home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!uwm.edu!ee!bloc1469
- From: bloc1469@ee.ee.uwm.edu (Gregory R Block)
- Newsgroups: comp.sys.amiga.hardware
- Subject: Re: Data/Instruction Cache & BURST modes on 68030? Why/when?
- Date: 26 Dec 1992 19:42:04 GMT
- Organization: Electrical Engineering Dept. University of Wisconsin - Milwaukee
- Lines: 48
- Message-ID: <1hicicINN5t3@uwm.edu>
- References: <72192@cup.portal.com> <smcgerty.725067816@unix1.tcd.ie> <1992Dec23.212118.3625@nic.csu.net>
- NNTP-Posting-Host: 129.89.2.33
-
- In article <1992Dec23.212118.3625@nic.csu.net> matthewd@csufres.CSUFresno.EDU (Matthew Davidian) writes:
- >The main rationale I've read for the "Harvard" or split cache setup is that
- >they can both be accessed at the same time; in pipelined architectures this
- >can be especially good, since in one stage you might be fetching & decoding
- >an instruction, while another stage might be fetching the operand for another
- >instruction; both at the same time.
-
- Right. Harvard architecture allows for a more useful pipeline. To
- say the LEAST. :)
-
- >A side effect of having data and instruction caches is that self modifying
- >code (which is not a good thing anyway) won't work--your writes to modify
- >the code will go into the data cache (because it's assumed to be data) and
- >the instruction caches doesn't get the change.
-
- Well, again, we find self-modifying code a problem. Just as a
- side-point to make, C= and Apple have both said to avoid
- self-modifying code. I have never found a purpose for self-modifying
- code, except in the case of copy protection, which I think shouldn't
- exist.
-
- >In an article I read before the P5/586/Pentium was announced, the writer
- >swore that Intel would never produce an 80x86 with a Harvard architecture
- >because so many DOS programs use self-modifying code--including Windows.
- >In order to pull this off, they must have added some extra stuff to check
- >data writes against the instruction cache, or some such magic, or else
- >the caches will have to be disabled for self-modifying code.
-
- How quickly things change when performance becomes an issue over
- compatibility, eh? I doubt seriously they have--if anything, if you
- fall back into compatibility mode (if it is true that the P5's native
- mode isn't compatible with the 486 and previous), you probably lose
- both caches.
-
- Probably.
-
- Adding a ton of kludges wouldn't be beyond Intel, but it does appear
- that for once they're trying to shrug off the kludges of their past
- and start anew--I don't believe they'd do this by introducing Yet
- Another Kludge.
-
- Greg
-
- --
- (: (: (: (: Have you overdosed on smileys today? Why NOT!?! :) :) :) :)
- (: "Our father, who art in Iowa, Hollow be thy head, Thy ideas run :)
- (: Thy will be done, At Commodore as it is at Apple" -Dan Barrett :)
- (: (: (: (: (: (: (: (: (: (: (: (: () :) :) :) Wubba, the Dark Angel :)
-