diff options
Diffstat (limited to 'benchmarks/stream/index.html')
| -rw-r--r-- | benchmarks/stream/index.html | 241 |
1 files changed, 241 insertions, 0 deletions
diff --git a/benchmarks/stream/index.html b/benchmarks/stream/index.html new file mode 100644 index 0000000..1da4651 --- /dev/null +++ b/benchmarks/stream/index.html @@ -0,0 +1,241 @@ +<!doctype html public "-//w3c//dtd html 4.0 transitional//en"> +<html> +<head> + <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1"> + <meta name="description" content="Sustainable memory bandwidth benchmark, with results on a wide variety of computer systems, from Mac's and PC's to most current and recent workstations to Cray supercomputers."> + <meta name="keywords" content="bandwidth,memory bandwidth,benchmarks,high performance computing, computer performance measurements,computer performance modelling,machine balance,vector processing"> + <meta name="GENERATOR" content="Mozilla/4.7 [en] (X11; I; Linux 2.2.15pre3 ppc) [Netscape]"> + <title>MEMORY BANDWIDTH: STREAM BENCHMARK PERFORMANCE RESULTS</title> +</head> +<body bgcolor="#FFFFFF"> +<img SRC="stream_logo.gif" ALT="STREAM Logo (Image)" SGI_SRC="./stream_logo.gif" height=240 width=320 align=RIGHT><b><a href="http://www.cs.virginia.edu">Department +of Computer Science</a></b> +<br><a href="http://www.cs.virginia.edu/~seas/">School of Engineering and +Applied Science</a> +<br><a href="http://www.virginia.edu/">University of Virginia</a>, <a href="http://www.virginia.edu/cville.html">Charlottesville, +Virginia</a> +<br> +<hr> +<h1> +STREAM: Sustainable Memory Bandwidth in High Performance Computers</h1> + +<hr> +<h2> +Table of Contents</h2> + +<ul> +<li> +<a href="new.html">What's New?</a></li> + +<ul> +<li> +Version 5 of the STREAM benchmark is available in Fortran! <a href="new.html">Check +it out.</a></li> + +<li> + <a href="ref.html">STREAM FAQ</a> has been significantly revised +and enhanced, with more discussion of how to run the STREAM benchmark. +Check it out!</li> + +<li> +The Archives of Original Submissions of STREAM benchmark results have been +updated through April 18, 2000! All of the results have made it into the +tables, but the links from the tables back to the original submissions +are still being (slowly & manually) added.</li> +</ul> + +<li> +<a href="#ListofTables">Here are the current RESULTS!</a></li> + +<li> +<a href="ref.html">STREAM FAQ</a></li> + +<li> +<a href="analyses.html">Analyses, Commentary, etc....</a></li> + +<li> +<a href="stream_mail/index.html">Hypermail archives of original contributions</a></li> + +<li> +FTP access to code and data:</li> + +<br>Many people have reported that the anonymous ftp service is often unavailable, +especially on weekends. I do not know what is happening, but the service +does get restarted, and seems to be up more often during the work week, +so keep on trying! +<ul> +<li> +<a href="ftp://ftp.cs.virginia.edu/pub/stream/">Top of tree</a> (includes +source code and raw database files).</li> + +<li> +<a href="ftp://ftp.cs.virginia.edu/pub/stream/Code/">Source Code</a></li> +</ul> +</ul> + +<hr> +<h2> +<a NAME="ListofTables"></a>Here are the current RESULTS!</h2> + +<h3> +Top 10 Results for Shared-Memory Systems!</h3> +This set of results includes the top 10 shared-memory systems, ranked by +STREAM TRIAD performance. The results are currently presented in the following +tables: +<ul> +<li> +<a href="top10/Bandwidth.html">Main Table - Bandwidth in MB/s</a></li> + +<li> +<a href="top10/Balance.html">Machine Balance - relative cost of memory +accesses vs arithmetic</a></li> + +<li> +<a href="top10/MFLOPS.html">Equivalent MFLOPS</a></li> +</ul> + +<hr> +<h3> +Standard Results</h3> +The "standard" set of results presents the results of the C or Fortran +versions of the STREAM benchmark running with 64-bit data types on production +hardware. The "standard" set of results <b>excludes</b>: +<ul> +<li> +Cases with 32-bit operands or operations.</li> + +<li> +Cases with significantly modified code or assembly language.</li> + +<li> +Simulated results.</li> + +<li> +Results from experimental or non-production hardware or software.</li> +</ul> +The results are currently presented in the following tables: +<ul> +<li> +<a href="standard/Bandwidth.html">Main Table - Bandwidth in MB/s, alphabetized +by vendor (mostly)</a></li> + +<li> +<a href="standard/Balance.html">Machine Balance - relative cost of memory +accesses vs arithmetic</a></li> + +<li> +<a href="standard/MFLOPS.html">Equivalent MFLOPS</a></li> +</ul> + +<hr WIDTH="100%"> +<h3> +PC-compatible Results</h3> +This set of tables summarizes the "standard" test cases, but restricted +to IBM PC-compatible computers. +<p>Users are free to re-compile the <a href="ftp://ftp.cs.virginia.edu/pub/stream/Code/">code</a> +or use the new <a href="ftp://ftp.cs.virginia.edu/pub/stream/Contrib/MasonCabot/win32/wstream.exe">NT</a> +or <a href="ftp://ftp.cs.virginia.edu/pub/stream/Contrib/MasonCabot/linux/stream_l">Linux</a> +binaries. Use of the old DOS binaries is discouraged. +<p>The results are currently presented in the following tables: +<ul> +<li> +<a href="peecee/Bandwidth.html">Main Table - Bandwidth in MB/s</a></li> + +<li> +<a href="peecee/Balance.html">Machine Balance - relative cost of memory +accesses vs arithmetic</a></li> + +<li> +<a href="peecee/MFLOPS.html">Equivalent MFLOPS</a></li> +</ul> + +<hr WIDTH="100%"> +<h3> +Macintosh-Compatible Results</h3> +This set of tables summarizes the "standard" test cases, but restricted +to Macintosh and compatible computers. +<p>Users are free to re-compile the <a href="ftp://ftp.cs.virginia.edu/pub/stream/Code/">code</a> +or use the <a href="ftp://ftp.cs.virginia.edu/pub/stream/STREAM.sea.hqx">contributed +binaries.</a> +<p>The results are currently presented in the following tables: +<ul> +<li> +<a href="mac/Bandwidth.html">Main Table - Bandwidth in MB/s</a></li> + +<li> +<a href="mac/Balance.html">Machine Balance - relative cost of memory accesses +vs arithmetic</a></li> + +<li> +<a href="mac/MFLOPS.html">Equivalent MFLOPS</a></li> +</ul> + +<h3> + +<hr WIDTH="100%"></h3> + +<h3> +Experimental/Nonstandard Results</h3> +These tables include <b>only</b> results that are +<ul> +<li> +Assembly language coded, or</li> + +<li> +Simulated, or</li> + +<li> +Based on experimental or non-production hardware, or</li> + +<li> +Based on <a href="#note">partially depopulated</a> systems.</li> +</ul> +<a NAME="note"></a><b>Note:</b> +<br>A "partially depopulated" system is one in which only a subset of the +cpus are used for the benchmark, and for which this subset is spread around +the machine to decrease contention. For example on the SGI Origin2000, +each node has 2 cpus sharing a single bus and memory subsystem. The results +in this table labelled "1 per node" are based on using only one cpu per +node board, and are considered a "nonstandard" way of using the machine. +Similarly, the Sun Ultra10000 has 4 cpus per node board, so results using +1, 2, or 3 cpus per node also go into this table of "nonstandard" results. <img SRC="new01.gif" ALT="NEW!" SGI_SRC="./new01.gif" > +<p>The results are currently presented in the following tables: +<ul> +<li> +<a href="experimental/Bandwidth.html">Main Table - Bandwidth in MB/s, alphabetized +by vendor (mostly)</a></li> + +<li> +<a href="experimental/Balance.html">Machine Balance - relative cost of +memory accesses vs arithmetic</a></li> + +<li> +<a href="experimental/MFLOPS.html">Equivalent MFLOPS</a></li> +</ul> + +<hr> +<h3> +32-bit Results</h3> +These tables include <b>only</b> results with 32-bit operands or operations. +Results using 64-bit operands that move the data in 32-bit "chunks" are +<b>not</b> +here. The results are currently presented in the following tables: +<ul> +<li> +<a href="thirty-two/Bandwidth.html">Main Table - Bandwidth in MB/s, alphabetized +by vendor (mostly)</a></li> + +<li> +<a href="thirty-two/Balance.html">Machine Balance - relative cost of memory +accesses vs arithmetic</a></li> + +<li> +<a href="thirty-two/MFLOPS.html">Equivalent MFLOPS</a></li> +</ul> + +<hr> +<address> +<i>John D. McCalpin <a href="mailto:mccalpin@cs.virginia.edu">mccalpin@cs.virginia.edu</a></i></address> + +</body> +</html> |