<?xml version="1.0"?>
<oembed><version>1.0</version><provider_name>Number Crunch</provider_name><provider_url>https://www.numbercrunch.de/blog</provider_url><author_name>Heiko Bauke</author_name><author_url>https://www.numbercrunch.de/blog/author/admin/</author_url><title>CUDA-aware MPI - Number Crunch</title><type>rich</type><width>600</width><height>338</height><html>&lt;blockquote class="wp-embedded-content" data-secret="c3k9fxx5zX"&gt;&lt;a href="https://www.numbercrunch.de/blog/2014/04/cuda-aware-mpi/"&gt;CUDA-aware MPI&lt;/a&gt;&lt;/blockquote&gt;&lt;iframe sandbox="allow-scripts" security="restricted" src="https://www.numbercrunch.de/blog/2014/04/cuda-aware-mpi/embed/#?secret=c3k9fxx5zX" width="600" height="338" title="&#x201C;CUDA-aware MPI&#x201D; &#x2014; Number Crunch" data-secret="c3k9fxx5zX" frameborder="0" marginwidth="0" marginheight="0" scrolling="no" class="wp-embedded-content"&gt;&lt;/iframe&gt;&lt;script type="text/javascript"&gt;
/* &lt;![CDATA[ */
/**
 * WordPress inline HTML embed
 *
 * @since 4.4.0
 * @output wp-includes/js/wp-embed.js
 *
 * Single line comments should not be used since they will break
 * the script when inlined in get_post_embed_html(), specifically
 * when the comments are not stripped out due to SCRIPT_DEBUG
 * being turned on.
 */
(function ( window, document ) {
	'use strict';

	/* Abort for ancient browsers. */
	if ( ! document.querySelector || ! window.addEventListener || typeof URL === 'undefined' ) {
		return;
	}

	/** @namespace wp */
	window.wp = window.wp || {};

	/* Abort if script was already executed. */
	if ( !! window.wp.receiveEmbedMessage ) {
		return;
	}

	/**
	 * Receive embed message.
	 *
	 * @param {MessageEvent} e
	 */
	window.wp.receiveEmbedMessage = function( e ) {
		var data = e.data;

		/* Verify shape of message. */
		if (
			! ( data || data.secret || data.message || data.value ) ||
			/[^a-zA-Z0-9]/.test( data.secret )
		) {
			return;
		}

		var iframes = document.querySelectorAll( 'iframe[data-secret="' + data.secret + '"]' ),
			blockquotes = document.querySelectorAll( 'blockquote[data-secret="' + data.secret + '"]' ),
			allowedProtocols = new RegExp( '^https?:$', 'i' ),
			i, source, height, sourceURL, targetURL;

		for ( i = 0; i &lt; blockquotes.length; i++ ) {
			blockquotes[ i ].style.display = 'none';
		}

		for ( i = 0; i &lt; iframes.length; i++ ) {
			source = iframes[ i ];

			if ( e.source !== source.contentWindow ) {
				continue;
			}

			source.removeAttribute( 'style' );

			if ( 'height' === data.message ) {
				/* Resize the iframe on request. */
				height = parseInt( data.value, 10 );
				if ( height &gt; 1000 ) {
					height = 1000;
				} else if ( ~~height &lt; 200 ) {
					height = 200;
				}

				source.height = height;
			} else if ( 'link' === data.message ) {
				/* Link to a specific URL on request. */
				sourceURL = new URL( source.getAttribute( 'src' ) );
				targetURL = new URL( data.value );

				if (
					allowedProtocols.test( targetURL.protocol ) &amp;&amp;
					targetURL.host === sourceURL.host &amp;&amp;
					document.activeElement === source
				) {
					window.top.location.href = data.value;
				}
			}
		}
	};

	function onLoad() {
		var iframes = document.querySelectorAll( 'iframe.wp-embedded-content' ),
			i, source, secret;

		for ( i = 0; i &lt; iframes.length; i++ ) {
			/** @var {IframeElement} */
			source = iframes[ i ];

			secret = source.getAttribute( 'data-secret' );
			if ( ! secret ) {
				/* Add secret to iframe */
				secret = Math.random().toString( 36 ).substring( 2, 12 );
				source.src += '#?secret=' + secret;
				source.setAttribute( 'data-secret', secret );
			}

			/*
			 * Let post embed window know that the parent is ready for receiving the height message, in case the iframe
			 * loaded before wp-embed.js was loaded. When the ready message is received by the post embed window, the
			 * window will then (re-)send the height message right away.
			 */
			source.contentWindow.postMessage( {
				message: 'ready',
				secret: secret
			}, '*' );
		}
	}

	window.addEventListener( 'message', window.wp.receiveEmbedMessage, false );
	document.addEventListener( 'DOMContentLoaded', onLoad, false );
})( window, document );

/* ]]&gt; */
&lt;/script&gt;
</html><description>CUDA and MPI provide two different APIs for parallel programming that target very different parallel architectures. While CUDA allows to utilize parallel graphics hardware for general purpose computing, MPI is usually employed to write parallel programs that run on large SMP systems or on cluster computers. In order to improve a cluster&#x2019;s overall computational capabilities&hellip; Continue reading</description></oembed>
