Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!purdue!decwrl!sun-barr!apple!rutgers!rochester!pt.cs.cmu.edu!ius3.ius.cs.cmu.edu!ralphw
From: ralphw@ius3.ius.cs.cmu.edu (Ralph Hyre)
Newsgroups: comp.arch
Subject: Interested in memory bandwidth figures for DEC PMAX & Sparcstation 1
Message-ID: <4993@pt.cs.cmu.edu>
Date: 15 May 89 17:44:43 GMT
Organization: Carnegie-Mellon University, CS/RI
Lines: 20

I'm evaluating platforms for our vision work, which can be characterized
as memory bandwidth intensive.

A pointer to a document sustained and burst for various transfer sizes (up 
to 1Meg/xfer) would be ideal.  Are there any benchmarks out there that try
to measure various dimensions of memory bandwidth?  (DMA vs. CPU block moves,
bcopy times, time to xfer a 1M file from disk out through the ethernet.)

Currently leaning toward Sparcstation 1 because we have a chance to add
framegrabbers to it and DMA is well-supported.  Only moderately disappointed
that FP performance isn't better (why wasn't the TI 8847 FP chip used?
Cost, I presume.)  Somewhat concered about the smaller cache size
(64K vs. 128K for DEC)

Will DEC have an R3000 box before Sun gets the Sparcstation 2 out the door?
-- 
					- Ralph W. Hyre, Jr.
Internet: ralphw@{ius{3,2,1}.,}cs.cmu.edu    Phone:(412) CMU-BUGS
Amateur Packet Radio: N3FGW@W2XO, or c/o W3VC, CMU Radio Club, Pittsburgh, PA
"You can do what you want with my computer, but leave me alone!8-)"
--