Path: utzoo!attcan!uunet!munnari!ditmela!george
From: george@ditmela.oz (George michaelson)
Newsgroups: comp.arch
Subject: memory speed & futurology
Message-ID: <2179@ditmela.oz>
Date: 24 Aug 88 02:02:09 GMT
Organization: CSIRO Division of Information Technology, Australia
Lines: 71


	The August edition of 'electronics' is about memory technology
current & futures. it had this table suggesting price/performance trends
in the next 10 years. I suspect we see another sub-generation of workstations
coming online before this, but a 5 year design timescale might mean the sun-6
series and its peers could take advantage of these changes.

(from Electronics, Aug '88 page 70, reproduced without permission)

	HOW SYSTEM STORAGE WILL CHANGE 
	==============================

		Cache		Main Mem	Magnetic Disk	Optical Disk
Cost/Mb:	
 1987		$500		$250		$6.30		$3.50 - $5.50
 1995		$ 60		$ 30		$1.75		  under 50c
Av access time:
 1987		45-55ns		120-150ns	15-20ms		35-100ms
 1995		15-20ns		under 25ns	 10ms		20-25ms


My questions:

(1)	how plausible are their figures? I have no other info to hand.

(2)	main memory gets 6 times faster and 8 times cheaper. cache gets
	2-3 times faster and 8 times cheaper. Thus their speeds almost
	converge. Does that make cache sufficiently unattactive to stop 
	being used?

	-price/speed ratio looks ugly (twice the cost for saving 5-10ns)
	-overall access speed to memory is 2 X current speed to cache so
	for existing processor architectures you can possibly do without it
	unless there are `logical' reasons for using cache other than
	buffering for slower memory.

(3)	do the new speeds still look good alongside predicted clock speeds
	for CPU or do we have another development lag here?

	there is another table in the mag showing possible access times
	for existing 32bit cpus & available speeds from different memory
	technology but it's hard to reproduce. It implies memory access
	delay is one of many  bottlenecks, I think these speedups might cure 
	it but only for existing clockspeeds.

(4)	at $30/Mb do we start to get ramdisk coming back into fashion?

	$3k for a 100meg `drive' looks pretty neat, assuming there are
	architectural reasons for not making it simply look like resident
	memory.

(5)	do we start to get 32/64Mb by default in our workstations? 
	does the opsys change its memory usage when that much memory
	is around? 

(6)	with optical disk getting down to current access times for `real'
	disk do they become standard or are the disadvantages still too
	great? does unix sprout file version numbers? -I'm assuming WORM speeds
	are similar to pre-recorded speeds here...

(7)	dual ported memory & video rams: do they benifit too?

(8)	put your own question here!

	-george

-- 
        George Michaelson, CSIRO Division of Information Technology

ACSnet: G.Michaelson@ditmela.oz                      Phone: +61 3 347 8644
Postal: CSIRO, 55 Barry St, Carlton, Vic 3053 Oz       Fax: +61 3 347 8987