April 2004 - openSUSE AMD64 - openSUSE Mailing Lists

GA-K8VT800 Pro
by Jan Hendrik Berlin 19 Apr '04

19 Apr '04

Hi! There are 4 IDE-conectors on my mainboard. Thw first two are normaly! But the others are for a raid! Can I use this two conectors for may CD-R/RW? If I conect a cd-r oder cd-rw to the thrid or fourth conector, it isnt found by the system! What is the problem? regards Jan

2 1

[Fwd: Re: [suse-amd64] Questions about 9.1]
by Aurelio Caliaro 18 Apr '04

18 Apr '04

Yes, ATI 9200 is supported. I have an ATI 9800Pro - a special fanless edition from Sapphire, with a HUGE heat-spreader - but these new ATI cards are not supported in 64-bit by that open source driver with its 3D functions, just the simpler 2D functionalities. See what RDI says about this on http://dri.sourceforge.net/cgi-bin/moin.cgi/ATIRadeon . Excerpt: "Open source 3D acceleration is available on all radeons up to and including the 9200 (rv280). 3D support for r300 based cards (9500 and above) is only available from ati's binary driver." 3D drivers in 64 bit are not available at all for newer ATI cards, be it Windows or Linux. Greetings Aurelio -------- Original Message -------- Hi, On Saturday 17 April 2004 00:11, Laurent GUERBY wrote: > I've been told open source 3D driver was available for ATI 9200, > not top of the line, but the most powerful model with open source 3D I > could find by asking around (fedora list in this case, don't > know if SuSE includes a working DRI/3D supporting this). Yes, the readme for the X version that will come with 9.1 does say so. I took the gamble and bought my AMD64 with a ATI9200SE The other reason for it is that I wanted a silent system for reasonable price and this card is fanless. BB, Arjen -- Check the List-Unsubscribe header to unsubscribe For additional commands, email: suse-amd64-help(a)suse.com

1 0

Questions about 9.1
by Örn Hansen 17 Apr '04

17 Apr '04

I'm wondering about the nvidia support in SuSE 9.1. Currently, I'm unable to use nvidia driver with kernel 2.6, the compilation of the newest AMD64 5332 driver, breaks ... and when fixed, the kbuild script breaks. Another issue, is that I used the 2.4.21 kernel for a while, but found out that games that use the 3D extensions, looked screwed up, if they loaded at all. Probably due to the fact, that these games are 32bit, and the balance between 32bit and 64bit 3D rendering environment isn't too good. 1. Will there be a wroking nvidia driver, with SuSE 9.1. 2. Will 3D environment, be working for both 32bit and 64bit apps. Hope someone knows the answers.

9 10

YOU update
by Detlef Heinrich 17 Apr '04

17 Apr '04

Hello all, I don't have access to "ftp.suse.com" anymore. Anonymous Login and YOU fail. Do you have some guidance for me? kind regards ... Detlef

3 2

Fonts in gtk/gnome ...
by Örn Hansen 16 Apr '04

16 Apr '04

Anyone know of a quick way to restore fonts to gnome, I'm fumbling in the dark ... all texts are displayed empty, which is really annoying. Was wondering if there's anyway to create a sanity check, so that I won't be stumbling in the dark when a gtk application fires up.

1 0

amd64 / linux / java fud?
by suseamd64＠rgare.com 16 Apr '04

16 Apr '04

hello -- i'm sorry if this is not the appropriate place for this question. an hp loyalist has convinced some decision makers that we can't use opterons for an in house app because he claims there's no linux jvm that uses more than 3.8GB of memory. i couldn't find anything on the blackdown.org site backing this up, and couldn't find anything googling around. is he just spreading fud? does anyone here have experience and/or documentation that *nix 1.4+ java can use more than 3.8GB of ram? i just cringe when i see us purchasing super expensive quad itanium-2 boxes when our own in-house testing had our 2-way suse/opteron box out-performing a 4-way itanium(1) machine by 10%+. again, appologies if this would be better suited on the blackdown.org mailing list. info appriciated, -- john

4 4

Re: [suse-amd64] SuSE Updates: was -> [suse-amd64] apt, DVD (and avi) question.
by Robert_Thilsted＠ord.uscourts.gov 15 Apr '04

15 Apr '04

Thanks for your reply, Yes, I have tried both updated drivers with no luck. Next (Friday I hope), I will try the original SLES drivers and see if they work. It is an interesting problem as boot procedes normaly to the point of the mount of the real root filesystem. At this point it panics complaining it cannot mount / on ... It seems like it might be looking for the wrong (or non-existent) partition. sdb1 is a swap partition and sdb2 is the / partition These translate to 0x817 and 0x818, but it seems like I remember other numbers than 0x818. I will have to check. Could the System map be bad? The /dev files are correct. Boot options to grub are correct (and I have tried varying the boot partition to see if I could fix it). I flashed the SCSI card bios to the latest and greatest but this had no effect. I am currently running memtest86 (36 hours) so should see if there are any problems there.

1 0

Fw: [suse-amd64] Bandwidth on Quad Opteron
by Kevin_Gassiot＠veritasdgc.com 15 Apr '04

15 Apr '04

----- Forwarded by Kevin Gassiot/HOU/VES/VDGC on 04/15/2004 02:33 PM ----- ALBERTO D SCOTTI <ascotti(a)email.un c.edu> To Kevin_Gassiot(a)veritasdgc.com 04/15/2004 01:16 cc PM Subject Re: [suse-amd64] Bandwidth on Quad Opteron Kudos to Kevin for fishing out the -Mnontemporal flag! Since many people have asked me the asm files, I have posted them on ftp://ftp.unc.edu/pub/marine/ascotti/asm/ You will need to compile second_wall.c separately if you plan to use them. I have put the statically linked executables, in case you want to try them out and you do not have PGI fortran. Here's a summary of the results. Compiling without any optmization flag gives about 1.7 GB/s (stream_d.s) Compiling with -fastsse (shorthand for -Mvect=sse -Mscalarsse -Mcache_align -Mflushz) gives 2.1 GB/s (stream_d_sse.s) Adding -Mnontemporal brings it to 2.9 GB/s (stream_d_sse_nta.s) By replacing (by hand) prefetcht0 with prefetchnta, and increasing the prefetch offset to 256 from 64 bytes, the machine screams at 3.7 GB/s (stream_d_sse_nta_tweaked.s) Some comments 1) -Mconcur and -Mnontemporal will give broken code when running on more tha one CPU. It is probably due to the fact that nta operations require 16 byte alignment, which might be broken when distributing the blocks. I have not tried, but using MPI should be ok (since every threads allocates its own memory, and allocate() seems to always align on 16 bytes). 2) PGI refuses to use nontemporal opcodes when operating on float (STREAM uses double). This is a bummer. What's worse, it rearranges loops so that more often than not ntazing by hand does not work due to aligment problems. 3) -O3 does not seem to make a big difference. Well, that's it for now. I better go back and do some real work. It has been fun! Alberto Kevin_Gassiot(a)veritasdgc.com wrote: > Well, I found some references to compiling the code with the following > options : > > pgf77 ?O3 ?mp ?fastsse -Mnontemporal > I have compiled it this way to see if this gives me the same results. I > don't know enough about the PGI compiler to know if the -Mnontemporal > flag will do the same thing as what you did to the asm, but it sounds > like it might.... > > I am running a full lmbench run right now, so can't try the streams code > yet, but it will be interesting to see if this reproduces the desired > results :) > > > Kevin Gassiot > Advanced Systems Group > Visualization Systems Support > > > Veritas DGC > 10300 Town Park Dr. > Houston, Texas 77072 > 832-351-8978 > kevin_gassiot(a)veritasdgc.com <mailto:kevin_gassiot@veritasdgc.com> > > -----ascotti(a)email.unc.edu wrote: ----- > > To: Kevin_Gassiot(a)veritasdgc.com > From: ascotti(a)email.unc.edu > Date: 04/14/2004 07:45PM > cc: suse-amd64(a)suse.com > Subject: Re: [suse-amd64] Bandwidth on Quad Opteron > > Well, > > last night I have looked into the asm code generated by pgi fortran. > Using > the guidelines in the AMD software optimization publication for Opteron > (pub # 25112, p. 125), I have tweaked the asm just a little. (see diff > below for the copy loop). Essentially, I have replaced prefetch and > store > operations with their non temporal counterpart. To my surprise, the > bandwidth jumped for 2 GB/s for the compiled code to 3.7 GB/s for the > asm-tweaked code. I actually wonder if the executable that Daniel > sent me > (which gave 2.8 GB/s BTW) was not tweaked in a similar fashion. 3.7 > GB/s is > close to the value that AMD reports for this kind of operations. > Incidentally, a better results (15% according to AMD) could be > obtained by > grouping a bunch of prefetch operations, fill up L1 and then use > nontemporal store, rather than interleaving... > So it seems that my bandwidth problem was after all a compiler > issue. It > seems that STREAM actually can tell you the maximum bandwidth a > compiler > can give you (which, a posteriori, is kind of obvious). COmpilers > need to > be conservative, necessarily, but it would be nice to have the > possibility > of using directives (or pragmas in C) to instruct the compiler to > use for > example nontemporal instructions on a given loop. Also, it is > crucial that > arrays be aligned on a cache line, or else one gets a segfault. From > this > point of view, I am still not sure if allocate() in PGI fortran does > that > consistently... > If anybody is interested, I can e-mail the tweaked asm code for > stream... > > Cheers > > Alberto > > > *** 401,413 **** > # c(j) = a(j) > # 30 CONTINUE > movapd (%rsi,%rcx),%xmm0 > ! prefetcht0 64(%rsi,%rcx) > ! movapd %xmm0,(%rdx,%rcx) > movapd 16(%rsi,%rcx),%xmm1 > ! movapd %xmm1,16(%rdx,%rcx) > addq $32,%rcx > subl $4,%eax > jg .LB1_719 > # lineno: 179 > # t = mysecond() - t > # c(n) = c(n) + t > --- 401,414 ---- > # c(j) = a(j) > # 30 CONTINUE > movapd (%rsi,%rcx),%xmm0 > ! prefetchnta 256(%rsi,%rcx) > ! movntpd %xmm0,(%rdx,%rcx) > movapd 16(%rsi,%rcx),%xmm1 > ! movntpd %xmm1,16(%rdx,%rcx) > addq $32,%rcx > subl $4,%eax > jg .LB1_719 > + sfence > # lineno: 179 > # t = mysecond() - t > # c(n) = c(n) + t > > > > --On Wednesday, April 14, 2004 2:18 PM -0500 > Kevin_Gassiot(a)veritasdgc.com > wrote: > > > > > > > > > > > Well, I guess I am in the same boat. I was using a pre-compiled > > executable provided by someone else. I am trying to compile the code > > myself to see if I can reproduce the results.... > > > > Kevin > > > > > > > > Kevin Gassiot > > Advanced Systems Group > > Visualization Systems Support > > > > > > Veritas DGC > > 10300 Town Park Dr. > > Houston, Texas 77072 > > 832-351-8978 > > kevin_gassiot(a)veritasdgc.com > > > > > > > > ascotti(a)email.unc > > .edu > > > > To 04/13/2004 06:26 > Kevin_Gassiot(a)veritasdgc.com > > PM > > cc suse-amd64(a)suse.com > > > > Subject Please respond to Re: [suse-amd64] > > Bandwidth on Quad ascotti(a)email.unc Opteron > > .edu > > > > > > > > > > > > > > > > > > > > > > Interesting. Daniel Kedger sent me the STREAM executable > statically linked > > that he used on his machine. With that, I get 2.8 GB/s, which is > > believable, considering the slower CPUs on my box. However, I > cannot get > > the same numbers when I compile STREAM myself. What compiler/compiler > > switches did you use? > > > > Alberto > > > > > > --On Tuesday, April 13, 2004 4:14 PM -0500 > Kevin_Gassiot(a)veritasdgc.com > > wrote: > > > >> > >> > >> > >> > >> I just got a quad cpu - 32 GB system in, and am running STREAM > to confirm > >> bandwidth. I just got this machine loaded this afternoon, but I am > > seeing > >> numbers similar to what Daniel Kidger is seeing. After turning > off the > >> node interleaving in the BIOS, I get the following numbers : > >> > >> single cpu - 3170 MB/sec > >> 4 cpus - 12790 MB/sec > >> > >> I am running the 2.4.21-193-smp SuSE 9.0 professional kernel. > >> > >> Hardware - > >> > >> Quartet Motherboard > >> Phoenix BIOS - 09/26/2003 > >> 4 Opteron 846 cpus > >> 32 GB memory - 16 2 Gb DIMMs > >> > >> > >> BIOS - > >> > >> Dram Bank Interleave [AUTO] > >> Node Memory Interleave [Disabled] > >> ECC [Enabled] > >> > >> > >> I will probably take the covers off tomorrow to check out the > innards :) > >> > >> > >> > >> Kevin Gassiot > >> Advanced Systems Group > >> Visualization Systems Support > >> > >> > >> Veritas DGC > >> 10300 Town Park Dr. > >> Houston, Texas 77072 > >> 832-351-8978 > >> kevin_gassiot(a)veritasdgc.com > >> > >> > >> > >> ascotti(a)email.unc > >> .edu > >> > >> To 04/10/2004 01:14 suse-amd64(a)suse.com > >> PM > >> cc > >> > >> Subject Please respond to [suse-amd64] > Bandwidth on > >> Quad ascotti(a)email.unc Opteron > >> .edu > >> > >> > >> > >> > >> > >> > >> > >> > >> > >> > >> Hi all, > >> > >> I thought I had a pretty good system in terms of Bandwidth (2 > GB/s on a > >> single node, using STREAM benchmark), until Daniel Kidger, with a > >> considerable amount of understatement, pointed out that on their > system > >> (with faster CPUs but otherwise same memory type and > motherboard), they > >> get > >> > >> a "somewhat" higher value, 3.5 GB/s. 75% increase in bandwidth > is too > >> large > >> > >> not to investigate the matter further. Plus, all of the codes we > run on > >> this machine are bandwidth limited, so any improvement > translates in days > >> shaved from runs. The system is as follows > >> > >> Qartet MotherBoard > >> BIOS upgraded to version PQTDX0-B (9/26/2003). The original BIOS > gave > >> about > >> > >> 700 MB/s on 1 CPU! > >> 4 OPTERON 840 CPUs (1.4 GHz) > >> SLES8 SP3 > >> kernel 2.4.21-207-numa > >> 8 1Gb 333MHz PC2700 DIMMS (2 DIMMs per node, INFINEON brand) > >> STREAM compiled with pgf77 -fastsse -Mvect=prefetch > >> > >> The BIOS settings are > >> > >> Dram Bank Interleave [AUTO] > >> Node Memory Interleave [Disabled] > >> ECC [Enabled] > >> > >> Enabling Node Memory Interleave does not change the measured > bandwidth > >> significantly. I have also taken out 1 DIMM per node (which > should ensure > >> that the memory works in standard single channel mode). The > bandwidth > >> drops > >> > >> to 1.5 GB/s. In other words, dual channel seems to give a 35% > boost. I > >> have > >> > >> also tried a number of kernels (2.4.19, 2.4.24, 2.4.25, 2.6.4) with > >> similar > >> > >> results. Note that the system scales up well with NUMA kernels > (up to > >> about > >> > >> 7.5 GB/s with 4 CPUs). > >> > >> My questions are: > >> > >> 1) Can/should I do better? > >> 2) If there is a problem, where should I look for a solution? Is it > > likely > >> to be a hardware problem (e.g. lousy DIMMs), a BIOS problem (as I > >> mentioned > >> > >> earlier, the original BIOS gave me 700 MB/s), incorrect BIOS setting > > (what > >> else is out there, perhaps the SRAT table?) a kernel problem or a > > compiler > >> issue? > >> > >> I'd rather have your input before pestering the vendor and > Celestica. > >> > >> Thank you for your time > >> > >> Alberto Scotti > >> > >> > >> > >> > >> > >> > >> > >> > >> -- > >> Check the List-Unsubscribe header to unsubscribe > >> For additional commands, email: suse-amd64-help(a)suse.com > >> > >> > >> > > > > > > > > > > > > -- > > Check the List-Unsubscribe header to unsubscribe > > For additional commands, email: suse-amd64-help(a)suse.com > > > > > > > > > > > -- Alberto Scotti Asst. Prof. Dept. of Marine Sciences CB 3300 University of North Carolina Chapel Hill, NC 27599-3300 919-962-9454 (w) 919-962-1254 (f)

1 0

RE: [suse-amd64] amd64 / linux / javva fud?
by suseamd64＠rgare.com 15 Apr '04

15 Apr '04

Thank you for the information Joe, Andi, and Richard! :D -- john

1 0

SSHD Segmentation Fault on AMD Opteron 248 SUSE 9.0
by Michael Gasch 15 Apr '04

15 Apr '04

hi there... i recently experienced some weird things with latest openssh and different ssh linux clients if i connect or scp to my server (AMD64 Suse 9.0) from different Suse 8.2/9.0 clients with no given username (so the current username of the user, who executes ssh on his client is used) everything works fine if i give a username (user@server or -l user - no matter with or without pubkey) i get Have a lot of fun... connection closed in the logs i see something like Apr 15 19:01:05 nevanpdc sshd[2902]: Accepted keyboard-interactive/pam for root from ::ffff:192.168.0.1 port 49435 ssh2 Apr 15 19:01:05 nevanpdc kernel: sshd[2905]: segfault at 0000002a00000000 rip 0000002a966b5870 rsp 0000007fbfffe218 error 4 normally it should work ?!?!? is it a known problem? strange is: after a reboot it works for some time, until somebody tries a connection with pubkey to an other account then his own (e.g. root) then i get seg fault and after this the problems as described above start thx micha -- "Matrix - more than a vision" ************************************************** Michael Gasch - Central IT Department - Max Planck Institute for Evolutionary Anthropology Deutscher Platz 6 04103 Leipzig Germany **************************************************

1 0