From: Digestifier To: Linux-Misc@senator-bedfellow.mit.edu Reply-To: Linux-Misc@senator-bedfellow.mit.edu Date: Sat, 24 Sep 94 00:13:37 EDT Subject: Linux-Misc Digest #811 Linux-Misc Digest #811, Volume #2 Sat, 24 Sep 94 00:13:37 EDT Contents: Re: Linux on 40,000 FREE(ish) CD's (Duncan THOMSON) Re: More Memory = Slow Linux?? (Pete Chown) Function Keys (Chris Burnette) LILO (Chris Burnette) Re: More Memory = Slow Linux?? (Hugh Johnson x6549) Re: Ghostscript, resume printing from page n (Matt Weber) howto use telnet/ftp under term? (Lars L. Madsen) Re: Don't use Linux or it's to academic! (David E. Fox) Re: Dynamic PPP Dial-in (Al Longyear) Free Linux CD's (Adam J. Richter) Re: ** autoconf.h? ** (Michael_Nelson) Re: DOOM linux with TERM (Alex R. Moon) Re: Don't use Linux or it's to academic! (Sean Gilley) Re: P5-90 MHz beats SGI R4000-100MHz. (Ove Ewerlid) Re: P5-90 MHz beats SGI R4000-100MHz. (Robert Stockmann) Sound problems and others (Thomas Ward Hanselman) Re: Pentium-optimized compiler is slower then 486 gcc-2.5.8 on P5 !!! (Michael Griffith) ---------------------------------------------------------------------------- From: duncan@lightning.eee.strath.ac.uk (Duncan THOMSON) Subject: Re: Linux on 40,000 FREE(ish) CD's Date: 23 Sep 1994 16:03:48 GMT In article rsmurf@ritz.mordor.com (Rasta Smurf) writes: It's the UK edition of PC Plus, the October edition (out now) -Duncan -- Duncan C Thomson, Instrumentation & Measurement | .__ _.._ .__._ _ , .___ _ University of Strathclyde, Glasgow G1 1XW, UK | |_ (_ |_)|_ |_)/_\|\| | / \ +44 41 552 4400 X2205 Fax: +44 41 552 2487 | |__._)| |__| \| || | | \_/ | -==- lingvo internacia -==- http://lightning.eee.strath.ac.uk/~duncan/ | ------------------------------ From: pc@dale.dircon.co.uk (Pete Chown) Subject: Re: More Memory = Slow Linux?? Date: Fri, 23 Sep 1994 16:17:54 GMT In article <35qu9m$18et@yuma.ACNS.ColoState.EDU> pyeatt@cervesa.cs.colostate.edu (Larry Pyeatt) writes: >In article <35pd26$2ft@fs7.ece.cmu.edu>, garcia@ece.cmu.edu (Brad Matthew Garcia) writes: >> I keep seeing posts made by people who have added memory >> to their computers and subsequently experienced a drop in >> performance under Linux. >Most PC motherboards have really cheesy cache setups. The cache works >fine as long as you don't put too much RAM in. The other point to remember is that some BIOSes think (for reasons best known to themselves) that you might not want to cache all your RAM. If you have one of these, you have to make sure that you have told the BIOS to cache all your RAM and not just the first part. (AMI is like this, as I found out the hard way a while ago.) ------------------------------ From: cburnett@nvl.army.mil (Chris Burnette) Subject: Function Keys Reply-To: cburnett@nvl.army.mil Date: Mon, 19 Sep 1994 18:31:45 GMT I am wanting to make my PC keyboard more like a workstation keyboard, ie with the functions keys down the side which do the Find, Copy, Paste, etc. I cannot seem to get this to work. I have tried xmodmap but can't seem to get that to do it. Anyone have any suggestions? E-mail them to cburnett@nvl.army.mil Thanks, Chris Burnette E-OIR Measurements, Inc. ------------------------------ From: cburnett@nvl.army.mil (Chris Burnette) Subject: LILO Reply-To: cburnett@nvl.army.mil Date: Mon, 19 Sep 1994 18:37:55 GMT Anyone know of a way to make LILO boot up DOS on default instead of Linux? E-Mail: cburnett@nvl.army.mil Chris Burnette E-OIR Measuements, Inc. ------------------------------ From: hugh@asdi.saic.com (Hugh Johnson x6549) Subject: Re: More Memory = Slow Linux?? Date: Thu, 22 Sep 1994 18:17:25 GMT Larry Pyeatt (pyeatt@cervesa.cs.colostate.edu) wrote: : In article <35pd26$2ft@fs7.ece.cmu.edu>, garcia@ece.cmu.edu (Brad Matthew Garcia) writes: : |> : |> I keep seeing posts made by people who have added memory : |> to their computers and subsequently experienced a drop in : |> performance under Linux. - - - - - - - < snip > - - - - - - - - : OK. here is the scoop: : Most PC motherboards have really cheesy cache setups. The cache works : fine as long as you don't put too much RAM in. When you get too much : RAM, the cache does not work on the upper part, so on a 16 Meg system, : you may only have caching on the lower 8 Meg. If your cache RAM is 15ns : and your DRAM is 70ns, and your cache hit rate is 90% for the lower 8 Meg - - - - - - - - < snip > - - - - - - - - : The fix: upgrade your cache to 256K. Don't forget to upgrade the : cache tag RAM as well, or you will still get no benefit from the : additional cache. Some motherboards are so cheesy that it is : impossible to cache the whole address space, even with the maximum : cache RAM. Well put, although not _entirely_ accurate, but in essence, correct. Some bios (AMI, as an example) allow you to determine how much RAM to cache. Some don't (Bummer). Also, alot of hardware designs have very lousey design for DMA above 16 Megs, (read that as: if it's a 386 or if its a mature 486, don't expect DMA above 16 Megs.) Most high speed I/O may want to use DMA to memory rather than the old 'Test-n-Transfer' mode. (This applies to the BIOS drivers for floppies {Yes, floppies} and hard drives.) The Pentium grade machines and some of the later '486 designs now allow DMA above 16 Megs. Cache for these machines are larger in size and allow cache'ng for more than 16 Meg ram. (Do I get a receipt for my $0.02?) -- hugh@asdi.saic.com or | Opinions? I doan' have no steenkin' hugh@seada.com or | opinions. They doan' let me have 76317.2234@compuserve.com | none. ------------------------------ From: mattw@meaddata.com (Matt Weber) Subject: Re: Ghostscript, resume printing from page n Date: 21 Sep 1994 17:30:30 GMT Bontridder Dirk (bontridd@btmpds) wrote: : Hello, : A few days ago, I was printing with ghostscript a manual (postscript format) of 200 pages. After 60 pages my printer was resetted. How can I resume my printing without printing the first 60 pages of my document? : Thanks. : -DB- Try pulling up the file in ghostview select pages 61-200 the use the "print marked pages" option under file. -- ######################################################################### Matt Weber ------------------------------ From: madsen@polymer.ucsb.edu (Lars L. Madsen) Crossposted-To: comp.os.linux.help Subject: howto use telnet/ftp under term? Date: 23 Sep 1994 16:25:50 GMT Dear netters Could some kind soul please explain 'term' to me? I have managed to build and run term ver. 2.1.1, and I am able to telnet from my local machine to a remote machine on campus. I do the following: tredir 2023 remotehost.ucs.edu:23 telnet localhost 2023 this will give me a telnet session on the remote host! good, but what if I want to telnet to another host? Same questions regarding ftp. What is the difference between 'tredir 2023 remotehost.ucsb.edu:23' and 'tredir 2023 23' ? How do I use a name server ? can I use 'tredir 2042 nameserver:42' ? I would like to run mosaic from home, I tried to run mosaic on the remote computer using txconn, and that worked, but running an x-server over a 14.400 modem is not exactly flying. I know that this is a little premature, since I have not ftp'ed mosaic yet (I can't ftp to ncsa... yet, you know) what do I need to get mosaic to work with term? Which docs are a good starting point ? I think my problems stem from the fact that I don't fully understand 'term'. I have tried to read the README, INSTALL, term.HOWTO, and the tcp/ip chapter in "Essential System Administration". If some of you could try explain term or answer my concrete problems I will be very great-full. Any pointer to further docs on term are also appreciated. sincerely Lars L. Madsen -- +----------------------------------------------+----------------------------+ | Lars L. Madsen | (805) 893-4325 | | | (805) 893-4349 (laboratory)| | University of California | (805) 893-4731 (FAX) | | Department of Chemical & Nuclear Engineering | | | Santa Barbara, CA 93106-5080 | madsen@junction.ucsb.edu | +----------------------------------------------+----------------------------+ | Home: | (805) 964-1159 | | 4320 Modoc Road #M | | | Santa Barbara, CA 93110 | | +----------------------------------------------+----------------------------+ ------------------------------ From: root@belvedere.sbay.org (David E. Fox) Subject: Re: Don't use Linux or it's to academic! Date: Fri, 23 Sep 1994 04:41:35 GMT Svein Erik Brostigen (serik@oslonett.no) wrote: : 1. There is no support for the more sofisticated disk handlings like : RAID 5 or STRIPING. In DOS or OS/2? You've got to be kidding that such support will ever happen. At least in the case of Linux, it's an open-developed system, and sooner or later this'll get supported, if it's something worth supporting. : 2. There are no programs available that does Word-processing, spreadsheets, : databases, presentation graphics and so on, as we have under MS-DOS : and Windows. I can think of several. Andrew, for instance. They may not look as snazzy as commercial software for DOS, but they do the job. In some cases, the free tools are better. : 3. There is currently no easy way to interconnect to LAN Server or Netware : natively, i.e. from Linux. : 4. There is no support for MCA-based machines. Well, there probably won't be. MCA is a joke. Newer designs exist that give more throughput (local bus, for example) than MCA ever did. Besides, the support for MCA, even in the DOS or hardware arena, is rather minimal. MCA-based cards are a rarity. Apart from the extant MCA-based computers, there's little reason to even consider it. : 5. There is no support for Token-Ring (some ALPHA code is floating around : on the Net, but....) I'll let someone answer the net stuff. : 6. Nationalized version of the few programs existing are not available. : 7. Nationalized versions of Linux is not available. Not yet. : 8. Support for modern graphics accelerators like Mach64, Stealth64 etc, : is not scheduled until maybe next year. Again, not yet. : 9. Printed manuals and easy 'Get started' manuals does not exist. Well, in all of this you have to realize that Linux is an evolving thing. Documentation does exist (like the Linux Bible, and the Linux getting started guide by Matt Welsh) but more is helpful. Instead of flaming linux, consider improving it. : 10. Support for Mulit-media is shaky. Andrew. : 11. Support for ISDN is not generally available. : 12. There is no disk-compression. There is now (double). I wouldn't ever trust data to the DOS disk compression stuff. It is NOT reliable. : When I'm recommending some system to my customers it should be something : that has a proven record of durability, Netware and Lan Server both have : that when it comes to NOS'es. MS-DOS and OS/2 have it when it comes to : PC OS'es. Netware is a JOKE compared to Unix networking. : My customers are like all other customers, they will use wordprocessors, : spreadsheets and so on. They are now used to be able to see the same thing Fine. If you want a data appliance, get a Mac. If you want a computer, get something else. : --- : Svein Erik Brostigen, Tech. manager : Rønning Netverk Systemer AS Pho. +47 22 37 04 00 : P.O. Box 6730, Rodeløkka Fax. +47 22 37 03 70 : N-0503 Oslo Cel. +47 92 03 00 74 : Norway -- David Fox root@belvedere.sbay.org 5479 Castle Manor Drive San Jose, CA 95129 Thanks for letting me change 408/253-7992 magnetic patterns on your hard disk. ------------------------------ From: longyear@netcom.com (Al Longyear) Subject: Re: Dynamic PPP Dial-in Date: Fri, 23 Sep 1994 14:01:03 GMT raf@datatamers.com (Richard Farrar) writes: >I am looking for a FAQ or HOWTO on setting up some modems for dial in >using PPP with Dynamic IP number for terminal serving. >The modems work fine for VT100 but I would like to be able to use PPP. This is fairly easy. There is no need for a HOWTO. Since you are asking how to setup a modem . . . . Make sure that the modems present the proper status of DCD, CTS, and DSR. Ensure that the modems honor DTR and RTS. Don't use a cheap cable which does not pass these signals. Now, if you wanted to know how to setup PPP for use . . . Read the files README and README.linux in the source package. If you have questions about specific sections in the documentation, I'll be happy to answer them. Please quote the section that you don't understand so that I know what must be changed. The documentation in the PPP is not designed to be an all-encompassing PPP document. Nor is it meant to be the general discussion on how to set up your network configuration. There are other documents for those. The README file is generated by the package author, Paul Mackerras. The README.linux file is brought to you by the linux PPP working group. You will find additional information in the Networking Admin Guide and the NET-2-HOWTO. Please look there as well. I am sorry to simply say RTFM, but, in the case of "gee, I just was using vt100 terminal connections and I would like to use PPP. How do I do it?" is too broad of a question to answer specifically. So, please Read The Fine Manuals. A great deal of time was put into writing them. -- Al Longyear longyear@netcom.com ------------------------------ From: adam@yggdrasil.com (Adam J. Richter) Subject: Free Linux CD's Date: 21 Sep 1994 17:33:11 GMT We're cleaning up the office and have gathered about 100 pounds of old Linux releases that we want to get rid of. We have our Fall 1993 release, some copies of our beta release before that and also discs from other Linux vendors. They are free to whoever wants to drop by and pick them up. (We will not ship them.) Take as many as you want. -- Adam J. Richter Yggdrasil Computing, Incorporated (408) 261-6630 "Free Software For The Rest of Us." ------------------------------ From: nelson@seahunt.imat.com (Michael_Nelson) Subject: Re: ** autoconf.h? ** Date: 23 Sep 1994 15:55:38 GMT Reply-To: nelson@seahunt.imat.com Rob Janssen (rob@pe1chl.ampr.org) wrote: -> OTOH, it is not a good idea to include /usr/src/linux/include/config.h -> in an application program... -> Apparently the application depends on configuration details of the kernel, -> but those can change without the application knowing about it! I agree. There must be a better way of determining the current system configuration without counting on a file that may or may not exist at a given moment. -- Michael Nelson nelson@seahunt.imat.com San Francisco, CA FAX: 1-415-621-2608 ------------------------------ From: moon@symphony.cc.purdue.edu (Alex R. Moon) Crossposted-To: alt.games.doom,comp.os.linux.help Subject: Re: DOOM linux with TERM Date: 21 Sep 1994 17:38:45 GMT In article wrote: : damianf@wpi.edu (Damian Frank) writes: : : >: man tredir : : >: RTFM : : >Yes, I'd forgotten that. It MIGHT work; : : Won't work; Was tested in the past. : Something to do with the way Doom uses sockets. : : Casey : -- : Casey Ryder Vrije Universiteit Amsterdam ++312503-16844 CET : http://www.cs.vu.nl/~cvrij The MultiMedia Experience Have you tried tudpredir? I hear somewhere that linuxxdoom uses udp ports instead of tcp ports. --Alex moon@symphony.cc.purdue.edu ------------------------------ From: slg@slgsun.cb.att.com (Sean Gilley) Subject: Re: Don't use Linux or it's to academic! Reply-To: sean.l.gilley@att.com Date: Fri, 23 Sep 1994 20:43:40 GMT In article , Ian McCloghrie wrote: >It goes to buy you a little nametag on the front which says "Compaq". >Oh, and, at least in older Compaqs, a case put together with >nonstandard screws, so that you can't open it without a special tool, >so that you're more likely to use their service department than >a random PC hacker. > >On a more serious note, Compaqs and other name-brand quotes are, >perhaps, a little more reliable than hole-in-the-wall made-in-Taiwan >stuff. But the extra thousands of dollars you spend on them don't >get you any more I/O, not unless you specifically buy a system with >a faster/wider bus. This doesn't belong here, really, but hey, neither did the above article. The Compaq name tag buys you one thing that *could* be worth the money. Customer service. I bought a used Compaq 386DX20. Old machine when I bought it, extremely expensive when it was new, I'm sure. After I'd had it a year or two, I wanted to install a 3.5 inch drive. Bought one, and the durn thing would work. Took it back, got another, and it wouldn't work. Called Compaq customer service. The guy on the other side of the phone was *extremely* helpful, and further, he *overnighted*, at no cost to me, Compaq installation disks, a manual, and some other PC programs by Compaq. That little trick bought them a friend for life. I piece together my own systems now, but I'll recommend Compaq to people who want a good system and can afford a little extra for it. Sean. --- Sean L. Gilley The Information Super Highway is sean.l.gilley@att.com really just a rough gravel road with 614 860 9053 (h), 614 860 5743 (w) wonderful roadsigns. ------------------------------ From: ewerlid@frej.teknikum.uu.se (Ove Ewerlid) Subject: Re: P5-90 MHz beats SGI R4000-100MHz. Date: 23 Sep 1994 22:34:02 GMT In article ewerlid@frej.teknikum.uu.se (Ove Ewerlid) writes: > What about ECC checked memory ? What about a >200MB/sec bus ? Yup! Got one in front of me that does more than 200Mb/sec (64 bits wide) >> What about ECC checked memory ? What about a >200MB/sec bus ? >Yup! Got one in front of me that does more than 200Mb/sec (64 bits wide) Someone mailed me and said that the PCI bus on my pentium system is 32 bits and capable of 132 Mb/sec. This may well be correct and the PCI bus is probably what we are talking about. However, I'm primarily concerned about the bandwidth of the onboard DRAM. This IS 64 bits wide and is interleaved in two-bank blocks. Does any one know if there is a separate bus for the DRAM? I did some benchmarks some time ago that seemed to indicate more then 200 Mb/s but I could have f**ked someting up ... (like not using enough working set size to trick the cache ...) ------------------------------ From: stock@dutsh7.tudelft.nl (Robert Stockmann) Subject: Re: P5-90 MHz beats SGI R4000-100MHz. Date: Fri, 23 Sep 1994 22:04:15 GMT In <35v63m$3u4o@yuma.ACNS.ColoState.EDU>, Larry Pyeatt (pyeatt@cervesa.cs.colostate.edu) wrote: : In article , lera@zeus.chem.wvu.edu (Valery Petrov) writes: : |> Some benchmarks comparison: : |> : |> DELL XPS-90 SGI with R4000 cpu (100MHz): : |> Integer: 19.2 sec. 23.3 sec. : |> Floating point: 200 sec. 199 sec. : |> : |> I used gcc-2.5.8 with Linux-1.1.51 on DELL's Pentium and C 3.18 with : |> Irix 5.2 on Silicon Graphics machine. Programs were written in plain : |> C using double precision for floating point. Considering : |> the price difference (similarly equipped SGI is ~3 times more expensive) : |> I wonder who whould like to buy those Indigos nowdays. : ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ : No one. The SGI you are describing is outdated and not sold anymore. : Funny how a top of the line PC is being compared to something that : is too slow to be a workstation anymore. : 1. the MIPS R4000 is hardly the fastest processor made by MIPS. The : Indego line is several years old whereas Pentium 90 is new. If you : want to compare apples to apples, use a new Indy with a MIPS R4600 : processor at 166 Mhz. It blows the doors off Pentium 90. Would you : like to send your code to me so I can run it on a new SGI machine? : 2. SGI Indy does not cost ~3 times more than a top of the line DELL. : Let's look at the numbers, shall we: : Dell XPS 90, 16Meg ram, 17" monitor, #9 graphics card, : CD Rom Drive, 1G IDE disk, ethernet card, 3 year warranty. : Price: ~$4400 : SGI Indy, 16 Meg Ram, 17" monitor, accelerated graphics : 1G FAST SCSI Drive, ethernet, CCD camera, 3 year warranty. : Price: ~$6500 : You should note the the Indy has a MIPS R4600 processor, which is : much faster than Pentium. Also, the SCSI Drive is faster and more : expandable than the Dell IDE. Overall, the Indy will have much : higher throughput and lower price/performance. OK we bought an SCI Indy with the R4000 100MHz and 16 Mb RAM... We bought it in oktober '93 so its already "old" however *IT SUCKS* It has for instance no external cache.... here is the hinv output: 1 100 MHZ IP22 Processor FPU: MIPS R4010 Floating Point Chip Revision: 0.0 CPU: MIPS R4000 Processor Chip Revision: 3.0 On-board serial ports: 2 On-board bi-directional parallel port Data cache size: 8 Kbytes Instruction cache size: 8 Kbytes Main memory size: 16 Mbytes Integral Ethernet: ec0, version 1 Disk drive: unit 1 on SCSI controller 0 Integral SCSI controller 0: Version WD33C93B, revision D Iris Audio Processor: version A2 revision 4.1.0 Graphics board: Indy 8-bit Vino video: unit 0, revision 0, Indycam connected logging in on console takes over 1 minute...We use insightII from BIOSYM, actually we bought the software (insightII and ...) from BIOSYM, with it came the machine for a price of Dfl 20.000,= (US$ 12000,=). the system has no c-compiler and I can go on like this for a minute. Running insightII is the biggest dissaster happening to you... Ok the demo which was installed was alright, but this machine is in practice a failure.... Robert Stockmann stock@dutsh7.tudelft.nl He your email adress is not valid! Ever tried to set up some intelligent networking (e.g. sub netting) from within the sysadmin X program on a Indy? it doesn't work. So I had to use a normal xterm and vi and some tough hacking in those over 500 lines long corrupted scripts (My Linux experience helped me out) to get it going...IRIX is certainly not my favorit UNIX... ------------------------------ Crossposted-To: comp.os.linux.help From: Thomas Ward Hanselman Subject: Sound problems and others Date: Fri, 23 Sep 1994 02:43:23 GMT I have two problems/questions: First, the easy (or short) one. When I run top right after a cold reboot, it shows most of my main memory is being used (less than 600K) free. I just recompiled my kernel, I now have version 1.1.50. Also, I don't know if this is related, but I was looking in my /var/adm/messages file, and the kernel had reported several times that bdflush wasn't running. What is bdflush? Any ideas as to what's causing this? (my system info is at the end of this letter). Second, I have a major sound problem. I've got an SB16, and I'm using version 2.9 patch2 of the sound driver. Anyway, whenever I try to play anything that has 16-bit sound, my computer either locks up, reboots, or the kernel panics. This happens for things like DOOM and when I try to play MODs at 16-bits with s3mod. Here is the debugging output from the latest panic (this is after running X, compiling a few programs, and playing a few MODs at 8-bits): Ooops: 0000 EIP: 0010:00110dfb EFLAGS: 00010006 eax: 00000007 ebx: 0018d3bc ecx: 0018d3bc edx: fffffc18 esi: 00001000 edi: 001c76f8 ebp: 001ac8fc esp: 0018d338 ds:0018 es:0018 fs:002b ss:0018 Process swapper (pid:0, process nr:0, stack page=0018c394) Stack 00000007 00180018 00113683 001a0010 00000246 Code: 18 00 00 00 8e 1a 8e c2 e4 01 eb 00 eb 00 80 0d 68 06 18 00 Kfree of non-kmalloced memory: 0018d38c, next:00000000, order:0 task[0] (swapper) killed: unable to recover Kernel panic: Trying to free up swapper memory space in swapper task -not syncing Here's my setup: AMD 386dx40 Math coproc (forget name) SB16 Basic w/ ASP chip Turtle Beach Rio waveblaster daughterboard ATI Graphics Wonder 1meg ISA video card AMI bios 1 seagate 120 meg IDE hard drive 1 Western Digital 420 meg IDE hard drive 8 megs of RAM (8 1-meg 30 pin SIMMS -- 70ns)] Zoom faxmodem installed on COM2 14.4 kbps MS mouse 2.0 on COM1 I have no problems with 16-bit sound under DOS, BTW. Also, does anyone out there know of a MIDI player for linux that'll use the Rio daughterboard? Thanks in advance, =======* Tom Hanselman *===* Indiana University *===* Bloomington, IN *======= GCS -d+ H s g+(-) p?+ au a-- w(+) v C++ UL+ P? L+(++) 3- E+ N+(++) K W M(-) V -po+ Y t+ 5 j- R G+ tv D++ B--- e+ u(+) h f !r n(+) y? =================* Email: thanselm@silver.ucs.indiana.edu *================= ------------------------------ From: grif@corsa.ucr.edu (Michael Griffith) Subject: Re: Pentium-optimized compiler is slower then 486 gcc-2.5.8 on P5 !!! Date: 23 Sep 1994 01:57:53 GMT In article , Valery Petrov wrote: >Hello Linuxers, >I thought it would be interesting to know: > >I've just got a DELL XPS-90 and I installed and tested pentium-optimized gcc-i2.4.0p and the patched version found on sunsite.unc.edu (gcc-i2.5.8p). I've done some benchmarking for >integer and floating point calculations (plain C) and results for "pentium-optimized: -O4 -mpenium" code turned out to be even slower (5% on integer, 20% on floating point calculations) then the standart distribution gcc-2.5.8 code !!! So, don't mess with that Intel optimized compiler. I found different results. A few months ago, when I was playing with an mpeg decoder, I regularly saw 10-15% faster results with 2.4.0p. YMMV. -- Michael A. Griffith (grif@cs.ucr.edu) Department of Computer Science University of California, Riverside ------------------------------ ** 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-Misc-Request@NEWS-DIGESTS.MIT.EDU You can send mail to the entire list (and comp.os.linux.misc) via: Internet: Linux-Misc@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 sunsite.unc.edu pub/Linux End of Linux-Misc Digest ******************************