home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.linux
- Path: sparky!uunet!cis.ohio-state.edu!zaphod.mps.ohio-state.edu!moe.ksu.ksu.edu!usenet-feed.cc.umr.edu!mcs213f.cs.umr.edu!dminer
- From: dminer@mcs213f.cs.umr.edu (Dan Miner)
- Subject: Compilation speeds (was Re: No patchlevel 3.}
- References: <1992Sep7.215429.7866@umr.edu> <1992Sep8.002147.790@wam.umd.edu>
- Date: Tue, 8 Sep 1992 18:17:30 GMT
- Nntp-Posting-Host: mcs213f.cs.umr.edu
- Organization: University of Missouri - Rolla
- Sender: cnews@umr.edu (UMR Usenet News Post)
- Message-ID: <1992Sep8.181730.21601@umr.edu>
- Lines: 49
-
- In article <1992Sep8.002147.790@wam.umd.edu> joel@wam.umd.edu (Joel M. Hoffman) writes:
- >In article <1992Sep7.215429.7866@umr.edu> dminer@mcs213e.cs.umr.edu (Dan Miner) writes:
- >>wait a little (~100K? for that one). Now, I got pl4 and ready for
- >>an all night compile (it takes over 5 hrs on my machine.. :( )
- > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
- >
- >Along the lines of my last questions, about the progress of Linux over
- >the Summer, I've been seeing a lot of this (5 hours to recompile
- >kernel). Is this because the kernel is bigger and GCC slower than
- >they were when we had 0.96, or are the auth}rs just running slow
- >machines? I seem to recall about 15 minutes for a recompilation.
- >
-
- I being one of those authors. :)
-
- I have a 386sx, 2 megs, Paradise SuperVGA, 80 meg HD, and running
- at 20 MHz. Now, this is NO way a great setup. In the 0.12 days,
- it took about 15 mins for me. I had 5 megs swap and gcc 1.40.
- Now, as the kernel got bigger, it took longer. No problem. But
- the kernel expanded in memory usage too over this time. With it
- new release I had less memory and more SWAPPING. It has finally
- downgraded to 5 hrs. I compiled 0.97.4 last night. It took 5 hrs
- and 20 mins. I do believe things will speed up. With it new kernel,
- most operations are just a tad faster. The only answer is for me
- to buy more memory and/or upgrade motherboards.
-
- I came up with a possible speed increase yet. 0.97.4 sports
- sharing of swap pages. Why not "throw" the text segments of a program
- out. When needed again you can "demand load" it. All that is vital is the
- data pages... This could at almost divide swap time by 2. Swapping is
- an everyday occurance for me. (Few things DON'T swap).
-
- Now don't get me wrong, I'm complaining about compile times. I can do little
- about it. I like Linux alot and will be with it for along time. It has
- made me respect PCs a bit more (now that you can do something :). Linux
- memory requirements are a little high for my liking but it works well.
- I have been "pushing" it to all my friends who have 386/486...
-
- Anyway, this is getting too long.... :)
- Dan Miner
- Hackers' Guide Coordinator
-
- P.S. I upgrade kernel only every 4 or 5 patches or new releases, if
- you haven't guessed. :)
- --
- Dan Miner
- email: dminer@cs.umr.edu
- Computer Science
- University of Missouri-Rolla "It all begins with a glimmer."
-