Re: Xorg 7.0
Hans wrote: Am Donnerstag, 13. April 2006 08:40 schrieb Gudjon I. Gudjonsson: Hi Gudjon ! Same to me ! I had just upgraded to 7.0, when I read your message. Now my X does not start any more. Your message just came 2 minutes too late. I hope, that the missing packages will be handy soon. In 32-bit they are already there, but 64-bit will last a little bit I suppose. You can get nearly all xorg packages from ftp://ftp.de.debian.org. Just add the following line to /etc/apt/source.list: deb http://ftp.fi.debian.org/debian unstable main contrib However, xbase-clients is missing from the archives, which is needed to have the correct keymap installed. I have compiled a version from sources, which you can find here: http://www.fugmann.net/~afu/amd64/xbase-clients_7.0.0-2_amd64.deb Remember also to install xkb-data, and modify the xorg configuration file (/etc/X11/xorg.conf) to use xorg keyboard rules (may installations uses xfree86, which xorg did not complain about before). You keyboard section should look something like this: Section "InputDevice" Identifier "Generic Keyboard" Driver "kbd" Option "CoreKeyboard" Option "XkbRules" "xorg" Option "XkbModel" "pc102" Option "XkbLayout" "dk" Option "XkbOptions""ctrl:nocaps" EndSection (change the model, layout and options to suit your needs). best regards Hans Regards Anders Fugmann -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: turion/mobile athlon benchmark
A. P. Kennedy wrote: ** Turion below but with 32 bit kernel *** # /usr/lib/atlas3-test/xzinvtst NREPS ORDER UPLO NLDA TIME MFLOP RESID = = = = = 0Col GE100100 0.008996.37 6.066196e-03 0Col GE200200 0.045 1419.78 5.083446e-03 0Col GE300300 0.154 1401.06 4.577686e-03 0Col GE400400 0.353 1449.29 5.339619e-03 0Col GE500500 0.684 1461.11 4.390384e-03 0Col GE600600 1.172 1473.71 4.389517e-03 0Col GE700700 1.865 1470.75 4.753297e-03 0Col GE800800 2.793 1466.06 5.002779e-03 0Col GE900900 3.912 1490.02 5.070376e-03 0Col GE 1000 1000 5.278 1515.10 5.078879e-03 10 cases: 10 passed, 0 skipped, 0 failed # cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 15 model : 36 model name : AMD Turion(tm) 64 Mobile Technology ML-32 stepping: 2 cpu MHz : 1800.307 cache size : 512 KB fdiv_bug: no hlt_bug : no f00f_bug: no coma_bug: no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt lm 3dnowext 3dnow pni lahf_lm bogomips: 3601.38 It would be great is you were able to run these tests using a 64 bit kernel and a 64 bit version of atlas3. It seems that there is a huge performance win for atlas under 64 bit, as I cannot imagine that the difference in 2. level cache (one half of that of a turion MT) can give this huge difference in MFlops. Gathering some of the data (Dual system excluded): MFLOPS MHz MFLOPS/MHz System 3113.9 20101.55Opteron 2879.13 20091.43Std amd64 3367.67 25561.33Newcastle 2148.33 16071.34Turion MT 3317.38 24111.38Std amd64 1515.1 18000.84Turion ML - 32 bit Reading this table (disregarding the 32 bit test) it seems that the Turion processors is just as fast as a standard opteron in terms of MFlops/MHz. The opterons beat every other AMD based CPU in this regard, which would be expected. One should note that the Turion MT-30 has a max power consumption of 25 Watt, where as the Opteron tops at 89 Watt IIRC - Huge gain in FLOPS / Watt when using a turion processor. Regards Anders Fugmann -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: No swap on my Debian Sid system
Jack Malmostoso wrote: On Sun, 22 Jan 2006 19:31:09 +0100, Jack Malmostoso wrote: I have added two swap entries in /etc/fstab but have not rebooted yet to try them (work in progress :()... Upon reboot swap was activated: [EMAIL PROTECTED]:~$ dmesg | grep swap Adding 498004k swap on /dev/sda2. Priority:-1 extents:1 across:498004k Adding 498004k swap on /dev/sdb2. Priority:-2 extents:1 across:498004k Would it give any advantage to RAID the swap too or is it just good like this? As pointed out by others on the list, having swap on a raid partition allows you system to survive a disc crash. It will, however, be a bit slower than swapping on the harddrive directly. Even more speed can be gained by specifying the same priority for both swap partitions: eg: /dev/sda2 none swap sw,pri=1 0 0 /dev/sdb2 none swap sw,pri=1 0 0 Which will tell the vm that it should span the swapped out data over the two partitions, which will yeild higher read performance. Regards Anders Fugmann -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: problem with firefox and mozilla
Dan Merillat wrote: On 8/23/05, Harald Wenninger <[EMAIL PROTECTED]> wrote: Hi, I have problems with mozilla+firefox. I am using the unstable distribution. Every time I want to start firefox or mozilla, no window is opened. The .mozilla-dir and its subdirectories get created, though. There's no error displayed in the terminal from which I start the browsers. Does anybody know somethin about this problem? I've been reporting this problem for a few months now, but nobody seems able to reproduce it. I surely can reproduce it. Native 64, running chrooted X apps. Simple things (Xterm, Clock, etc) work. Konquerer, Firefox, Mozilla fail. I finally tried compiling the debian 0.92.0 wmaker package with gcc-3.4 (3.4.5 20050821), and behold - firefox and thunderbird starts every time. I had to disable MMX and X86 asm as the instructions used were illegal for the 64bit platform (according to gcc). It anyone else wants to try the package, you can retrieve it at: http://fugmann.net/~afu/amd64/wmaker_0.92.0-1_amd64.deb I will try later to compile using gcc-4.0 to see if the instructions used are the cause of the problems or if compiling with gcc-4.0 somehow triggers a problem in wmaker. I'm using unstable btw. Regards Anders Fugmann -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: ASUS A8V Deluxe, SATA disk on the Promise port?
Lennart Sorensen wrote: The second sata is a promise 20378. I highly doubt it will work with the sil driver. Me neither. Any driver not made to drive the promise controller would not work I guess. I have not found a free driver yet (the ft3xx driver is supposed to work but appears to be binary only). How about the promise sata driver? I have been using this for a while now with no problems at all for a A8V Deluxe motherboard. Of course if you can find an add in pci sil3112a card that would work. Seems to be one of the best supported chipsets so far. No need to. A standard kernel compiled with SCSI_SATA_PROMISE=[y|m] should do the trick. Len Sorensen Regards Anders Fugmann -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Nvidia problems... again
James Titcumb wrote: Hello all, I've managed to set up nvidia drivers before on this amd64 machine, but since I've upgraded to my new kernel 2.6.11 with realtime-lsm, nvidia module doesn't want to work. Have you tried using a vanilla kernel (e.g. 2.6.11.7), without the realtime-lsm patches? Also I have noticed that NVidia does not work well with unordered IO. (CONFIG_UNORDERED_IO). I would check to see is this was set in your kernel congiguration. (**) NVIDIA(0): Option "NvAGP" "1" (**) NVIDIA(0): Option "TwinView" "true" (**) NVIDIA(0): Option "TwinViewOrientation" "RightOf" (**) NVIDIA(0): Option "SecondMonitorHorizSync" "30.0-75.0" (**) NVIDIA(0): Option "SecondMonitorVertRefresh" "50.0-85.0" (**) NVIDIA(0): Option "MetaModes" "1280x1024, 1280x1024; 1280x1024, 1280x1024;" Is the culprit the AGP setting? Try setting it to 2 (use Linux own AGPgart driver) or 0 (no not initialize AGP) for debugging purposes. I never understood why NVidia supplied their own AGP driver, as I have never made it work properly. Also make sure that the graphics card has an IRQ (look at lspci -v output). If you have compiled with ACPI, you system log should print a line like: ACPI: PCI Interrupt :01:00.0[A] -> GSI 16 (level, low) -> IRQ 16 NVRM: loading NVIDIA Linux x86_64 NVIDIA Kernel Module 1.0-7174 Tue Mar 22 06:45:40 PST 2005 When loading the nvidia kernel module. Looking further at your log, I see that you try loading the dri and glcore modules. These should be replaced by GLX module when using the 'nvidia' driver. This, however, should not result in the error you are seeing. Hope it helps Anders Fugmann -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: nfs server problems with gcc-3.4
Andreas Jochens wrote: I just uploaded two different new libc6 versions (-19.0.0.2.gcc4 and -19.0.0.3.gcc4) to http://debian-amd64.alioth.debian.org/gcc-3.4/glibc It would be nice if someone could check if one or maybe both of those versions fix the nfs problems before I make an upload to the gcc-3.4 archive. Regards Andreas Jochens Both versions works on my system. No long wait when mounting nfs, or staring the nfs-kernel-server. Great work! Regards Anders Fugmann
[OT] - LM-sensors settings.
Running a bit off topic, I changed the subject. Sebastian Steinlechner wrote: On Fri, 2004-10-15 at 21:48, Anders Peter Fugmann wrote: Other than the -12V, I see no bogus values, other than the stopped fans. (Maybe the V5SB is a bit off, but it can be your power supply.) I also mean in regard to the min/max settings... and a high-value of +4°C for mainboard temperature doesn't seem to make much sense either. Whatever, I'm positively surprised that it actually works - that I get a read-out at all. All min max settings are controlled from /etc/sensors.conf. These can easily be set to something sane. I know that it's usually not a good idea to write such a script on my own when there are better solutions already available, but my understanding from e.g. the fancontrol manpage was that it needs the fan rpm readout - and that doesn't work (yet). Actually, the fancontrol program does not require RPM readings from the fan - it is optional. You can leave it out safely. Regards Anders Fugmann
Re: Asus K8V - acpi/cpufreq
Sebastian Steinlechner wrote: On Sun, 2004-10-10 at 17:56, Anders Peter Fugmann wrote: -- VCore: +1.10 V (min = +1.71 V, max = +1.89 V) +3.3V: +3.30 V (min = +3.14 V, max = +3.47 V) +5V: +5.00 V (min = +4.76 V, max = +5.24 V) +12V: +11.25 V (min = +10.82 V, max = +13.19 V) -12V: +0.30 V (min = -13.18 V, max = -10.80 V) -5V: +5.10 V (min = -5.25 V, max = -4.75 V) V5SB: +5.51 V (min = +4.76 V, max = +5.24 V) VBat: +0.02 V (min = +2.40 V, max = +3.60 V) fan1:0 RPM (min = 37500 RPM, div = 2) fan2:0 RPM (min = 337500 RPM, div = 2) temp1: +29°C (high =+4°C, hyst =+0°C) sensor = thermistor temp2: +30.0°C (high = +80°C, hyst = +75°C) sensor = thermistor alarms: Chassis intrusion detection ALARM beep_enable: Sound alarm disabled -- It's easy to see there are some completely bogus values in there. I Other than the -12V, I see no bogus values, other than the stopped fans. (Maybe the V5SB is a bit off, but it can be your power supply.) don't know about the fan speed - the bios correctly reports it, but it doesn't show here (and I'm _really_ interested in seeing a fan that does 337500 rpm as a minimum...). Whatever, I found the two measured Maybe you need to increase the 'div' on the fans to get a reading. Try something like: echo 32 > /sys/bus/i2c/devices/0-0290/fan1_div echo 32 > /sys/bus/i2c/devices/0-0290/fan2_div to read out the fan rotation when they are running at low speeds. temperatures to be correct. temp1 is motherboard temp, temp2 is cpu temp. fan2 is cpu fan here, and I can throttle it to as low as 120, which stabilizes cpu temperature at about 32°C during normal typing work (so, idle most of the time) on a 3200+ using powernowd. Probably I could even switch it off, but air flow in my box isn't too good (and frankly, I don't care - on 120, the fan isn't audible anyway.) I wonder what happens to the CPU temperature when you run 'while true; do true:done' for ten minutes, or if you try render something with povray. Static fan settings can be dangerous, when set too low. Luckily the AMD64 processors all have thermal protection. Regards Anders Fugmann
Re: Asus K8V - acpi/cpufreq
Sebastian Steinlechner wrote: I don't know about fan speed - both /proc/acpi/thermal_zone and /proc/acpi/fan are empty here, so I have no idea how to get the necessary information. On the other hand, the K8V bios has a "Q-Fan control" option that's supposed to do fan throttling. However, I don't know what thermal trip points it uses so I'm a bit worried to turn it on... I have only experience with the Asus A8V board, but I expect it to be similar wrt. Q-FAN and Fan Throttling to the K8V board. My experience is that Q-FAN is useless. It's lowest voltage is 11/16*12 8 volt which it way too high for Fan (Using water cooling with a big radiator). Like you, I cannot find any information on trip points other that for The P4, which runs at much higher temperatures than the AMD64. Fans are not controlled through ACPI (AFAIK, only notebooks uses this). However, lm-sensors let you control the fan speeds from userspace. To use this, you should install the lm-sensors package, and paste the following lines into /etc/modules: # I2C adapter drivers i2c-isa # I2C chip drivers w83627hf (Or run sensors-detect, which will probe for the chip installed on you motherboard, and instruct you how to install the correct modules). To see if it worked, try running the program 'sensors' which will print out temperatures, voltages and fan speeds. When the modules are loaded, the fans are set to full throttle, clearing the BIOS Q-FAN settings. To control the fans, write an integer in the range [0;240] to /sys/bus/i2c/devices//fan[n]_pwm, where 'id' is the id of the chip (usually there is only one directory), and 'n' is the number of your fan (usually between 1 and 3). E.g. echo 0 > /sys/bus/i2c/devices/0-0290/fan2_pwm to stop fan 2 (my case fan). use 240 for full throttle. Instead of controlling the fan speed manually, I the use the program 'fancontrol' (/usr/sbin/fancontrol) which is part of the lm-sensors package. The program is a simple shell script which adjusts the fan speeds based on CPU/MB temperature. It can be configured through the program /usr/sbin/pwmconfig. See /usr/share/doc/lm-sensors/doc/fancontrol.txt for more information. Anyways, thanks everyone for your help! In search of a low noise setup, I found that my AMD64 3500+ accepts much lower core voltage than the specified 1.55 V. Without overclocking it runs stable at 2GHz (PowerNow disabled, as it does not let me control the core voltage) using a core voltage of 1.30, and thus running much cooler (If anyone knows how to control the core voltage using powernow, I would be happy). Hope it helps. Anders Fugmann
Re: Promise or VIA?
Pete Harlan wrote: On Wed, Sep 22, 2004 at 09:37:53AM -0400, Lennart Sorensen wrote: The Asus A8V I bought has both a VIA and a Promise SATA controller, and both work fine with Linux. The Promise is better supported under Linux (or possibly just a better controller; it does TCQ under Linux, where the VIA doesn't (yet?)) as far as I could tell from the SATA compatibility page. The status page referenced elsewhere in this thread outlines the capabilities of the hardware. The promise controller supports TCQ, where as the VIA does not. In kernel support is, however, not ready yet, but is in the pileline as I understand it. I would advice to configure the system to use the promise controller if possible, as the Linux-kernel _will_ support TCQ in the future, and at that point the VIA controller will lack behind, and a hardware reconfiguration would be nessesary to take advantage of the TCQ if initially configured using the VIA controller. Regards Anders Fugmann
Re: gcc-3.4 builds
Thanks for your reply. I hope you will give a status of when the port is ready for public test :-). Regards Anders Fugmann P.s. Your ip address is listed in SORBS (http://www.nl.sorbs.net/), as beeing a dynamically assigned ip address. You should consider using you ISP as mail relay for sending mails, or send SORBS a note. Andreas Jochens wrote: On 04-Jul-26 19:05, Anders Peter Fugmann wrote: Hi, I see a new directory on alioth named gcc-3.4, which seems to contain alot of packages. I guess that this is all the packages compiled with gcc-3.4, which generates better code for x86-64 than gcc-3.3. Is this a complete port (compared to pure64 on alioth) or should I sill have pure64 in my sources.list for packages that are not yet compiled using gcc-3.4? Currently all packages are being recompiled with gcc-3.4 as the default compiler. The resulting packages are being uploaded to the new repository in the gcc-3.4 directory. Every package is being compiled in clean chroot environment with only the specified Build-Depends installed. Building about 8500 packages takes some time even on an amd64 machine. At the moment the gcc-3.4 repository has about 1200 packages and it includes all packages with priorities 'Required', "Important' and 'Standard' as well as all packages from the base system which are necessary to debootstrap a chroot environment. However, the upload of the missing ~7000 packages will take a few more days. Also, is there any issues to be aware of when mixing gcc 3.4 compiled packages with gcc 3.3 compiled packages? I read that gcc 3.4 has changed ABI from gcc 3.3, but I'm not sure which architectures this affectes. I would recommend _not_ to use the packages from the gcc-3.4 repository until the conversion of the archive has been completed. C++ programs compiled with g++-3.3 use the library from the libstdc++5 package while programs compiled with g++-3.4 use the one from libstdc++6. There are (minor) incompatibilities between those two libraries. Regards Andreas Jochens
gcc-3.4 builds
Hi, I see a new directory on alioth named gcc-3.4, which seems to contain alot of packages. I guess that this is all the packages compiled with gcc-3.4, which generates better code for x86-64 than gcc-3.3. Is this a complete port (compared to pure64 on alioth) or should I sill have pure64 in my sources.list for packages that are not yet compiled using gcc-3.4? Also, is there any issues to be aware of when mixing gcc 3.4 compiled packages with gcc 3.3 compiled packages? I read that gcc 3.4 has changed ABI from gcc 3.3, but I'm not sure which architectures this affectes. Regards Anders Fugmann
Re: working mozilla, ffox and tbird packages in unstable pure64.
Gerard H. Pille wrote: Unfortunately, I have only been to school till 260856, so maybe if you could elaborate a little? He is referring to a bug number in the debian bug tracking system. See: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=260857 To patch download the attached patch and do: $ patch /usr/bin/mozilla-1.7.1 < /tmp/mozilla.patch (assuming you downloaded the patch as /tmp/mozilla.patch) Regards Anders Fugmann
Re: installation report/problems with xsane and subversion
Alex Perry wrote: Unless I'm missing something, those packages depend on things that are neither in that directory nor available on alioth: $ sudo dpkg -i mozilla-browser_1.7-5_amd64.deb Password: Selecting previously deselected package mozilla-browser. (Reading database ... 55818 files and directories currently installed.) Unpacking mozilla-browser (from mozilla-browser_1.7-5_amd64.deb) ... dpkg: dependency problems prevent configuration of mozilla-browser: mozilla-browser depends on libgcc1 (>= 1:3.4.1-1); however: Version of libgcc1 on system is 1:3.3.4-3. mozilla-browser depends on libstdc++6 (>= 3.4.1-1); however: Package libstdc++6 is not installed. mozilla-browser depends on libnspr4 (= 2:1.7-5); however: Version of libnspr4 on system is 2:1.7.1-1. dpkg: error processing mozilla-browser (--install): dependency problems - leaving unconfigured Errors were encountered while processing: mozilla-browser $ apt-cache showpkg libstdc++6 Package: libstdc++6 Versions: Reverse Depends: mozilla-browser,libstdc++6 3.4.1-1 Dependencies: Provides: Reverse Provides: $ Sorry, you will also need the gcc 3.4.1 packages found in http://debian-amd64.alioth.debian.org/pure64/tmp/ Regards Anders Fugmann