{"id":8289,"date":"2014-02-18T17:26:30","date_gmt":"2014-02-18T16:26:30","guid":{"rendered":"http:\/\/www.planet3dnow.de\/cms\/?p=8289"},"modified":"2014-02-18T22:01:50","modified_gmt":"2014-02-18T21:01:50","slug":"pgi-compiler-sammlung-openacc-2-0-amd-gpu-apu","status":"publish","type":"post","link":"https:\/\/www.planet3dnow.de\/cms\/8289-pgi-compiler-sammlung-openacc-2-0-amd-gpu-apu\/","title":{"rendered":"PGI-Compiler-Sammlung mit Unterst\u00fctzung f\u00fcr OpenACC 2.0 und AMD<\/span> GPUs ver\u00f6ffentlicht"},"content":{"rendered":"

Die Com\u00adpi\u00adler-Exper\u00adten von The Port\u00adland Group<\/em> (PGI<\/span>), die mitt\u00adler\u00adwei\u00adle zu NVIDIA<\/span> geh\u00f6\u00adren, haben ihre PGI<\/span> 2014 Com\u00adpi\u00adlers &<\/span> Tools<\/em> ver\u00ad\u00f6f\u00adfent\u00adlicht, mit denen Unter\u00adst\u00fct\u00adzung f\u00fcr die aktu\u00adel\u00adle Ver\u00adsi\u00adon 2.0 des Open\u00adACC-Stan\u00addards Ein\u00adzug h\u00e4lt und zudem erst\u00admals die APUs und GPUs von AMD<\/span> unter\u00adst\u00fctzt wer\u00adden. Ent\u00adspre\u00adchend freut sich AMDs Ser\u00adver-Mana\u00adger Suresh Gopa\u00adl\u00ada\u00adkrish\u00adn\u00adan, dass neben Intel Xeon Phi und NVIDIA<\/span> Tes\u00adla jetzt auch die Beschleu\u00adni\u00adger von AMD<\/span> \u00fcber die\u00adses Pro\u00adgram\u00admier\u00admo\u00addell ange\u00adspro\u00adchen wer\u00adden k\u00f6nnen.<\/p>\n

\u201c<\/span>We applaud PGI<\/span>\u2019s abili\u00adty to extra\u00adct per\u00adfor\u00admance from AMD<\/span> dis\u00adcrete GPUs and APUs using Open\u00adACC. It will help break down the remai\u00adning bar\u00adriers to wide-sca\u00adle acce\u00adle\u00adra\u00adtor adop\u00adti\u00adon, and decou\u00adple the choice of acce\u00adle\u00adra\u00adtor pro\u00adgramming model from the choice of acce\u00adle\u00adra\u00adtor hard\u00adware.\u201d \u2014 Suresh Gopa\u00adl\u00ada\u00adkrish\u00adn\u00adan, Cor\u00adpo\u00adra\u00adte Vice Pre\u00adsi\u00addent und Gene\u00adral Mana\u00adger des Ser\u00adver Busi\u00adness bei AMD<\/span><\/p><\/blockquote>\n

Open\u00adACC ist eine im HPC-Umfeld sehr belieb\u00adte API<\/span>, die \u00e4hn\u00adlich wie OpenMP auf Direk\u00adti\u00adven basiert und spe\u00adzi\u00adell f\u00fcr den Ein\u00adsatz von Beschleu\u00adni\u00adger\u00adkar\u00adten von CAPS<\/span>, Cray, NVIDIA<\/span> und PGI<\/span> erschaf\u00adfen<\/a> wur\u00adde. Im Fal\u00adle der AMD-Hard\u00adware setzt der Com\u00adpi\u00adler die Direk\u00adti\u00adven dann nicht durch Nut\u00adzung von CUDA<\/span>, son\u00addern Open\u00adCL um. Die Zusam\u00admen\u00adar\u00adbeit zwi\u00adschen PGI<\/span> und AMD<\/span><\/a> zur Imple\u00admen\u00adtie\u00adrung des Open\u00adCL-Back-Ends f\u00fcr den Open\u00adACC-Com\u00adpi\u00adler wur\u00adde Ende Okto\u00adber 2012 bekannt\u00adge\u00adge\u00adben. Aller\u00addings hat NVIDIA<\/span> kurz nach der Bereit\u00adstel\u00adlung der ers\u00adten Beta\u00adver\u00adsi\u00adon Ende Juli 2013<\/a> PGI<\/span> auf\u00adge\u00adkauft. Wie es mit der Unter\u00adst\u00fct\u00adzung von AMD-Hard\u00adware nach der Fer\u00adtig\u00adstel\u00adlung des gemein\u00adsa\u00admen Pro\u00adjekts wei\u00adter\u00adgeht, bleibt daher abzuwarten.<\/p>\n

In der Pres\u00adse\u00admit\u00adtei\u00adlung beschreibt NVIDIA<\/span> jeden\u00adfalls wei\u00adter\u00adhin das Ziel, eine platt\u00adform\u00adun\u00adab\u00adh\u00e4n\u00adgi\u00adge Com\u00adpi\u00adler-Samm\u00adlung zu ent\u00adwi\u00adckeln, die eine her\u00adaus\u00adra\u00adgen\u00adde Per\u00adfor\u00admance auf unter\u00adschied\u00adli\u00adchen Sys\u00adte\u00admen bie\u00adten soll. Ein \u00e4hn\u00adli\u00adches Pro\u00adgram\u00admier\u00admo\u00addell ver\u00adsucht aktu\u00adell AMD<\/span> in Koope\u00adra\u00adti\u00adon mit SUSE<\/span> bei der Imple\u00admen\u00adtie\u00adrung eines HSA-Back-Ends in die GNU<\/span> Com\u00adpi\u00adler Coll\u00adec\u00adtion<\/em> (GCC<\/span>) zu nut\u00adzen. Der gew\u00e4hl\u00adte Weg f\u00fchrt hier \u00fcber die OpenMP-Direk\u00adti\u00adven in der Ver\u00adsi\u00adon 4.0, die bereits seit vie\u00adlen Jah\u00adren zur Par\u00adal\u00adle\u00adli\u00adsie\u00adrung auf Mehr\u00adkern- und Mul\u00adti-Pro\u00adzes\u00adsor\u00adsys\u00adte\u00admen zum Ein\u00adsatz kom\u00admen. AMD<\/span> und SUSE<\/span> ori\u00aden\u00adtie\u00adren sich hier\u00adbei an den von Open\u00adACC genutz\u00adten Direktiven.<\/p>\n

\u201c<\/span>The use of acce\u00adle\u00adra\u00adtors in high per\u00adfor\u00admance com\u00adpu\u00adting is now main\u00adstream. With PGI<\/span> 2014, we are taking ano\u00adther big step toward our goal of pro\u00advi\u00adding plat\u00adform-inde\u00adpen\u00addent, mul\u00adti-core and acce\u00adle\u00adra\u00adtor pro\u00adgramming tools that deli\u00adver out\u00adstan\u00adding per\u00adfor\u00admance on mul\u00adti\u00adple plat\u00adforms wit\u00adhout the need for exten\u00adsi\u00adve, device-spe\u00adci\u00adfic tuning.\u201d \u2014 Dou\u00adglas Miles, Direc\u00adtor of PGI<\/span> Soft\u00adware bei NVIDIA<\/span><\/p><\/blockquote>\n

Alle wei\u00adte\u00adren Neue\u00adrun\u00adgen der PGI<\/span> 2014 Com\u00adpi\u00adlers &<\/span> Tools<\/em> k\u00f6n\u00adnen PGI 2014 Compilers &amp; Tools<\/em>\" target=\"_blank\">in den offi\u00adzi\u00adel\u00adlen Release Notes<\/a> des Her\u00adstel\u00adlers nach\u00adge\u00adle\u00adsen werden.<\/p>\n

Quel\u00adle:<\/strong> Pres\u00adse\u00admit\u00adtei\u00adlung<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

ie Com\u00adpi\u00adler-Exper\u00adten von The Port\u00adland Group (PGI<\/span>), die mitt\u00adler\u00adwei\u00adle zu NVIDIA<\/span> geh\u00f6\u00adren, haben ihre PGI<\/span> 2014 Com\u00adpi\u00adlers &<\/span> Tools ver\u00ad\u00f6f\u00adfent\u00adlicht, mit denen Unter\u00adst\u00fct\u00adzung f\u00fcr die aktu\u00adel\u00adle Ver\u00adsi\u00adon 2.0 des Open\u00adACC-Stan\u00addards Ein\u00adzug h\u00e4lt und zudem erst\u00admals die APUs und GPUs von AMD<\/span> unter\u00adst\u00fctzt wer\u00adden. (\u2026) Wei\u00adter\u00adle\u00adsen \u00bb<\/a><\/p>\n","protected":false},"author":8,"featured_media":8290,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"wp_typography_post_enhancements_disabled":false,"ngg_post_thumbnail":0,"footnotes":""},"categories":[12,11],"tags":[966,970,539,109,540,992,538,537],"class_list":["post-8289","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-aktuelles","category-news","tag-amd","tag-apu","tag-compiler","tag-gpu","tag-hpc","tag-nvidia","tag-openacc","tag-pgi","entry"],"share_on_mastodon":{"url":"","error":""},"_links":{"self":[{"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/posts\/8289","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/comments?post=8289"}],"version-history":[{"count":4,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/posts\/8289\/revisions"}],"predecessor-version":[{"id":8305,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/posts\/8289\/revisions\/8305"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/media\/8290"}],"wp:attachment":[{"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/media?parent=8289"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/categories?post=8289"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.planet3dnow.de\/cms\/wp-json\/wp\/v2\/tags?post=8289"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}