Re: i860 performance (was: Re: World's Most Powerful Computing Sites)

Article View: comp.sys.super

Article #485

#485

From: John D. McCalpin
Date: Fri, 22 Jan 1993 22:03

42 lines
3077 bytes

In article <C19s04.3yF@pgroup.com> lfm@pgroup.com (Larry Meadows) writes:
>>The problem is that peak mflops performance doesn't give the whole story --
>memory bandwidth is much more important.  Perhaps machines should be
>rated by how long it takes to do a vector add, or some such measure; then
>"super-vector" performance occurs when values can be left in vector registers.

This was the motivation for my "stream" benchmark.  I have not added any
new results lately, but the code, documentation, and lots of results are
in the anonymous ftp area of perelandra.cms.udel.edu in bench/stream/.

Some tidbits:

	 Memory Bandwidth results from STREAM_S and STREAM_D 
	 John D. McCalpin --- mccalpin@perelandra.cms.udel.edu

                    Bytes          Bandwidth (MB/s)    
Machine             /word      Copy     SSCAL       Sum     SAXPY   
---------------     -----  --------  --------  --------  --------   
Cray YMP/C90 16 cpu     8  105497.4  104656.4  101736.1  103812.8   
Cray YMP/C90  1 cpu     8    6965.4    6965.4    9378.7    9500.7   
Cray Y/MP 8 cpu         8   19291.6   19294.2   26588.9   26802.2   
Cray Y/MP 1 cpu         8    2426.4    2426.2    3454.4    3396.9   
CM-200  8k 10 MHz       8    5251.4    5251.3    5931.4    5653.2   
CM-2    8k 7 MHz        8    3224.5    3224.5    3657.4    3494.8   
NEC SX3/44 1 cpu        8   16941.0   15640.7   22436.5   21972.2   
NEC SX3/12 1 cpu        8    8302.2    7670.9    9712.8    9712.9   
ETA-10E                 8    2974.7    3010.5    4357.5    4214.8   
IBM 3090J/VF 1 cpu      8     251.9     250.1     288.5     264.6   
Convex C3420 2 cpu      8     289.1     273.1     302.7     300.3   
Convex C3410 1 cpu      8     177.6     172.5     179.0     178.4   
Convex C3220 2 cpu      8     334.3     333.0     337.3     336.5   
Convex C3210 1 cpu      8     180.1     179.4     181.4     180.8   
MIPS RC6280             8      53.5      58.2      56.5      56.5   
IBM RS6000-950          8     193.9     177.8     195.9     192.0   
HP 9000/730             8      53.3      48.0      55.4      55.4   
HP 9000/720             8      43.6      40.0      45.0      42.3   
---------------     -----  --------  --------  --------  --------   
-- 
--
John D. McCalpin                        mccalpin@perelandra.cms.udel.edu
Assistant Professor                     mccalpin@brahms.udel.edu
College of Marine Studies, U. Del.      John.McCalpin@mvs.udel.edu

Technical Details

Message-ID: <C19zxw.4C8@news.udel.edu>

Path: arc1-ca.pugleaf.net!sparky!uunet!paladin.american.edu!darwin.sura.net!news.udel.edu!perelandra.cms.udel.edu!mccalpin

References: <1993Jan21.165159.10149@meiko.com> <1993Jan22.015827.26653@nas.nasa.gov> <C19s04.3yF@pgroup.com>

Show Full Headers

Complete Article Headers

Newsgroups: comp.sys.super
Path: arc1-ca.pugleaf.net!sparky!uunet!paladin.american.edu!darwin.sura.net!news.udel.edu!perelandra.cms.udel.edu!mccalpin
From: mccalpin@perelandra.cms.udel.edu (John D. McCalpin)
Subject: Re: i860 performance (was: Re: World's Most Powerful Computing Sites)
Message-ID: <C19zxw.4C8@news.udel.edu>
Sender: usenet@news.udel.edu
Nntp-Posting-Host: perelandra.cms.udel.edu
Organization: College of Marine Studies, U. Del.
References: <1993Jan21.165159.10149@meiko.com> <1993Jan22.015827.26653@nas.nasa.gov> <C19s04.3yF@pgroup.com>
Date: Fri, 22 Jan 1993 22:03:31 GMT
Lines: 42

Reply to Parent Message