[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Gabriel,DN10000 Benchmarks

	From att!TITAN.KSC.NASA.GOV!DUMOULIN Tue Mar  6 13:11:23 EDT 1990
	Received: by att.att.com; Tue Mar  6 16:40:23 1990
	Received: from Warbucks.AI.SRI.COM by IU.AI.SRI.COM via SMTP with TCP;
		  Tue, 6 Mar 90 10:46:31-PST
	Received: from TITAN.KSC.NASA.GOV by Warbucks.AI.SRI.COM with INTERNET
		  ; Tue, 6 Mar 90 10:42:40 PST
	Received: by TITAN id <00000227141@TITAN.KSC.NASA.GOV> ; Tue,
		  6 Mar 90 13:36:45 EDT
	Date: Tue,  6 Mar 90 13:11:23 EDT
	Subject: Gabriel,DN10000 Benchmarks
	To: slug@Warbucks.AI.SRI.COM
	X-VMS-Mail-To: EXOS%"slug@ai.sri.com"
	Message-ID: <900306131123.00000227141@TITAN.KSC.NASA.GOV>
	  In response to the message about "Why is the Symbolics so slow"...
	  You are comparing:
	           Symbolics 3645 with 12Mbytes made in 1986 with unspecified
	           amount of paging (probably on an ST-506 disk)
	           Symbolics 3675 with 16Mbytes made in 1986 with unspecified
	           amount of paging (probably on an SMD disk)
	           Symbolics 3650 with 32Mbytes made in 1986 with unspecified
	           amount of paging (probably on an ESDI disk)
	           Each of these systems are not up to the latest microcode or
	           patch releases.
	    To a   Apollo/HP Domain DN10000 running with 49 Mbytes of memory,
	           two CPU's and a BETA version of the Software dated a few
	           months ago!
	     PERHAPS comparing your DN10000 to an XL400 or XL1200 with the
	     same amount of wired memory, same paging size AND same type
	     of disk interface (ie both SMD, ESDI or SCSI) AND running the
	     latest microcode and S/W release would be fairer.
I'm aware that the Symbolicses were not the most recent versions.  My
question was not why the Symbolicses were so much slower than the DN10000
but why the 3650 was so slow compared with the 3675 and 3645.

David Loewenstern
AT&T Bell Laboratories
Whippany, NJ 07981
email: davel@whutt.att.com || whutt!davel
at&t: 201-386-6516