{"json_modified": "2016-10-23T02:59:57.936369", "uuid": "5cd90974-bc21-4800-9aa8-ddeee42f5ed4", "title": "Les types SIMD", "url": "/fr/docs/Web/JavaScript/Types_SIMD", "tags": ["Avanc\u00e9", "SIMD", "JavaScript"], "translations": [{"uuid": "9f61a658-f658-4f51-98e6-7bb4c595c3b4", "title": "SIMD types", "url": "/en-US/docs/Web/JavaScript/SIMD_types", "tags": ["SIMD", "JavaScript"], "summary": "The experimental JavaScript SIMD API introduces vector objects that utilize SIMD/SSE instructions on supporting CPUs; <a class=\"glossaryLink\" href=\"/en-US/docs/Glossary/SIMD\" title=\"SIMD: SIMD (pronounced &quot;seem-dee&quot;) is short for Single Instruction/Multiple Data which is one classification of computer architectures. SIMD allows one same operation to be performed on multiple data points resulting in data level parallelism and thus performance gains, for example for 3D graphics and video processing, physics simulations or cryptography, and other domains.\">SIMD</a> is short for Single Instruction/Multiple Data.<em> SIMD operations</em> are methods that process multiple data with a single instruction. In contrary, <em>scalar operations</em> (<a class=\"glossaryLink\" href=\"/en-US/docs/Glossary/SISD\" title=\"SISD: SISD is short for Single Instruction/Single Data which is one classification of computer architectures. In SISD architecture, a single processor executes a single instruction and operates on a single data point in memory.\">SISD</a>) process only one individual data with a single instruction.", "localization_tags": [], "locale": "en-US", "last_edit": "2016-10-20T13:23:21", "review_tags": []}, {"uuid": "fb4d8b51-0aee-456f-9ce6-7c993b97dc8c", "title": "SIMD types", "url": "/ja/docs/Web/JavaScript/SIMD_types", "tags": [], "summary": "The experimental JavaScript SIMD API introduces vector objects that utilize SIMD/SSE instructions on supporting CPUs; <a class=\"glossaryLink\" href=\"/en-US/docs/Glossary/SIMD\" title=\"SIMD: SIMD (pronounced &quot;seem-dee&quot;) is short for Single Instruction/Multiple Data which is one classification of computer architectures . SIMD allows one same operation to be performed on multiple data points resulting in data level parallelism and thus performance gains, for example for 3D graphics and video processing, physics simulations or cryptography, and other domains.\">SIMD</a> is short for Single Instruction/Multiple Data.<em> SIMD operations</em> are methods that process multiple data with a single instruction. In contrary, <em>scalar operations</em> (<a class=\"glossaryLink\" href=\"/en-US/docs/Glossary/SISD\" title=\"SISD: SISD is short for Single Instruction/SIngle Data which is one classification of computer architectures. In SISD architecture, a single processor executes a single instruction and operates on a single data point in memory.\">SISD</a>) process only one individual data with a single instruction.", "localization_tags": ["inprogress"], "locale": "ja", "last_edit": "2015-07-26T07:03:46", "review_tags": []}, {"uuid": "7aa6666b-5e13-4af1-955a-f9c54a30e41b", "title": "SIMD types", "url": "/zh-CN/docs/Web/JavaScript/SIMD_types", "tags": [], "summary": "\u5173\u4e8eSIMD\u7684\u7f3a\u70b9\uff0c\u8fd9\u4e5f\u662f\u4e3a\u4f55\u7b97\u6cd5\u8981\u9488\u5bf9SIMD\u8fdb\u884c\u8bbe\u8ba1\u3002\u5f53\u4f60\u7684\u7b97\u6cd5\u8981\u5206\u522b\u5904\u7406\u4e0d\u540c\u7684\u6570\u636e\u65f6\uff0c\u6570\u636e\u96c6\u4e2d\u7684\u4e0d\u540c\u6570\u636e\u5e76\u4e0d\u80fd\u88ab\u533a\u5206\u5904\u7406\u3002\u5728\u540e\u9762\u7684\u6587\u7ae0\u4e2d\u6211\u4eec\u5c06\u77e5\u9053\u5982\u4f55\u4f7f\u7528\u8499\u677f\u4ee5\u53ca\u5982\u4f55\u5c06\u6570\u636e\u91cd\u65b0\u5bf9\u9f50\u6765\u89e3\u51b3\u8fd9\u4e00\u95ee\u9898\u3002", "localization_tags": ["inprogress"], "locale": "zh-CN", "last_edit": "2015-12-01T01:41:18", "review_tags": []}], "modified": "2016-10-23T02:59:57", "label": "Les types SIMD", "localization_tags": [], "locale": "fr", "id": 151967, "last_edit": "2016-10-23T02:59:50", "summary": "Traiter un ensemble de donn\u00e9es avec une seule instruction peut apporter des gains de performances cons\u00e9quents \u00e0 votre application. Ces gains d\u00e9pendront, entre autres, de la taille de l'ensemble de donn\u00e9es, ou vecteur, qu'on peut manipuler. Pour cette raison, les op\u00e9rations SIMD sont largement utilis\u00e9es pour le graphisme en 3D, le traitement audio/vid\u00e9o, les simulations physiques, la cryptographie et d'autres domaines.", "sections": [{"id": "Quick_Links", "title": null}, {"id": "SIMD_et_JavaScript", "title": "SIMD et JavaScript"}, {"id": "R\u00e9aligner_des_donn\u00e9es_pour_utiliser_des_vecteurs_SIMD", "title": "R\u00e9aligner des donn\u00e9es pour utiliser des vecteurs SIMD"}, {"id": "Parall\u00e9liser_des_branches_conditionnelles", "title": "Parall\u00e9liser des branches conditionnelles"}, {"id": "Cr\u00e9er_des_branches_appliquer_des_masques_s\u00e9lectionner", "title": "Cr\u00e9er des branches, appliquer des masques, s\u00e9lectionner"}, {"id": "D'autres_algorithmes_et_cas_d'utilisation_pertinents_pour_SIMD", "title": "D'autres algorithmes et cas d'utilisation pertinents pour SIMD"}, {"id": "Statut_d'impl\u00e9mentation_pour_SIMD_en_JavaScript", "title": "Statut d'impl\u00e9mentation pour SIMD en JavaScript"}, {"id": "Voir_aussi", "title": "Voir aussi"}], "slug": "Web/JavaScript/Types_SIMD", "review_tags": []}