diff --git a/doc/html/news.shtml b/doc/html/news.shtml index f6787b8521923a963f7f13e2d89d116c613c1664..4b8ed1b9fa7ceaad2cbcb221459b62d1651efcab 100644 --- a/doc/html/news.shtml +++ b/doc/html/news.shtml @@ -4,29 +4,12 @@ <h2>Index</h2> <ul> -<li><a href="#24">SLURM Version 2.4, June 2012</a></li> <li><a href="#25">SLURM Version 2.5, November 2012</a></li> <li><a href="#26">SLURM Version 2.6, scheduled for April 2013</a></li> <li><a href="#27">SLURM Version 2.7 and beyond</a></li> <li><a href="#security">Security Patches</a></li> </ul> -<h2><a name="24">Major Updates in SLURM Version 2.4</a></h2> -<p>SLURM Version 2.4 was release in June 2012. -Major enhancements include: -<ul> -<li>Major improvement in job throuhgput: Up to 500 jobs per second (actual -throughput depends upon configuration, hardware, etc.).</li> -<li>Support for BlueGene/Q computers.</li> -<li>Resource reservations with a node count specification will select -those nodes optimized for the system topology.</li> -<li>Support for job and step constraints with resource counts (e.g. a job -request for two nodes with a feature of "Intel" plus four nodes with a feature -of "AMD").</li> -<li>New tool <i>sdiag</i> reports scheduling statistics.</li> -<li>Support for SLURM commands operating over a LoadLeveler resource manager -(in branch of SLURM code, not in main release).</li> -</ul> <h2><a name="25">Major Updates in SLURM Version 2.5</a></h2> <p>SLURM Version 2.5 was release in November 2012. @@ -34,14 +17,15 @@ Major enhancements include: <ul> <li>User control over CPU frequency of each job step.</li> <li>Recording power usage information for each job.</li> -<li>Job specification of desired CPU frequency, which will effect it's -power usage and possibly its execution cost.</li> <li>Advanced reservation of cores rather than whole nodes.</li> <li>Integration with IBM's <a href="http://www-03.ibm.com/systems/software/parallel/">Parallel Environment</a> including POE (Parallel Operating Environment) and NRT (Network Resource Table) API.</li> <li>Highly optimized thoughput for serial jobs in a new "select/serial" plugin.</li> +<li>CPU load is information available</li> +<li>Configurable number of CPUs available to jobs in each SLURM partition, +which provides a mechanism to reserve CPUs for use with GPUs.</li> </ul> <h2><a name="26">Major Updates in SLURM Version 2.6</a></h2> @@ -49,15 +33,16 @@ plugin.</li> Major enhancements planned include: <ul> <li><a href="http://en.wikipedia.org/wiki/MapReduce">MapReduce</a> support -(launches ~1000x fater, runs ~10x faster).</li> +(launches ~1000x faster, runs ~10x faster).</li> <li>Scheduling optimized for energy efficiency.</li> <li>Integration with <a href="http://en.wikipedia.org/wiki/FlexNet_Publisher">FLEXlm (Flexnet Publisher)</a> license management.</li> <li>Faster and more powerful job step management support (e.g. step dependencies).</li> -<li>Improved control over scheduling of CPUs and GPUs with multiple jobs per -node (<i>MaxCPUsPerNode</i> allocated to all jobs in a partition/queue).</li> +<li>Support for Intel MIC (Many Integrated Core) processor.</li> +<li>Finer-grained BlueGene resouce management (partitions/queues and advanced +reservations containing less than a whole midplane).</li> </ul> <h2><a name="27">Major Updates in SLURM Version 2.7 and beyond</a></h2>