From: Digestifier To: Linux-Activists@senator-bedfellow.mit.edu Reply-To: Linux-Activists@senator-bedfellow.mit.edu Date: Sun, 22 Aug 93 16:13:08 EDT Subject: Linux-Activists Digest #156 Linux-Activists Digest #156, Volume #6 Sun, 22 Aug 93 16:13:08 EDT Contents: Re: Tractatus Linuxicus Newbius (Stephen Harris) Re: Linux CD-ROMs in general --> Yggdrasil in particular (Adam J. Richter) Remote task-starting (Harald Finnaas) Re: Compiling kernel (Joel Jeffery) Re: SCSI Performance (Yet Again) (John Henders) Re: Tractatus Linuxicus Newbius (John Henders) Re: The meaning of DOS (KZUPAN@LSTC2VM.stortek.com) errors when compiling xfig under linux 1.01 (sh7131220@ntuvax.ntu.ac.sg) Re: Has anyone heard of Infomagic CDROM? (Kenneth Benson) Re: Tractratus Linuxicus Philosphicus Schmus... Re: Unix for sale (DDLEEDS@vm.cmp.ilstu.edu) Re: which (Charles T Wilson -- Personal Account) re: errors compiling xfig Re: HELP WITH MY MOUSE IN X PLEASE (Bill Mitchell) Re: SCSI Performance (Matthias Urlichs) Re: SCSI Performance (Matthias Urlichs) Re: Ethernet Cards informations needed Re: SCSI Performance (Piercarlo Grandi) Is my system OK for Linux? (T.X. Yuan) Re: SCSI Performance (Yet Again) (Piercarlo Grandi) ---------------------------------------------------------------------------- From: harris@teaching.physics.ox.ac.uk (Stephen Harris) Subject: Re: Tractatus Linuxicus Newbius Date: 22 Aug 93 15:18:04 BST : In <255ljc$5jm@ionews.io.org> las@io.org (Laszlo Herczeg) writes: : > I am a philosopher myself, and I really think no one should be allowed : >to wield power over technology who cannot pass a course in literary : >criticism. Literary critisism? Hmm. WOnder how many Unix Admin people are strictly Unix admin adn deal with people that understand basic computing, and so can cope with hackerese ? There are two sorts of people: Those that can do Those can can teach (Note: regardless of the "fortune" quote, these groups are *NOT* exclusive. People *can* do and teach at the same time). Now, why should those people who "do" be forced to translate between current technical "jargon" and a form that others can understand. This would slow down developement thousandfould (I don't exagerate!). Where I work I have to do both developement and teaching roles, and know that the teaching side of things is soooooo slooooowww that if I had to deal in those terms almost constantly no work would be done. Jargon (ie those terms that would fail any course in literar criticsim) is essential in any "deep knowledge" situation just to make people who have "the knowledge" use that knowledge effectively. After all, how many people have been to a car garage and understood exaclty 100% of what the mechanic has told you? You may pretend or guess the meaning, but this doesn't mean that you *do* know what they mean. Hmm, not really linux based - more philosophy based :-) oh well.... Stephen Harris harris@teaching.physics.ox.ac.uk Opinions are just opinions, and the facts are the facts. But what are what? ------------------------------ From: adam@netcom.com (Adam J. Richter) Subject: Re: Linux CD-ROMs in general --> Yggdrasil in particular Date: Sun, 22 Aug 1993 14:41:07 GMT In article <1993Aug19.172528.15787@valis.ampr.ab.ca> user1@valis.ampr.ab.ca (Kevin Fluet) writes: >That Yggdrasil group seems to have disappeared, too (at least I haven't >heard anything about their final release CD). At the outset, let me congratulate you for being the first USENET poster to spell "Yggdrasil" correctly on the first try. Yggdrasil Computing, Inc. is alive and well and very busy. We have sold about 3,100 LGX betas (not including freebies), which I gather is about twice as many copies as Jana and SLS combined. We have more than that number of advance orders for the first production release, which will feature, among other things, a "fill in the blanks" graphical user interface for doing things like installing an ethernet card or a printer, plus support for almost every CDROM drive. You may not have heard much from Yggdrasil lately because we've been busy trying to cut the first production release and also doing other things to build company infrastructure in anticipation of rolling out support services with the first production release. For example, if all goes well, we will be moving into new offices in the heart of silicon valley with a shared high speed internet connection on September 1, at which time we will be able to set up our own FTP site. We will also soon have a 900 number for general Linux support, including support for users of other distributions and questions of a tutorial nature (e.g., "how do I use this program?"). Also, We are in a unique position to offer on-demand support services for hardware-related questions because we've already invested in so much different PC hardware for testing. I hope that I have given you a reasonable sense of the work that is going on at Yggdrasil right now. To some extent all of the Linux CDROM companies seem to be victims of their own success. It is a very challenging problem to transform a small business into a serious company in the space of, say, 800 business hours, while not ignoring the operation of the business during that time or spending money beyond your means. I think that I speak for all of the Linux CD companies when I say that the quiet that you're hearing does not mean that we've gone away. It's more like the calm before the storm. -- Adam J. Richter Yggdrasil Computing, Incorporated 409 Evelyn Ave., Apt. 312, Albany CA 94706 PO Box 8418, Berkeley CA 94707-8418 (510) 528-3209 (510) 526-7531, fax: (510) 528-8508 adam@netcom.com yggdrasil@netcom.com Another member of the League for Programming Freedom (lpf@uunet.uu.net). ------------------------------ From: harald-f@fidonet.bbs.no (Harald Finnaas) Subject: Remote task-starting Date: Sun, 22 Aug 93 15:02:29 GMT This isn't really Linux specific, but I'd guess it'll do. I would like to call a Linux/UNIX machine via modem, start a regular shell script (.sh) and be able to hang up without the script beeing killed before it's finished. Is this possible? And some Linux-related questions: Are there drivers for 3com 507 available? How do I go about setting up TCP/IP? I've got a 3com 503 that I can use if the 507 isn't supported, so that shouldn't be a problem. _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ _/ Harald Finnaas (Stinger) | Romeo November ...... +47-4-536698 _/ _/ | N-4026 Stavanger, Norway. _/ _/ harald-f@fidonet.bbs.no | FIDO: 2:211/5@fidonet.org _/ _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ ------------------------------ From: Joel.Jeffery@purplet.demon.co.uk (Joel Jeffery) Subject: Re: Compiling kernel Date: Sat, 21 Aug 1993 17:26:00 +0000 * In a message originally to All, Richard Kasperowski said: RK> It's true. I went so far as to download the _Kernel_Hacker's_Guide_, RK> hoping to find out what to do after compiling a new kernel, but RK> there's no information there either. RK> So what do you do after compiling a new kernel in order to RK> use it? Erm: cat zImage >/vmlinux or cat Image >/vmlinux followed by /etc/lilo/install assuming that you've set up /etc/lilo/config properly... ------------------------------ From: jhenders@jonh.wimsey.bc.ca (John Henders) Subject: Re: SCSI Performance (Yet Again) Date: Sun, 22 Aug 1993 15:03:38 GMT root@hip-hop.suvl.ca.us (Remco Treffkorn) writes: >Listen, the point here is simply that some people are wondering if the >SCSI performance is all the way up there *or* if it could be improved. >Natually they are comparing notes. The problem is only that nobody >can agree on the correct way to measure performance. Iozone is certainly >not perfect, but gives you a relative measure if compiled correctly. Exactly. In my comparisons, I even noted that the numbers I see for performance under DOS for my Adaptec 1452b are similar, though slightly higher. I'm curious, not whining. I'd be satisfied with some sort of explanation, no specificly any promise of improvement. SO far, it seems to me that the theoretical 5.7 meg/sec of the Adaptec is just that, theory, and only a few people get even 1 meg/sec. Again, on my ST, there was a program that bypassed the filesystem completely to test raw disk i/o, and somehow they even made it non-destructive so you didn't have to use a clean partition to do tests. It seems there is no equivalent under *nix. >Again, nobody (I hope!-) is whining. I am subjectively happy with the >disk performance, but after having seen some people with IDE presenting >numbers that beat the sh*t out of my setup with an Adaptec SCSI controller >I am certainly motivated to look into the matter a bit further. I'm happy enough with performance, as the filesystem speed of even the minix filesystem beats the crap out of the FAT based fiesystem on my ST. news unbatching is probably 10x faster on my linux box than on my ST. I won't even mention the difference in deletes on 100+ files. It is a bit mystifying to me that scsi, which I thought was supposed to be superior to ide, gets lower numbers compared to a similar ide drive though. I'd really like to see a comparion between 2 of the same model drive, one scsi, one ide, like the Quantum 105 or a Maxtor. -- John Henders GO/MU/E d* -p+ c+++ l++ t- m--- s/++ g+ w+++ -x+ ------------------------------ From: jhenders@jonh.wimsey.bc.ca (John Henders) Subject: Re: Tractatus Linuxicus Newbius Date: Sun, 22 Aug 1993 15:06:49 GMT las@io.org (Laszlo Herczeg) writes: > The Linux system administrator's guide (SAG) was a very good first >step in making Linux more accessible, but it is in .ps format, and how >do you expect someone coming to Linux from the PC world to be able >to read .ps if they don't even know what a compiler is ? I believe a text release of the SAG is just waiting on a working release of the TeX to ascii program Lametex. -- John Henders GO/MU/E d* -p+ c+++ l++ t- m--- s/++ g+ w+++ -x+ ------------------------------ From: KZUPAN@LSTC2VM.stortek.com Subject: Re: The meaning of DOS Date: Thu, 19 Aug 1993 14:32:55 GMT In article <1993Aug19.121336.28699@taylor.uucp> mark@taylor.uucp (Mark A. Davis) writes: > >simmons@borg.uucp (David Simmons) writes: > >>I have come to a revelation about DOS, after a long night >>of assmebling a Linux network... > >>DOS does have its uses. DOS serves as a low level operating >>system that can be used to perform maintenance on linux >>machines easily because of its small size. An excellent >>analogy would be the PROM monitor on Sun machines compared >>to the SunOS.. :) > >Linux is DOS. DOS is Disk Operating System. MS-"DOS" is a single user, >single tasking, single processing program loader; which is hardly an OS >at all. In fact, one could say that Linux=DOS and MS-"DOS"<>DOS! :) :) >-- > /--------------------------------------------------------------------------\ > | Mark A. Davis | Lake Taylor Hospital | Norfolk, VA (804)-461-5001x431 | > | Sys.Administrator| Computer Services | mark@taylor.wyvern.com .uucp | > \--------------------------------------------------------------------------/ Mark Go home, drink a beer, and go back to bed!!! ========================================================================= | L IIIII NN N U U X X Out to fix what Windog MS'd up!!! | | L I N N N U U X X | | L I N N N U U X Another Word From Kevin Zupan | | L IIIII N NN UUU X X Kzupan@lstc2vm.stortek.com | | LLLLL X X | ========================================================================= ------------------------------ From: sh7131220@ntuvax.ntu.ac.sg Subject: errors when compiling xfig under linux 1.01 Date: Sun, 22 Aug 1993 15:38:31 GMT Hello, I have a little problem with compiling programs under Linux version 1.01. The programs that I have tried includes Tgif, xfig,xloadimage and xv all of which I have gotten from sunsite.unc.edu. When I tried to make the Makefile included in the programs, tonnes of error messages just poured out. Some of the most frequently encountered ones include " parse error"and "undeclared but used ". I couldn't believe that so much errors could be generated because the programs are supposed to be already thoroughly debugged and tested. I suspect that certain files ( e.g *.h ) in my linux are corrupted. Anybody who has any idea as to what is going on,please email to me or send replies thru' usenet. email address : sh7131220@ntuvax.ntu.ac.sg THANKS !!! ------------------------------ Crossposted-To: alt.cd-rom Reply-To: kbenson@fireborn.win.net (Kenneth Benson) From: kbenson@fireborn.win.net (Kenneth Benson) Date: Sat, 21 Aug 1993 00:22:03 GMT Subject: Re: Has anyone heard of Infomagic CDROM? In article <1993Aug19.133432.8016@turtle.apana.org.au>, Fozzie Bear (fozzie@turtle.apana.org.au) writes: >A friend of mine asked me today about a CDROM called Infomagic. It >apparently has Linux on it, BSD and heaps of other Unix things. > >Any further info would be appreciated. > >fozzie > >-- >--------------------------------------------------------------------------- >fozzie@turtle.apana.org.au I'm looking for an ascii turtle picture. > If you've got one, email it to me! > You can contact them at info@infomagic.com... they have a number of different CD's. One has 386BSD, NetBSD and Linux 0.99.10 on it. Another is the Yggdrasil LINUX/GNU/X CD. The one I like reads as follows: >SOURCE CODE cd-rom >A wealth of source code including: the Berkeley Net/2 distribution (4.3BSD) >MACH, GNU software, InterViews, X11R5 (base MIT dist. including fixes 1-25 >and contributed software), Andrew windowing and XFree86 (port for 3/486 >machines), ports of GNU EMACS for DOS and Windows(tm) (DEMACS and WINEMACS) >djgpp (GNU C/C++ for DOS), Modula-3, Windows NT(tm) software and more. Of course, the MS-DOG and WINDOZE (NT) stuff I don't need, but the X11R5, MACH and Berkeley sources I'd love to have! Good Luck! Ken The FireBorn P.S. The flyer says the CD is $40. Pretty good, IMHO. Also, note that the above is from a flyer, I don't work for InfoMagic and it's for information sake only. It's NOT a commercial! ------------------------------ Date: Sun, 22 Aug 1993 20:25:27 CDT From: Subject: Re: Tractratus Linuxicus Philosphicus Schmus... Laszlo Herczeg, las@io.org writes: Las> I am a philosopher myself, and I think no one should be allowed Las> to wield power over technology who cannot pass a course in literary Las> critisism. As a philosopher yourself, you surely have already passed your course in literary critisism. Thus you are allowed to wield power over technology. A shoe-salesman is never allowed to wield power over technology. Bill Gates was a shoe-salesman. Then he attended a course in literary critisism, and founded Microsoft. Without reading Kant, oh, and of course Wittgenstein, a dumb CS student should not be allowed to wield power of technology. /herp ------------------------------ Crossposted-To: comp.unix.misc,comp.unix.sys5.r3,comp.unix.sys5.r4,comp.unix.sys5.misc,comp.unix.pc-clone.16bit,comp.unix.pc-clone.32bit,comp.os.linux.misc From: DDLEEDS@vm.cmp.ilstu.edu Subject: Re: Unix for sale Date: Sun, 22 Aug 93 13:26:20 EDT In article <1993Aug22.020132.6155@cseg03.engr.uark.edu> dws@cseg03.engr.uark.edu (David W. Summers) writes: > > >Xenix for sale. Nothing wrong with it. I used it for years! I upgraded >to Linux. > > >I have two Xenix 286 for sale and two Xenix 386 for sale. > >Please e-mail me for details. > >I am offering these at rock-bottom prices! I can't find anyone to buy >them! > > > - David Summers > >-- > "Never under-estimate the bandwidth of a station-wagon >David Summers full of tapes, hurtling down the highway." >dws@engr.uark.edu - Tanenbaum, "Computer Networks" ------------------------------ From: ctwilson@rock.concert.net (Charles T Wilson -- Personal Account) Subject: Re: which Date: 22 Aug 1993 18:30:40 GMT I've used something posted to comp.sources called att-which since 0.12; it's compiled 'out of the box' since then. It works well and I'll be glad to post it to sunsite and tsx-11 if people want it. The one thing it does that shell solutions don't is print its search paths when it fails to find the executable you're checking. It's also small and fast. It was at f.ms.uky.edu about 2 years ago and probably still is. -- /-----------------------------------------------------------------------\ | Tom Wilson | "I can't complain, but sometimes | | ctwilson@rock.concert.net | I still do." | | | -Joe Walsh | ------------------------------ Date: Sun, 22 Aug 1993 20:45:07 CDT From: Subject: re: errors compiling xfig umh, I had not a single error when compiling xfig. I can't remember if it did even give warnings. After compile was finished, I started it and ... it worked! does someone want the binary? :-) /herp ------------------------------ From: mitchell@mdd.comm.mot.com (Bill Mitchell) Subject: Re: HELP WITH MY MOUSE IN X PLEASE Reply-To: mitchell@mdd.comm.mot.com (Bill Mitchell) Date: Sun, 22 Aug 1993 16:49:53 GMT in comp.os.linux, u339580@sparc17.ncu.edu.tw (* 3cMonster *) said: >Hi all UNIX admirers, > >Here I just come across the situation below: > > 1. I set everything OK with my X-window (XFree 1.3) and it works fine. > 2. As I am trying to move my mouse, I just find it move randomly on the > screen. > 3. I tried to find through the FAQs times and times..nothing about that > 4. my mouse is Microsoft compatible. > 5. I set it to Microsoft "/dev/cua0" in my Xconfig > 6. That's all... I've seen this, or something similar, on my system. What I did was to rm /dev/mouse and /dev/modem, and then ln -s /dev/ttyS0 /dev/mouse and ln -s /dev/ttyS1 /dev/modem. I haven't looked into why this worked, but it did. For now, that's sufficient. You might try something similar, with appropriate ttyS numbers, on your system. -- mitchell@mdd.comm.mot.com (Bill Mitchell) ------------------------------ From: urlichs@smurfnbg.smurf.sub.org (Matthias Urlichs) Subject: Re: SCSI Performance Date: 21 Aug 1993 21:08:15 +0200 In comp.os.linux, article <1993Aug20.211540.126@dde.dk>, kim@dde.dk (Kim Andersen) writes: > > >4) umount the filesystem > > Why should the filesystem be unmounted ? Is there some sync'ing needed which > doesn't get done with fsync ? > Quite the opposite; the unmount forces the buffer to be flushed so that no blocks remain cached. Otherwise your subsequent read will pull these bits out of the buffer cache, thus skewing the results. The fsync just forces your file to be written out, but the data may remain in the cache. -- Matthias Urlichs -- urlichs@smurf.sub.org -- Phone: NONE; use email or lose. Schleiermacherstrasse 12 -- 90491 Nuernberg -- Germany || Linux+Mac Consulting ------------------------------ From: urlichs@smurfnbg.smurf.sub.org (Matthias Urlichs) Subject: Re: SCSI Performance Date: 20 Aug 1993 11:46:33 +0200 In comp.os.linux, article , muts@compi.hobby.nl (Peter Mutsaers) writes: > > I would expect almost 1MB/s with SCSI. You can get 1 MB/s, just run iozone on two or three disks concurrently and add the results. This, and the fact that the CPU has less work to do for the same amount of I/O, is the advantage of SCSI (besides being able to connect more and different devices, of course, and that SCSI disks are capable of remapping bad blocks, which is a Good Thing), not raw speed to a single disk. -- Matthias Urlichs -- urlichs@smurf.sub.org -- Phone: NONE; use email or lose. Schleiermacherstrasse 12 -- 90491 Nuernberg -- Germany || Linux+Mac Consulting ------------------------------ From: cbarrera@austin.ibm.com () Subject: Re: Ethernet Cards informations needed Date: Sun, 22 Aug 1993 19:01:04 GMT Donald J. Becker (becker@super.org) wrote: : [[ My newsfeed is falling further behind each day, so this may not be timely.]] : The 3c509 driver is still in the alpha stage. : The 3c503 driver has been working for about nine months. It's not the : fastest ethercard around, but it's from 3Com. : The 3c501 driver once worked (for about three minutes), but was never So no one is working on 3c507? :( I would do it myself except I won't have my linux box installed for another month or so... -chris barrera I don't speak for this company, I am just a co-op :) ------------------------------ From: pcg@aber.ac.uk (Piercarlo Grandi) Subject: Re: SCSI Performance Reply-To: pcg@aber.ac.uk (Piercarlo Grandi) Date: Sun, 22 Aug 1993 19:51:26 GMT >>> On 20 Aug 93 21:15:40 GMT, kim@dde.dk (Kim Andersen) said: Kim> pcg@aber.ac.uk (Piercarlo Grandi) writes: pcg> The *only* accurate way to test effective read and write rates is: pcg> 1) mkfs an empty disk partitions pcg> 2) mount the filesystem pcg> 3) write the file to the filesystem a longer than total memory; fsync pcg> 4) umount the filesystem Kim> Why should the filesystem be unmounted ? Is there some sync'ing Kim> needed which doesn't get done with fsync ? Fsync writes back the buffers, does not free them. Depending on how large the file is and what is the retention policy of the cache (FIFO, LRU, random) this means that perhaps the subsequent read will not be entirely from disk. Unfortunately it is rare (I know of none) to find a Unix FS that allows advising for expected access patterns, even if it would most probably give *huge* performance benefits, and would be rather easy to make the stdio library to use it transparently and effectively for 90% of the programs. pcg> This said, the ext2fs filesystem does not give the optimal pcg> bandwidth (even if I think it is better than the BSD UFS, as it is pcg> less complicated); Kim> Is it just the fact that the BSD UFS looks more complicated to you Kim> that you don't like it? Yes, and it is a very grave problem. Complication, especially in system level code, engenders two problems: it makes things harder to write/understand/maintain, and it generates complicated performance models that are tricky to tune. The UFS embodies a number of clever ideas that supposedly improve performance, but their value is very dubious, and it's in any case hard to gauge their effectiveness (not large). FS designs based on simpler strategies reportedly get much better average performance. Kim> Surely there are room for improvement in the BSD UFS (as used in Kim> 386BSD/NetBSD/FreeBSD), such as the clustering changes Larry McVoy Kim> made at Sun, I had once a nice discussion with Mc Voy about those. Perhaps I was being a bit too subtle, but I think he never got the gist of my argument, that he had conclusively proven that the central feature of the FFS, large blocks and fragments thereof, is pointless (besides having non trivial costs). He actually kept harking the merits of larger block sizes, which fact made me really wonder... Kim> or the LFS work done by Margot Selzer. Hasn't this idea been incorporated into the version of the UFS that is being sold for SVR4 by some company? Kim> One thing I like about BSD UFS is the robustness. I am usually not Kim> scared from turning of the power on a live and kicking fs. Because it keeps things in synch a lot, at some price in efficiency, depending on version. For example under some UFS versions copying/loading fs subtree is rather painful as it does every single inode update synchronously. Living a bit more dangerously seems a better idea, and logging an even better one, as fsck'ing a couple or more of 1GB disks can take a bit of time... BTW I think there are plenty of opportunities for Msc level projects on improving the Linux filesystems, which would look good and also be useful to a lot of people... Some of these could be extended into PhD level projects. ------------------------------ From: txy3340@ultb.isc.rit.edu (T.X. Yuan) Subject: Is my system OK for Linux? Date: Sun, 22 Aug 1993 19:21:31 GMT Hi Linuxies, I can't wait to run Linux and XWindow.. Can anyone tell me if my system is fit for it? I have a 486DX2/50 VLB/EISA/ISA with - Diamond Viper VLB video card (never seen a Weitek 9000 driver anywhere) - Conner CP30540 SCSI2 HDD - DTC31/3292 EISA SCSI2/FastSCSI2/SCSI controller - 20 meg memory - ZyXEL U-1496E 16.8k modem / V32bis / V42bis / 14.4k fax / voice-mail - SoundBlaster Pro 8 bit sound card - Z-Nix three buttons and MicroSoft compatible mouse - HP LaserJet IIIP with PS Level2 Thanks. ------------------------------ From: pcg@aber.ac.uk (Piercarlo Grandi) Subject: Re: SCSI Performance (Yet Again) Reply-To: pcg@aber.ac.uk (Piercarlo Grandi) Date: Sun, 22 Aug 1993 20:02:01 GMT >>> On Sat, 21 Aug 1993 16:47:58 GMT, root@hip-hop.suvl.ca.us (Remco >>> Treffkorn) said: Remco> Listen, the point here is simply that some people are wondering if the Remco> SCSI performance is all the way up there *or* if it could be improved. Oh yes, several. Remco> Natually they are comparing notes. The problem is only that nobody Remco> can agree on the correct way to measure performance. Iozone is certainly Remco> not perfect, but gives you a relative measure if compiled correctly. Yes, it gives you a relative measure, but perhaps it is not of disk performance. Actually one would want to have an idea of how good each of these components is *in isolation*; 1) the generic SCSI code 2) the HA specific SCSI code 3) the disk SCSI code 4) the filesystem code 5) the cache handling code. Also, it is rather useful to observe not only how long (elapsed) the operations take, but also the system CPU time taken. My *impression* is that the SCSI drivers are rather OK (e.g. by comparison with SVR4, which has some hideously inefficient drivers), in that they have relatively little delay and consume relatively little CPU, and that the biggest problems are, in order of decreasing importance: a) the filesystems don't do read/write clustering. b) advising for expected access patterns should be possible/automatic. c) the cache code should be eliminated (mmap'ing everywhere). d) the allocation clustering algorithm could be improved. Point c) is being done by Linus Torvalds; but it can have a severely negative performance impact, as Sun discovered a couple years ago, unless it is done after b). Doign a) and d) would not be too difficult, and would eliminate the need for blocks and fragments (a poor idea anyhow). ------------------------------ ** FOR YOUR REFERENCE ** The service address, to which questions about the list itself and requests to be added to or deleted from it should be directed, is: Internet: Linux-Activists-Request@NEWS-DIGESTS.MIT.EDU You can send mail to the entire list (and comp.os.linux) via: Internet: Linux-Activists@NEWS-DIGESTS.MIT.EDU Linux may be obtained via one of these FTP sites: nic.funet.fi pub/OS/Linux tsx-11.mit.edu pub/linux tupac-amaru.informatik.rwth-aachen.de pub/msdos/replace The current version of Linux is 0.99pl9 released on April 23, 1993 End of Linux-Activists Digest ******************************