home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.arch:9401 comp.parallel:2117
- Newsgroups: comp.arch,comp.parallel
- Path: sparky!uunet!gatech!hubcap!fpst
- From: olson@anchor.esd.sgi.com (Dave Olson)
- Subject: Re: sgi/sun shared memory machines question
- Message-ID: <1992Sep14.122957.16270@hubcap.clemson.edu>
- Apparently-To: comp-parallel@uunet.uu.net
- Sender: news@zuni.esd.sgi.com (Net News)
- Organization: Silicon Graphics, Inc. Mountain View, CA
- References: <1992Sep11.121830.17069@hubcap.clemson.edu>
- Date: Sun, 13 Sep 92 02:00:34 GMT
- Approved: parallel@hubcap.clemson.edu
- Lines: 33
-
- (I've removed the Distribution line; why limit things like this
- to the US?)
-
- In <1992Sep11.121830.17069@hubcap.clemson.edu> zilla@nj.nec.com (John Lewis) writes:
- | Someone told me that the SGI multiprocessor machines cannot run
- | different unix processes on different processors; they can only
- | run a single (but possibly parallelized) process at a given time.
-
- This is completely incorrect. You can run multiple threads of
- one process, or completely seperate processes on each CPU.
-
- | The reason was said to be that the machines do not implement
- | any mechanism for simultaneous writes.
-
- There is a bus called the MP or Sync bus that handles cache
- coherency issues. If a process with multiple threads, or
- seperate processes use shared memory, they need to provide
- some mechanism (normally the hardware spinlocks or higher
- level library routines based on them) to avoid attempts
- at simultaneous access to the same memory locations (if
- they care).
-
- | Is the new sun multiprocessor (4/690?) similar?
-
- It is 'similar' in that multiple processes can run on multiple
- CPUs. The exact methods used for semaphoring are different.
- --
- Let no one tell me that silence gives consent, | Dave Olson
- because whoever is silent dissents. | Silicon Graphics, Inc.
- Maria Isabel Barreno | olson@sgi.com
-
-
-
-