home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!gatech!paladin.american.edu!howland.reston.ans.net!wupost!spool.mu.edu!umn.edu!msus1.msus.edu!msus1.msus.edu!news
- Newsgroups: comp.sys.amiga.hardware
- Subject: Re: Data/Instruction Cache & BURST modes on 68030? Why/when?
- Message-ID: <1992Dec24.121546.1956@msus1.msus.edu>
- From: lkoop@TIGGER.STCLOUD.MSUS.EDU (LaMonte Koop)
- Date: 24 Dec 92 12:15:45 -0600
- Reply-To: lkoop@TIGGER.STCLOUD.MSUS.EDU
- References: <hellerS.724958427@batman> <72192@cup.portal.com><1992Dec22.212901.860@samba.oit.unc.edu>,<72290@cup.portal.com>
- Organization: SCS GP/Engineering Cluster
- Nntp-Posting-Host: tigger.stcloud.msus.edu
- Lines: 38
-
- In article <72290@cup.portal.com>, Tony-Preston@cup.portal.com (ANTHONY FRANCIS PRESTON) writes:
- >|>You should leave the setting alone! A small tutorial is in order!
- >|>SetPatch will setup the system correctly so that the caches and burst
- >|>are correct for the Amiga.
- >|>
- >|>several long words of data. This happens faster than if each long
- >|>word was a separate access. This pre-loads the cache and allows you
- >|>to get extra 'hits' on the instruction cache(or data cache). There
- >|>are times(code that branches) where this might slow down the cpu, but
- >|>in general it wil speed it up slightly.
- >|>
- >|>The 68020 has a single cache for data and instructions, the 68030
- >|>has separate ones. The cache size in this case is 256 bytes. The
- >|>68040 has 4096 byte caches for instruction and for data.
- >|>
- >|>There are no such things as stupid, dumb, or ignorant questions, just
- >|>stupid, dumb, or ignorant people that will not ask questions.
- >|
- >|Thanks a lot for this info. I have always known about the cache and when
- >|I use sysinfo it says data cache off so I put a command in my SS to turn
- >|it on thinking I was missing something. I guess I'll turn it off. Since
- >
- > If you turn on the data cache, you run the risk of crashes if the chipe
- > memory is data cached. This is due to the custom co-processors accessing
- > the chip memory and changing location while the cpu has the data cached.
-
- This is only on a badly designed system where cache management of such is not
- done in hardware. The CSA MMR DOES properly do this, so the data cache is
- not a problem in this respect, and should be used, and can be used when
- desired. The only setup I am aware of which had the problem of inproper cache
- managment of cacheable address spaces is the older Ronin Hurricane accelerators
- (the 68030 versions).
-
- ----------------------------------------
- LaMonte Koop -- SCSU Electrical/Computer Engineering
- Internet: lkoop@tigger.stcloud.msus.edu -OR- f00012@kanga.stcloud.msus.edu
- "You mean you want MORE lights on this thing???"
- ---------------------------------------------------------------------------
-