home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!destroyer!ncar!noao!arizona!saunders
- From: saunders@cs.arizona.edu (Richard T. Saunders)
- Newsgroups: comp.sys.amiga.hardware
- Subject: Re: Multiprocessing
- Message-ID: <26315@optima.cs.arizona.edu>
- Date: 12 Nov 92 17:04:53 GMT
- References: <1992Nov7.211453.21687@ultb.isc.rit.edu> <boing.721262416@mcl> <1992Nov10.234916.6311@ultb.isc.rit.edu>
- Sender: news@cs.arizona.edu
- Organization: U of Arizona CS Dept, Tucson
- Lines: 53
-
- In article <1992Nov10.234916.6311@ultb.isc.rit.edu> snm6394@ultb.isc.rit.edu (S.N. Mozumder ) writes:
- >In article <boing.721262416@mcl> boing@mcl.ucsb.edu (Geff Hanoian) writes:
- >>In <1992Nov7.211453.21687@ultb.isc.rit.edu> snm6394@ultb.isc.rit.edu (S.N. Mozumder ) writes:
- >>
- >>>To folks at CBM:
- >>
- >>>Are there any plans for a multiprocessing or RISC based Amiga? My last
- >>>rendering took 25 hours on Lightwave 3-D on a 030/882-25 MHz Amiga for
- >>
- >> [ ray tracing stuff cut out]
- >>
- >>>multiprocessing? Radius has such boards for 68040 multiprocessing on
- >>>Macintoshes. Already maximum performance on the macintosh has exceeded
- >>>the Amigas in ray tracing with the Radius boards. Help us videographers
- >>>out here. 25 hours is a whole day of not doing work on the computer.
- >>
- >> But see ... on an AMIGA it's not a whole day without doing
- >> work on the computer. We can multitask unlike the MAC.
- >But the Toaster Doesn't multitask while rendering in Lightwave 3-D.
- >
- >Bobby Mozumder
-
- Be careful to distuinguish between multi-TASKING and multi-PROCESSING.
- Multi-PROCESSING implies that you have >1 processors working in parallel
- (let's not count the blitter and Copper for the moment ... usually
- a multi-processor has "general purpose" processors). Multi-TASKING
- implies >1 "logical" processors, i.e., you don't HAVE to have >1
- processors.
-
- Why am I being so careful in my nomenclature? Well, if the Macintosh actually
- has plug in boards that allow you to "Multi-PROCESS", then I would say that
- is a serious advantage over the Amiga. Multi_PROCESSING is the wave of
- the future---it's the only way we'll be able to compute faster because
- soon processors will hit the "Speed of Light" barrier.
-
- What work is being done, or what work has been done, on upgrading the
- Amiga to multiprocessor status?
-
- (As an aside, I remember reading in the RKM:Hardware that TSET (a 680x0
- instruction used the in synchronization of multiprocessors on a shared-memory
- bus) is discouraged from use because it "ties up" (?) buses on the Amiga
- architecture. Without this "atomic" instruction, It seems like it would
- be difficult to have a multiprocessing Amiga, at least in the "Shared
- Memory Multiprocessor" sense. Any comments? Does anybody know why
- TSET is to be avoided?)
-
- Gooday,
-
- Richie
-
-
-
-
-