dd7c31"},"headline":"[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array","datePublished":"2022-09-12T18:50:36+00:00","mainEntityOfPage":{"@id":"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/"},"wordCount":147,"publisher":{"@id":"https:\/\/jassweb.com\/solved\/#organization"},"keywords":["c++","cuda","gpgpu","opencl","optimization"],"articleSection":["Solved"],"inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/","url":"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/","name":"[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array - JassWeb","isPartOf":{"@id":"https:\/\/jassweb.com\/solved\/#website"},"datePublished":"2022-09-12T18:50:36+00:00","breadcrumb":{"@id":"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/jassweb.com\/solved\/solved-comparing-2-different-scenarios-on-2-different-architectures-when-finding-the-max-element-of-an-array\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/jassweb.com\/solved\/"},{"@type":"ListItem","position":2,"name":"[Solved] Comparing 2 different scenarios on 2 different architectures when finding the max element of an array"}]},{"@type":"WebSite","@id":"https:\/\/jassweb.com\/solved\/#website","url":"https:\/\/jassweb.com\/solved\/","name":"JassWeb","description":"Build High-quality Websites","publisher":{"@id":"https:\/\/jassweb.com\/solved\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/jassweb.com\/solved\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/jassweb.com\/solved\/#organization","name":"Jass Web","url":"https:\/\/jassweb.com\/solved\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/jassweb.com\/solved\/#\/schema\/logo\/image\/","url":"https:\/\/jassweb.com\/wp-content\/uploads\/2021\/02\/jass-website-logo-1.png","contentUrl":"https:\/\/jassweb.com\/wp-content\/uploads\/2021\/02\/jass-website-logo-1.png","width":693,"height":132,"caption":"Jass Web"},"image":{"@id":"https:\/\/jassweb.com\/solved\/#\/schema\/logo\/image\/"}},{"@type":"Person","@id":"https:\/\/jassweb.com\/solved\/#\/schema\/person\/65c9c7b7958150c0dc8371fa35dd7c31","name":"Kirat","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/jassweb.com\/solved\/#\/schema\/person\/image\/","url":"https:\/\/jassweb.com\/solved\/wp-content\/litespeed\/avatar\/1261af3c9451399fa1336d28b98ea3bb.jpg?ver=1775798750","contentUrl":"https:\/\/jassweb.com\/solved\/wp-content\/litespeed\/avatar\/1261af3c9451399fa1336d28b98ea3bb.jpg?ver=1775798750","caption":"Kirat"},"sameAs":["http:\/\/jassweb.com"],"url":"https:\/\/jassweb.com\/solved\/author\/jaspritsinghghumangmail-com\/"}]}},"_links":{"self":[{"href":"https:\/\/jassweb.com\/solved\/wp-json\/wp\/v2\/posts\/8323","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/jassweb.com\/solved\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/jassweb.com\/solved\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/jassweb.com\/solved\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/jassweb.com\/solved\/wp-json\/wp\/v2\/comments?post=8323"}],"version-history":[{"count":0,"href":"https:\/\/jassweb.com\/solved\/wp-json\/wp\/v2\/posts\/8323\/revisions"}],"wp:attachment":[{"href":"https:\/\/jassweb.com\/solved\/wp-json\/wp\/v2\/media?parent=8323"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/jassweb.com\/solved\/wp-json\/wp\/v2\/categories?post=8323"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/jassweb.com\/solved\/wp-json\/wp\/v2\/tags?post=8323"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}