Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uwm.edu!spool.mu.edu!uunet!rbj From: rbj@uunet.UU.NET (Root Boy Jim) Newsgroups: comp.unix.internals Subject: Re: Compressed executables Message-ID: <120776@uunet.UU.NET> Date: 1 Feb 91 05:27:03 GMT Organization: UUNET Communications Services, Falls Church, VA Lines: 18 >In article <1991Jan30.123119.15448@chvax.uucp> rheiger@chvax.UUCP (Eiger Richard) writes: >:In article <5453@bwdls58.UUCP> hwt@bwdlh490.BNR.CA (Henry Troup) writes: >:>Of course, if you can implement compression in special hardware between the >:>disk and the CPU, you can do this with no impact on CPU performance. >:> >:I thought of the exact same thing. While we're at it, why not put >:hardware compressing onto the drives? The problem here is that disk drivers access data by block, optimizing sector rotation and cylinder head motion. If you're gonna send it all to the disk anyway, you don't save any transfer time. Compressing the data changes the boundarys where the sectors fall. Many people think it a bad idea to move layout policy out of the kernel; others would rather let a smart caching disk do it. -- Root Boy Jim Cottrell Close the gap of the dark year in between