home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.protocols.nfs
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!sdd.hp.com!hpscdc!cupnews0.cup.hp.com!tardy.cup.hp.com!raj
- From: raj@cup.hp.com (Rick Jones)
- Subject: Re: NFS I/O Ops/seconds
- Message-ID: <Bs2EJu.E4n@cup.hp.com>
- Date: Mon, 27 Jul 1992 20:35:52 GMT
- Organization: The Place Where Packets Go To Die
- X-Newsreader: Tin 1.1 PL4
- Lines: 17
-
- On the topic of saving 1000 instructions here or there, I would say that the
- number of instructions for an individual request on the surface doesn't
- look like much, but if you are talking about a big NFS server, it could
- gain you several score IOP/s.
-
- Saving trips up and down the protocl stack is a "good thing." If you can get
- there in a way that assists in data movement in the host is an "even better
- thing."
-
- 8KB in the cache that doesn't need to be there is a bad thing. Even if you
- have caches that are bigger than the core memories of most machines from
- 10 years ago.
-
- Waste is still waste, even in the face of plenty.
-
- rick jones
-
-