From ersoz ＠ cse.psu.edu Wed Sep 3 03:53:27 2003 From: ersoz ＠ cse.psu.edu (Deniz Ersoz) Date: Tue, 2 Sep 2003 14:53:27 -0400 Subject: [SCore-users-jp] [SCore-users] SCore Physical memory limit Message-ID: <20030902185327.GA10195@titan.cse.psu.edu> Hi, I am trying to see the effect of swapping in Score. I want to run three MG.8.C's (NAS Benchmark) at the same time on an 8 node cluster. Single MG.8.C needs around 450MB of memory and the nodes in the cluster have 1GB on them. However I wasn't able to get a logical result yet. First I tried changing queue limits with the command: limit queue all memory 300 after i did this I couldn't run anything. Even the first job I submit is rejected. then i tried first changing the machine limits as: limit cluster all memory 1500 then submit thhe jobs with "memorylimit=500" parameter this way I was be able to run three MG.8.C's at the same time. But this time it took more than 25 hours to finish all three jobs. This is weird because single MG.8.C takes around 150 seconds. Have you got any suggestions that can improve the performance with swapping in SCore? Or is it normal to get very long execution times? Thanks, Deniz Ersoz _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From igor_makunin ＠ nightmail.ru Wed Sep 3 18:53:32 2003 From: igor_makunin ＠ nightmail.ru (=?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?=) Date: Wed, 3 Sep 2003 13:53:32 +0400 Subject: [SCore-users-jp] [SCore-users] SCore patch on server Message-ID: <20030903135332.399f5bc3.igor_makunin@nightmail.ru> Hello. I have SCore 5.4.0 installed from sources on cluster (1 server and 8 compute hosts). Have I to patch kernel (SCore patch) on server host if it's not involved in calculations? Thanks. ---------- Igor Makunin. _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From kameyama ＠ pccluster.org Wed Sep 3 18:58:40 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Wed, 03 Sep 2003 18:58:40 +0900 Subject: [SCore-users-jp] Re: [SCore-users] SCore patch on server In-Reply-To: Your message of "Wed, 03 Sep 2003 13:53:32 JST." <20030903135332.399f5bc3.igor_makunin@nightmail.ru> Message-ID: <20030903095203.D99DC128944@neal.il.is.s.u-tokyo.ac.jp> In article <20030903135332.399f5bc3.igor_makunin ＠ nightmail.ru> =?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?= wrotes: > I have SCore 5.4.0 installed from sources on cluster (1 server and 8 compute > hosts). > > Have I to patch kernel (SCore patch) on server host if it's not involved in c > alculations? You don't change kernel on the server if the server dose not combined with Compute Host. from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From igor_makunin ＠ nightmail.ru Thu Sep 4 00:00:59 2003 From: igor_makunin ＠ nightmail.ru (=?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?=) Date: Wed, 3 Sep 2003 19:00:59 +0400 Subject: [SCore-users-jp] [SCore-users] Problems with rpmtest Message-ID: <20030903190059.608255f1.igor_makunin@nightmail.ru> Hello. I have SCore 5.4.0 installed from sources. I have one server "server" and two computation hosts "node-7" and "node-8". I have some problems with PM\Ethernet Test Procedure. According to http://pccluster.org/score/dist/score/html/en/installation/pm-testethernet.html I've done: 1st step: I've done "/etc/rc.d/init.d/pm_ethernet start " on hosts ("node-7" and "node-8") Result: [root ＠ node-7 root]# /etc/rc.d/init.d/pm_ethernet start Starting PM/Ethernet: device: eth0 [root ＠ node-7 root]# At "node-8" the same. 2nd step: 1st window on the server host: rpmtest node-7 ethernet -reply 2nd widown on the server host: rpmtest node-8 ethernet -dest 1 -ping PM_DEBUG was set to 5. Result in 1st window: [makunin ＠ server score]$ rpmtest node-7 ethernet -reply ethernet_open_device(): -config /var/scored/scoreboard/server.0000B400F-Kn pmEthernetOpenDevice: Library version $Id: pm_ethernet.c,v 1.64 2002/03/04 09:44:42 s-sumi Exp $ pmEthernetReadConfig(0x83b93d0, unit, 0): set unit number "0" (MAX: 4). pmEthernetReadConfig(0x83b93d0, backoff, 4800): set backoff "4800" usec. pmEthernetReadConfig(0x83b93d0, checksum, 0): set checksum "0" off. pmEthernetGetNodeByNumber(0x83b93d0, 0, 0xbffff4f8): not found pmEthernetGetNodeByNumber(0x83b93d0, 0, 0xbffff4f4): not found store host xx:xx:xx:xx:xx:xx: dev{0] mac 0x0000000xxxxxxxxx pmEthernetGetNodeByNumber(0x83b93d0, 1, 0xbffff4f8): not found pmEthernetGetNodeByNumber(0x83b93d0, 1, 0xbffff4f4): not found store host xx:xx:xx:xx:xx:xx: dev{0] mac 0x0000000xxxxxxxxx pmEthernetOpenDevice("/var/scored/scoreboard/server.0000B400F-Kn", 0xbffffa04): pmEthernetMapEthernet(0, 0xbffff858): 0 Ethernet(0): fd=512 self node-7 n 0 of 2 nodes pm_ethernetCalibrateTimer(): loop t:1.912734e+07, vt: 1.247400e-02 pm_ethernetCalibrateTimer(): loop t:3.060018e+07, vt: 1.995700e-02 pm_ethernetCalibrateTimer(): end loop t:3.060018e+07, vt: 1.995700e-02 pm_ethernetCalibrateTimer(): d0:1.533377e+09, d1:1.533306e+09 pm_ethernetCalibrateTimer(): clk:1533, clock 1.533341e+03 pmEthernetOpenDevice: Driver version $Id: pm_ethernet_dev.c,v 1.4 2003/02/03 03:11:42 kameyama Exp $ ethernet_open_device(): success [0](0) PM/Ethernet CTX map Ctx(0x83ba3a0): send=0x40161000, recv=0x40181000, shared=0x40016000 success [0](0)pmEthernetRegisterProc(): proc 8717(8717), tid 8717 [0](0) pmEthernetAssociateNodes(0x83ba3a0, 0x83b9ae0, 2):ndev=1 [0](0) pmEthernetBindChannel(0x83ba3a0, 0, 0): called Result in 2nd window: [makunin ＠ server score]$ rpmtest node-8 ethernet -dest 1 -ping ethernet_open_device(): -config /var/scored/scoreboard/server.0000B400F-Kn pmEthernetOpenDevice: Library version $Id: pm_ethernet.c,v 1.64 2002/03/04 09:44:42 s-sumi Exp $ pmEthernetReadConfig(0x83b93d0, unit, 0): set unit number "0" (MAX: 4). pmEthernetReadConfig(0x83b93d0, backoff, 4800): set backoff "4800" usec. pmEthernetReadConfig(0x83b93d0, checksum, 0): set checksum "0" off. pmEthernetGetNodeByNumber(0x83b93d0, 0, 0xbffff4d8): not found pmEthernetGetNodeByNumber(0x83b93d0, 0, 0xbffff4d4): not found store host xx:xx:xx:xx:xx:xx: dev{0] mac 0x0000000xxxxxxxxx pmEthernetGetNodeByNumber(0x83b93d0, 1, 0xbffff4d8): not found pmEthernetGetNodeByNumber(0x83b93d0, 1, 0xbffff4d4): not found store host xx:xx:xx:xx:xx:xx: dev{0] mac 0x0000000xxxxxxxxx pmEthernetOpenDevice("/var/scored/scoreboard/server.0000B400F-Kn", 0xbffff9e4): pmEthernetMapEthernet(0, 0xbffff838): 0 Ethernet(0): fd=512 self node-8 n 1 of 2 nodes pm_ethernetCalibrateTimer(): loop t:2.006119e+07, vt: 1.308300e-02 pm_ethernetCalibrateTimer(): loop t:3.060114e+07, vt: 1.995700e-02 pm_ethernetCalibrateTimer(): end loop t:3.060114e+07, vt: 1.995700e-02 pm_ethernetCalibrateTimer(): d0:1.533379e+09, d1:1.533354e+09 pm_ethernetCalibrateTimer(): clk:1533, clock 1.533366e+03 pmEthernetOpenDevice: Driver version $Id: pm_ethernet_dev.c,v 1.4 2003/02/03 03:11:42 kameyama Exp $ ethernet_open_device(): success [0](0) PM/Ethernet CTX map Ctx(0x83ba3a0): send=0x40161000, recv=0x40181000, shared=0x40016000 success [1](0)pmEthernetRegisterProc(): proc 6659(6659), tid 6659 [1](0) pmEthernetAssociateNodes(0x83ba3a0, 0x83b9ae0, 2):ndev=1 [1](0) pmEthernetBindChannel(0x83ba3a0, 0, 0): called [1](0) pmEthernetGetSendBuffer(0x83ba3a0, 1, 0xbffff9fc, 8): Send to myself 1 Ethernet PM context #0 information (unit 0) channel 0 descripter information rx_p=00000000, rx_c=00000000, rx_bp=00000000, rx_bc=00000000 tx_p=00000000, tx_c=00000000, tx_bp=00000000, tx_bc=00000000 === ack control of node 0 === seq= 00000001, s_acd= 00000000, stat=00000000, ssent= 00000000 s_ack=ffffffff, s_mxrv=00000000, sed= 00000001, rtime=00000000 st_txmit=00000000, st_rexmit=00000000, st_xmit_ctl=00000000 st_xmit_ack=00000000, st_xmit_lost=00000000, st_xmit_stop=00000000 st_received=00000000, st_rcv_valid=00000000, st_rcv_ackonly=00000000 st_rcv_ignre=00000000, st_rcv_lose=00000000, st_rcv_ov=00000000 st_rcv_stop=00000000, st_rcv_go=00000000 st_xmit_err=00000000 === end === === ack control of node 1 === seq= 00000000, s_acd= f7e40000, stat=00000000, ssent= 00000000 s_ack=00000000, s_mxrv=00000000, sed= 82750400, rtime=00000000 st_txmit=00000000, st_rexmit=00000000, st_xmit_ctl=00000001 st_xmit_ack=00000000, st_xmit_lost=00000000, st_xmit_stop=00000000 st_received=00000000, st_rcv_valid=00000000, st_rcv_ackonly=00000000 st_rcv_ignre=00000000, st_rcv_lose=00000000, st_rcv_ov=00000000 st_rcv_stop=00000000, st_rcv_go=00000000 st_xmit_err=00000000 === end === TX descriptor dump p 0 c 0 PC 0 (nil), 0 00000000, 00000000, 00000000 ......... e4 (nil), 0 00000000, 00000000, 00000000 select: Bad file descriptor [makunin ＠ server score]$ What's wrong? Test "scstest" works well. Thanks. --------------- Igor Makunin. _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From Motzkus ＠ ifr.de Thu Sep 4 03:18:17 2003 From: Motzkus ＠ ifr.de (=?iso-8859-1?Q?Andr=E9_Motzkus?=) Date: Wed, 3 Sep 2003 20:18:17 +0200 Subject: [SCore-users-jp] [SCore-users] Some probs when installing IBM HS20 Blades (Bladecenter) Linux Re dHat 7.3 engl. Message-ID: <76970123516DD611B30300105A66B2A80AA33F@IFRSERVER01> Hello @ all, i join your Score 5.4.0 by using the EIT. (I was creating the boot floppy (using Read Hat 7.3). When i come to the point for Installing the cluster hosts, i´ve got some trouble. when i was trying to boot the first HS20 Blade, it stops with the following message: "No dhcp server specified. Used broadcast SIOCSIFADDR: no such device try it again SIOCSIFADDR: no such device try it again SIOCSIFADDR: no such device try it again Configure network fails NFS mount: /mnt/runtime cannot mount exiting See the documentation for this trouble" What could it be? I´ve used the GigEthernet-Driver(e1000, bcm5700) (the Blades have a Broadcom NetXtreme Gigabit Ethernet Interface) If anybody could help, i would be absolutely happy about!!! Thanxx a lot, Andre _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From rene.storm ＠ emplics.com Thu Sep 4 04:02:29 2003 From: rene.storm ＠ emplics.com (Rene Storm) Date: Wed, 3 Sep 2003 21:02:29 +0200 Subject: [SCore-users-jp] AW: [SCore-users] Some probs when installing IBM HS20 Blades (Bladecenter) Linux Re dHat 7.3 engl. Message-ID: <29B376A04977B944A3D87D22C495FB2301279D@vertrieb.emplics.com> Hi Andre, I haven't tested Score 5.4 yet, but I remember there was some trouble with the e1000. Newer chips (e.g. 82545EM) aren't recognized by a standard pci_ids list. (Intel gives all e1000 cards (generations) their own identifier.) You will have to insert a new driver (from intel) to the gigabit boot disk. To your bcm5700 (tg3). Seems this module isn't supported by score. Same problem, same solution: Read: http://www.pccluster.org/score/dist/score/html/en/faq-tips/mk-eit-floppy.html Regards Rene Storm emplics ag -----Ursprüngliche Nachricht----- Von: André Motzkus [mailto:Motzkus ＠ ifr.de] Gesendet: Mittwoch, 3. September 2003 20:18 An: 'SCore-users ＠ pccluster.org' Betreff: [SCore-users] Some probs when installing IBM HS20 Blades (Bladecenter) Linux Re dHat 7.3 engl. Hello @ all, i join your Score 5.4.0 by using the EIT. (I was creating the boot floppy (using Read Hat 7.3). When i come to the point for Installing the cluster hosts, i´ve got some trouble. when i was trying to boot the first HS20 Blade, it stops with the following message: "No dhcp server specified. Used broadcast SIOCSIFADDR: no such device try it again SIOCSIFADDR: no such device try it again SIOCSIFADDR: no such device try it again Configure network fails NFS mount: /mnt/runtime cannot mount exiting See the documentation for this trouble" What could it be? I´ve used the GigEthernet-Driver(e1000, bcm5700) (the Blades have a Broadcom NetXtreme Gigabit Ethernet Interface) If anybody could help, i would be absolutely happy about!!! Thanxx a lot, Andre _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From M.Newiger ＠ deltacomputer.de Thu Sep 4 05:07:10 2003 From: M.Newiger ＠ deltacomputer.de (Martin Newiger) Date: Wed, 3 Sep 2003 22:07:10 +0200 Subject: [SCore-users-jp] AW: [SCore-users] Some probs when installing IBM HS20 Blades (Bladecenter) Linux RedHat 7.3 engl. Message-ID: Hi André, I haven't had any problems with the installation of SCore 5.4. I know that it supports the Intel 82546EB NIC without any problems. May I ask you which one of your NICs is eth0? SCore expects to be installed from eth0 as far as I know. If your Broadcom NIS is at eth0 then you must put it on eth1 or whatever. Regards Martin Newiger >-----Ursprüngliche Nachricht----- >Von: André Motzkus [SMTP:Motzkus ＠ ifr.de] >Gesendet am: Mittwoch, 3. September 2003 20:18 >An: 'SCore-users ＠ pccluster.org' >Betreff: [SCore-users] Some probs when installing IBM HS20 Blades >(Bladecenter) Linux RedHat 7.3 engl. > >Hello @ all, > >i join your Score 5.4.0 by using the EIT. >(I was creating the boot floppy (using Read Hat 7.3). >When i come to the point for Installing the cluster hosts, >i´ve got some trouble. > >when i was trying to boot the first HS20 Blade, it stops with >the following message: > >"No dhcp server specified. Used broadcast >SIOCSIFADDR: no such device >try it again >SIOCSIFADDR: no such device >try it again >SIOCSIFADDR: no such device >try it again >Configure network fails >NFS mount: /mnt/runtime >cannot mount >exiting >See the documentation for this trouble" > >What could it be? > >I´ve used the GigEthernet-Driver(e1000, bcm5700) >(the Blades have a Broadcom NetXtreme Gigabit Ethernet Interface) > >If anybody could help, i would be absolutely happy about!!! > >Thanxx a lot, > >Andre >_______________________________________________ >SCore-users mailing list >SCore-users ＠ pccluster.org >http://www.pccluster.org/mailman/listinfo/score-users _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From kameyama ＠ pccluster.org Thu Sep 4 09:14:50 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Thu, 04 Sep 2003 09:14:50 +0900 Subject: [SCore-users-jp] Re: [SCore-users] Some probs when installing IBM HS20 Blades (Bladecenter) Linux Re dHat 7.3 engl. In-Reply-To: Your message of "Wed, 03 Sep 2003 20:18:17 JST." <76970123516DD611B30300105A66B2A80AA33F@IFRSERVER01> Message-ID: <20030904000812.586C1128944@neal.il.is.s.u-tokyo.ac.jp> In article <76970123516DD611B30300105A66B2A80AA33F ＠ IFRSERVER01> =?iso-8859-1?Q?Andr=E9_Motzkus?= wrotes: > (I was creating the boot floppy (using Read Hat 7.3). > When i come to the point for Installing the cluster hosts, > ive got some trouble. > > when i was trying to boot the first HS20 Blade, it stops with > the following message: > > "No dhcp server specified. Used broadcast > SIOCSIFADDR: no such device I think SCore boot floppy dose not find any ethernet device. > Ive used the GigEthernet-Driver(e1000, bcm5700) > (the Blades have a Broadcom NetXtreme Gigabit Ethernet Interface) Do you select "1Gbps_Ethernet" when Select Boot Network Device window? Please look at following page arroun figure 14: http://www.pccluster.org/score/dist/score/html/en/installation/eit.html from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From hagiwaya ＠ pst.fujitsu.com Thu Sep 4 13:41:02 2003 From: hagiwaya ＠ pst.fujitsu.com (hagi) Date: Thu, 04 Sep 2003 13:41:02 +0900 Subject: [SCore-users-jp] SCoreのlibpthread Message-ID: <20030904134033.9EE8.HAGIWAYA@pst.fujitsu.com> 富士通プライムソフトテクノロジの萩原です。 いつもお世話になっております。 現在、SCore上でテストプログラムを作成しているのですが、 -lpthreadを指定してコンパイルした場合に/proc//*の ファイルにアクセスできません。 RedHat 7.3 + SCore5.4.0とRedHat 7.3 + SCore5.2.0で 発生することを確認しました。 SCore5.2.0の場合は、GNUコンパイラ・富士通コンパイラの 両方で試しましたがどちらを使用しても発生しました。 調べてみると、-lpthreadを指定した場合は、/proc//* のファイルのオーナがrootになっていました。 -lpthreadを指定しない場合は、一般ユーザのままです。 下記プログラムはプロセスのuid,pid を表示して、 /proc//exe をオープンしてみるだけの簡単なものです。 --a.cc---------------------------------- #include #include #include #include #include #include #include #include main(int argc, char *argv[]) { char procdir[256]; FILE *exe; mpcxx_spmd_initialize(argc, argv); printf("After initialize : getuid=%d\n", getuid()); printf(" : geteuid=%d\n", geteuid()); printf(" : getpid=%d\n", getpid()); sprintf(procdir, "/proc/%d/exe", getpid()); exe = fopen(procdir, "r"); if ( exe != NULL ) { printf("%s open success.\n", procdir); } else { printf("%s open failed. errno=%d\n", procdir, errno); } } ---------------------------------------- 上記プログラムを以下のように翻訳しますと正常にオープンできます。 $ mpc++ a.cc $ scrun -group=g2,nodes=2x1 ./a.out SCOUT: Spawning done. SCore-D 5.4.0 connected. <0:0> SCORE: 2 nodes (2x1) ready. After initialize : getuid=9999 : geteuid=9999 : getpid=18319 /proc/18319/exe open success. After initialize : getuid=9999 : geteuid=9999 : getpid=16706 /proc/16706/exe open success. SCOUT: Session done. ところが、これに SCore の libpthread.a をリンクするようにしますと、 $ mpc++ -o lpth a.cc -lpthread $ scrun -group=g2,nodes=2x1 ./lpth SCOUT: Spawning done. SCore-D 5.4.0 connected. <0:0> SCORE: 2 nodes (2x1) ready. After initialize : getuid=9999 : geteuid=9999 : getpid=18368 /proc/18368/exe open failed. errno=13 After initialize : getuid=9999 : geteuid=9999 : getpid=16755 /proc/16755/exe open failed. errno=13 SCOUT: Session done. このようにオープンに失敗します。 errno=13 (EACCES: Permission denied) なので 確認してみたとこと計算ノード上の/proc/18368/*と /proc/16755/* オーナが root になっていました。 同じプログラムを1ノードで実行すると、 $ scrun -group=g2,nodes=1 ./lpth SCOUT: Spawning done. SCore-D 5.4.0 connected. <0:0> SCORE: One node ready. After initialize : getuid=9999 : geteuid=9999 : getpid=1888 /proc/1888/exe open success. SCOUT: Session done. 正常に動作します。 コンパイル時にチェックポイント機能をはずしてみたところ、 なぜかこちらも正常に動作しました。 $ mpc++ -o nockpt -nockpt -lpthread a.cc $ scrun -group=g2,nodes=2x1 ./nockpt SCOUT: Spawning done. SCore-D 5.4.0 connected. <0:0> SCORE: 2 nodes (2x1) ready. After initialize : getuid=9999 : geteuid=9999 : getpid=4987 /proc/4987/exe open success. After initialize : getuid=9999 : geteuid=9999 : getpid=1688 /proc/1688/exe open success. SCOUT: Session done. コンパイル時にnockptオプションをつけた場合、 SCoreのチェックポイント機能が使用できないこと 以外になにか制限がつくのでしょうか。 SCore の libpthread をリンクすると/proc//* の ファイルのオーナが root になってしまうという現象は SCore のバグ、もしくは仕様なのでしょうか。 これを直していただくことはできないでしょうか。 以上、よろしくお願いいたします。 -- hagi From ersoz ＠ cse.psu.edu Sat Sep 6 02:24:02 2003 From: ersoz ＠ cse.psu.edu (Deniz Ersoz) Date: Fri, 5 Sep 2003 13:24:02 -0400 Subject: [SCore-users-jp] [SCore-users] SCore scheduling question Message-ID: <20030905172402.GA14511@titan.cse.psu.edu> Hi, Does anybody know if scored uses backfilling when scheduling jobs in multi-user mode? If yes, is there a way to disable this? Thanks, Deniz Ersoz _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From 607tpmtrlo ＠ yahoo.com Mon Sep 8 20:42:23 2003 From: 607tpmtrlo ＠ yahoo.com (Cara Woodson) Date: Mon, 08 Sep 03 11:42:23 GMT Subject: [SCore-users-jp] [SCore-users] Bigger tits equals better sex d mcyp Message-ID: HTMLの添付ファイルを保管しました... URL: From igor_makunin ＠ nightmail.ru Mon Sep 8 03:33:03 2003 From: igor_makunin ＠ nightmail.ru (=?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?=) Date: Sun, 7 Sep 2003 22:33:03 +0400 Subject: [SCore-users-jp] [SCore-users] SCI drivers from Dolphin Message-ID: <20030907223303.37c3cdf9.igor_makunin@nightmail.ru> Hi. Does the Omni OpenMP work only with Scali SCI drivers? I have SciOS/SciFS provided by Dolphins, and I'd like to use these drivers for computations. How can I do this? (I have SCore 5.4.0 installed from sources) Thanks. ---------- Igor Makunin _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From hori ＠ swimmy-soft.com Mon Sep 8 09:54:35 2003 From: hori ＠ swimmy-soft.com (Atsushi HORI) Date: Mon, 8 Sep 2003 09:54:35 +0900 Subject: [SCore-users-jp] Re: [SCore-users] SCore scheduling question In-Reply-To: <20030905172402.GA14511@titan.cse.psu.edu> References: <20030905172402.GA14511@titan.cse.psu.edu> Message-ID: <3145859675.hori0000@swimmy-soft.com> Hi, >Does anybody know if scored uses backfilling when scheduling jobs in >multi-user mode? SCore-D in multi-user mode schedules user jobs with gang scheduling by default. So, later submitted jobs scheduled immediately and this behavior looks like the backfilling. >If yes, is there a way to disable this? Use "exclusive scheduling mode" or add the scrun option "priority=0". ---- Atsushi HORI Swimmy Software, Inc. _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From msato ＠ is.tsukuba.ac.jp Mon Sep 8 11:48:04 2003 From: msato ＠ is.tsukuba.ac.jp (Mitsuhisa Sato) Date: Mon, 08 Sep 2003 11:48:04 +0900 (JST) Subject: [SCore-users-jp] Re: [SCore-users] SCI drivers from Dolphin In-Reply-To: <20030907223303.37c3cdf9.igor_makunin@nightmail.ru> References: <20030907223303.37c3cdf9.igor_makunin@nightmail.ru> Message-ID: <20030908.114804.730575297.msato@hpcs.is.tsukuba.ac.jp> Hi, From: Макунин Игорь Subject: [SCore-users] SCI drivers from Dolphin Date: Sun, 7 Sep 2003 22:33:03 +0400 > > Does the Omni OpenMP work only with Scali SCI drivers? No, the current version only supports PM-based platform. > I have SciOS/SciFS provided by Dolphins, and I'd like to use > these drivers for computations. How can I do this? Currently, Omni team is developing Omni/SCASH on MPI and TCP. it will be included the next release. Thanks. -- M. Sato. _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From kameyama ＠ pccluster.org Mon Sep 8 11:56:04 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Mon, 08 Sep 2003 11:56:04 +0900 Subject: [SCore-users-jp] Re: [SCore-users] SCI drivers from Dolphin In-Reply-To: Your message of "Sun, 07 Sep 2003 22:33:03 JST." <20030907223303.37c3cdf9.igor_makunin@nightmail.ru> Message-ID: <20030908024913.2216E128944@neal.il.is.s.u-tokyo.ac.jp> In article <20030907223303.37c3cdf9.igor_makunin ＠ nightmail.ru> =?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?= wrotes: > Does the Omni OpenMP work only with Scali SCI drivers? PM/SCI work only Scali SCI driver, because PM/SCI device driver use Scali SCI driver. from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From igor_makunin ＠ nightmail.ru Mon Sep 8 18:50:08 2003 From: igor_makunin ＠ nightmail.ru (=?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?=) Date: Mon, 8 Sep 2003 13:50:08 +0400 Subject: [SCore-users-jp] Re: [SCore-users] SCI drivers from Dolphin In-Reply-To: <20030908.114804.730575297.msato@hpcs.is.tsukuba.ac.jp> References: <20030907223303.37c3cdf9.igor_makunin@nightmail.ru> <20030908.114804.730575297.msato@hpcs.is.tsukuba.ac.jp> Message-ID: <20030908135008.292e9be1.igor_makunin@nightmail.ru> Hi. > Currently, Omni team is developing Omni/SCASH on MPI and TCP. Well, does Omni/SCASH work with SCI? Does SCASH a software DSM only? Thanks. ------ Igor Makunin. _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From kameyama ＠ pccluster.org Mon Sep 8 18:57:41 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Mon, 08 Sep 2003 18:57:41 +0900 Subject: [SCore-users-jp] Re: [SCore-users] SCI drivers from Dolphin In-Reply-To: Your message of "Mon, 08 Sep 2003 13:50:08 JST." <20030908135008.292e9be1.igor_makunin@nightmail.ru> Message-ID: <20030908095049.6F917128944@neal.il.is.s.u-tokyo.ac.jp> In article <20030908135008.292e9be1.igor_makunin ＠ nightmail.ru> =?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?= wrotes: > > > Currently, Omni team is developing Omni/SCASH on MPI and TCP. > > Well, does Omni/SCASH work with SCI? Does SCASH a software DSM only? PM/SCI support Omni OpenMP but single user mode only. http://www.pccluster.org/score/dist/score/html/en/installation/pmsci.html SCASH is A software DSM. http://www.pccluster.org/score/dist/score/html/en/overview/softarch.html > from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From PH1303 ＠ gmx.de Mon Sep 8 21:36:44 2003 From: PH1303 ＠ gmx.de (=?iso-2022-jp?b?UEgxMzAzIBskQiF3GyhCIGdteC5kZQ==?=) Date: Mon, 8 Sep 2003 14:36:44 +0200 (MEST) Subject: [SCore-users-jp] [SCore-users] PBS and Multi-User Environment Message-ID: <25640.1063024604@www20.gmx.net> Hi, I'm running a Multi-User-Environment with sc_watch (Version 5.4.0) but submitting jobs via qsub-command doesn't work. For example submitting a NAS parallel benchmark job via "qsub npb.sh" aborts with the message that no scout environment nor SCore-D server can be found. Starting the job via "mpirun -np 8 /benchmarks/NPB/mg.A.8" (without PBS) works fine. The SCORE_OPTIONS variable only contains the name of my scored server. What's wrong? Contents of npb.sh: ----------------------- mpirun -np 8 /benchmarks/NPB/mg.A.8 Contents of npb.sh. - stderr: -------------------------- No value for $TERM and no -T specified No value for $TERM and no -T specified FEP:ERROR Neither SCOUT environment nor SCore-D server. Patrick -- COMPUTERBILD 15/03: Premium-e-mail-Dienste im Test -------------------------------------------------- 1. GMX TopMail - Platz 1 und Testsieger! 2. GMX ProMail - Platz 2 und Preis-Qualitätssieger! 3. Arcor - 4. web.de - 5. T-Online - 6. freenet.de - 7. daybyday - 8. e-Post _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From lfhnllyupabr ＠ msn.com Mon Sep 8 22:57:11 2003 From: lfhnllyupabr ＠ msn.com (allison) Date: Mon, 8 Sep 2003 15:57:11 +0200 Subject: [SCore-users-jp] [SCore-users] Did you remember this place? Message-ID: <200309081356.h88DuVm19291@pccluster.org> HTMLの添付ファイルを保管しました... URL: From kameyama ＠ pccluster.org Tue Sep 9 08:55:39 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Tue, 09 Sep 2003 08:55:39 +0900 Subject: [SCore-users-jp] Re: [SCore-users] PBS and Multi-User Environment In-Reply-To: Your message of "Mon, 08 Sep 2003 14:36:44 JST." <25640.1063024604@www20.gmx.net> Message-ID: <20030908234845.B4C92128944@neal.il.is.s.u-tokyo.ac.jp> In article <25640.1063024604 ＠ www20.gmx.net> PH1303 ＠ gmx.de wrotes: > I'm running a Multi-User-Environment with sc_watch (Version 5.4.0) but > submitting jobs via qsub-command doesn't work. PBS dose not support multi user mode. PBS support single user mode only. from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From Motzkus ＠ ifr.de Tue Sep 9 21:14:00 2003 From: Motzkus ＠ ifr.de (=?iso-8859-1?Q?Andr=E9_Motzkus?=) Date: Tue, 9 Sep 2003 14:14:00 +0200 Subject: [SCore-users-jp] [SCore-users] Some probs when installing IBM HS20 Blades (Bladecenter) Linux Re dHat 7.3 engl. Message-ID: <76970123516DD611B30300105A66B2A80AA365@IFRSERVER01> > Hi, > first of all many thanx to Rene Storm, Martin Newiger and Kameyama Toyohisa for the quick response. I´ve tried to make a special boot floppy by using the doc´s about "Making EIT-Bootfloppy for unsupported NIC´s" So i include different bcm5700.o and t3.o Modules in the initrd.img But unfortunately without success! The IBM HS20 Bladecenter stops everytime with the message: "No dhcp server specified. Used broadcast SIOCSIFADDR: no such device try it again SIOCSIFADDR: no such device try it again SIOCSIFADDR: no such device try it again Configure network fails NFS mount: /mnt/runtime cannot mount exiting See the documentation for this trouble" So i consult an IBM technican, he told me that the Bladecenter are using usb-architecture for floppy and CD-Rom Support. Maybe the is the problem, why bootdisk hangs up? But i´ve no hope to create a well-working bootfloppy for the bladecenter :o( . So i search for an alternative clusternode-boot opportunity by using EIT. Have anybody an idea, how i can boot and connect to EIT-working Master-Server in any different way (than using boot floppy)? Can i configure the PXE function on "Broadcom BCM 5703 Tiga 3" NIC´s working on Blades ? Or maybe can i create an Bootpartition on the Blades IDE-HDD´s (using syslinux)? How should i modify the initrd.img? And (worst case workaround), if i install RedHat 7.3 native on every of the 42 Blades - maybe i can than connect to EIT-working Master-Server for taking rigth IP-address and the right Score-image? And what must i modify ? Last but not least... Has anybody some experiences with Score 5.4.0 (Kernel 2.4.19), ´cause my RedHat Linux has 2.4.18-3 . Can i built modules in Score Kernel by getting out of my RH 7.3? Thanxx a lot for every kind of response !!!!! Andre > i join your Score 5.4.0 by using the EIT. > (I was creating the boot floppy (using Read Hat 7.3). > When i come to the point for Installing the cluster hosts, > i´ve got some trouble. > > when i was trying to boot the first HS20 Blade, it stops with > the following message: > > "No dhcp server specified. Used broadcast > SIOCSIFADDR: no such device > try it again > SIOCSIFADDR: no such device > try it again > SIOCSIFADDR: no such device > try it again > Configure network fails > NFS mount: /mnt/runtime > cannot mount > exiting > See the documentation for this trouble" > > What could it be? > > I´ve used the GigEthernet-Driver(e1000, bcm5700) > (the Blades have a Broadcom NetXtreme Gigabit Ethernet Interface) > > If anybody could help, i would be absolutely happy about!!! > > Thanxx a lot, > > Andre _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From Motzkus ＠ ifr.de Tue Sep 9 21:42:31 2003 From: Motzkus ＠ ifr.de (=?iso-8859-1?Q?Andr=E9_Motzkus?=) Date: Tue, 9 Sep 2003 14:42:31 +0200 Subject: [SCore-users-jp] [SCore-users] Problems with EIT (Score 5.4.0) while installing IBM HS20 Blades (RedHat 7.3) Message-ID: <76970123516DD611B30300105A66B2A80AA366@IFRSERVER01> > Hi, > > first of all many thanx to Rene Storm, Martin Newiger and Kameyama > Toyohisa > for the quick response. > > I´ve tried to make a special boot floppy by using > the doc´s about "Making EIT-Bootfloppy for unsupported NIC´s" > > So i include different bcm5700.o and t3.o Modules in the initrd.img > But unfortunately without success! > > The IBM HS20 Bladecenter stops everytime with the message: > > "No dhcp server specified. Used broadcast > SIOCSIFADDR: no such device > try it again > SIOCSIFADDR: no such device > try it again > SIOCSIFADDR: no such device > try it again > Configure network fails > NFS mount: /mnt/runtime > cannot mount > exiting > See the documentation for this trouble" > > So i consult an IBM technican, he told me that the Bladecenter are > using usb-architecture > for floppy and CD-Rom Support. > Maybe the is the problem, why bootdisk hangs up? > > But i´ve no hope to create a well-working bootfloppy for the > bladecenter :o( . > > So i search for an alternative clusternode-boot opportunity by using > EIT. > > Have anybody an idea, how i can boot and connect to EIT-working > Master-Server > in any different way (than using boot floppy)? > > Can i configure the PXE function on "Broadcom BCM 5703 Tiga 3" NIC´s > working on Blades ? > > Or maybe can i create an Bootpartition on the Blades IDE-HDD´s (using > syslinux)? > How should i modify the initrd.img? > > And (worst case workaround), if i install RedHat 7.3 native on every > of the 42 Blades - maybe i can than > connect to EIT-working Master-Server for taking rigth IP-address and > the right Score-image? And > what must i modify ? > > Last but not least... > Has anybody some experiences with Score 5.4.0 (Kernel 2.4.19), ´cause > my RedHat Linux has 2.4.18-3 . > Can i built modules in Score Kernel by getting out of my RH 7.3? > > Thanxx a lot for every kind of response !!!!! > > Andre > > > > > > > > > > > > > > i join your Score 5.4.0 by using the EIT. > (I was creating the boot floppy (using Read Hat 7.3). > When i come to the point for Installing the cluster hosts, > i´ve got some trouble. > > when i was trying to boot the first HS20 Blade, it stops with > the following message: > > "No dhcp server specified. Used broadcast > SIOCSIFADDR: no such device > try it again > SIOCSIFADDR: no such device > try it again > SIOCSIFADDR: no such device > try it again > Configure network fails > NFS mount: /mnt/runtime > cannot mount > exiting > See the documentation for this trouble" > > What could it be? > > I´ve used the GigEthernet-Driver(e1000, bcm5700) > (the Blades have a Broadcom NetXtreme Gigabit Ethernet Interface) > > If anybody could help, i would be absolutely happy about!!! > > Thanxx a lot, > > Andre _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From bogdan.costescu ＠ iwr.uni-heidelberg.de Tue Sep 9 22:17:02 2003 From: bogdan.costescu ＠ iwr.uni-heidelberg.de (Bogdan Costescu) Date: Tue, 9 Sep 2003 15:17:02 +0200 (CEST) Subject: [SCore-users-jp] Re: [SCore-users] Some probs when installing IBM HS20 Blades (Bladecenter) Linux Re dHat 7.3 engl. In-Reply-To: <76970123516DD611B30300105A66B2A80AA365@IFRSERVER01> Message-ID: On Tue, 9 Sep 2003, [iso-8859-1] André Motzkus wrote: > So i include different bcm5700.o and t3.o Modules in the initrd.img > But unfortunately without success! It doesn't help if you do it blindly, especially if the modules were not matched with the kernel which probably led to module versioning mismatch. I suggest to install the RH 7.3 distribution standalone without SCore on one blade and check that the drivers that you are trying to use work; a standalone instalation can also be configured to use DHCP, so that you would get as close to the EIT node configuration as possible. > Maybe the is the problem, why bootdisk hangs up? USB preventing an on-board Ethernet from being recognized ? Very unlikely. > Can i configure the PXE function on "Broadcom BCM 5703 Tiga 3" NICŽs > working on Blades ? This would probably work, but you will be in the same situation as before after the kernel and modules load. > And (worst case workaround), if i install RedHat 7.3 native on every of > the 42 Blades ... You can use disk cloning to copy the disk after you have made one installation. You'll have to change the IP and name on each node afterwards. > ... - maybe i can than connect to EIT-working Master-Server for taking > rigth IP-address and the right Score-image? And what must i modify ? I think that you are complicating things too much, especially things that you don't understand well. If you manage to install RH 7.3 on each blade, I don't think that it makes sense to re-install SCore afterwards. You just have to follow the instructions for manual installation which will install the SCore bits on top of the existing RH 7.3. > Can i built modules in Score Kernel by getting out of my RH 7.3? I don't quite understand the question, but before messing up with kernels and modules, maybe you should do some reading, like the Linux Kernel FAQ. Precompiled kernels and modules might not work together well; it's probably best if the kernel and the modules come from the same source. -- Bogdan Costescu IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868 E-mail: Bogdan.Costescu ＠ IWR.Uni-Heidelberg.De _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From Motzkus ＠ ifr.de Tue Sep 9 22:37:23 2003 From: Motzkus ＠ ifr.de (=?iso-8859-1?Q?Andr=E9_Motzkus?=) Date: Tue, 9 Sep 2003 15:37:23 +0200 Subject: [SCore-users-jp] AW: [SCore-users] Some probs when installing IBM HS20 Blades (Bla decenter) Linux Re dHat 7.3 engl. Message-ID: <76970123516DD611B30300105A66B2A80AA368@IFRSERVER01> Hello Bogdan, thanks for your response. So i forgot to say, that i´m tried of course to use the right NIC driver Modules that match to the used kernel. The Prob here is, that nobody could say, if i should use the bcm5700.o or the tg3.o (both are Broadcom 5703 Tiga3 drivers) In this way i was installing a native RH 7.3 on a blade for checking the installed (well working)NIC-drivers. So i see that RH73 was installing the tg3.o driver. But my prob - focused on my question ´bout the Kernel 2.4.18 /2.4.19 - is : can i take the tg3.o module from the RH73 Source and put it on the Score Boot Image ? Does it mach? By the way... Yes you are right, i´m a newbie on this field, thatswhy i didn´t understand not all dependences as well. So i think its easier for me to use the EIT than the manual installation mode. @Cloning:yes, of course - with "worst case" i mean that way of installing not Disc-Jockeying ;-)... Regards, André _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From francisco.massetto ＠ poli.usp.br Wed Sep 10 05:01:54 2003 From: francisco.massetto ＠ poli.usp.br (Francisco Isidro Massetto) Date: Tue, 9 Sep 2003 17:01:54 -0300 Subject: [SCore-users-jp] [SCore-users] Problems with my first test using Score 5.4.0 Message-ID: <77E4B59F0FABF443A0E3344564C0FC2E5B834F@apl03.poli.usp.br> Dear all I've downloaded the Score 5.4.0 and followed the instructions about how to install it on a SMP machine (I have installed both server and compute packages - did I go wrong ?). During the installation, it was ok, but when I tried to run the 1st example (the mandel algorithm using MPICH), the following error was shown: # mpirun -np 4 /opt/score/demo/bin/pmandel <0> SCore-D:WARNING Unable to open PM ethernet/ethernet (error=19). <0> SCore-D:WARNING argv[0] -config <0> SCore-D:WARNING argv[1] /var/scored/scoreboard/sol0000V1001a20 <0> SCore-D:ERROR No PM device opened. Can anyone help me ?? Thanks a lot Best Regards Francisco Isidro Massetto Politechnic School - University of Sao Paulo / Brazil _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From bogdan.costescu ＠ iwr.uni-heidelberg.de Wed Sep 10 06:25:26 2003 From: bogdan.costescu ＠ iwr.uni-heidelberg.de (Bogdan Costescu) Date: Tue, 9 Sep 2003 23:25:26 +0200 (CEST) Subject: [SCore-users-jp] Re: AW: [SCore-users] Some probs when installing IBM HS20 Blades (Bla decenter) Linux Re dHat 7.3 engl. In-Reply-To: <76970123516DD611B30300105A66B2A80AA368@IFRSERVER01> Message-ID: On Tue, 9 Sep 2003, [iso-8859-1] André Motzkus wrote: > The Prob here is, that nobody could say, if > i should use the bcm5700.o or the tg3.o > (both are Broadcom 5703 Tiga3 drivers) From kameyama ＠ pccluster.org Wed Sep 10 09:30:53 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Wed, 10 Sep 2003 09:30:53 +0900 Subject: [SCore-users-jp] Re: [SCore-users] Problems with my first test using Score 5.4.0 In-Reply-To: Your message of "Tue, 09 Sep 2003 17:01:54 JST." <77E4B59F0FABF443A0E3344564C0FC2E5B834F@apl03.poli.usp.br> Message-ID: <20030910002356.519CD128944@neal.il.is.s.u-tokyo.ac.jp> In article <77E4B59F0FABF443A0E3344564C0FC2E5B834F ＠ apl03.poli.usp.br> "Francisco Isidro Massetto" wrotes: > During the installation, it was ok, but when I tried to run the 1st example ( > the mandel algorithm using MPICH), the following error was shown: > > # mpirun -np 4 /opt/score/demo/bin/pmandel > <0> SCore-D:WARNING Unable to open PM ethernet/ethernet (error=19). > <0> SCore-D:WARNING argv[0] -config > <0> SCore-D:WARNING argv[1] /var/scored/scoreboard/sol0000V1001a20 > <0> SCore-D:ERROR No PM device opened. errno 19 meens the device driver for PM/ethernet is not found. Do you change the kernel on the compute hosts? Do you execute following command on the compute hosts? /etc/rc.d/init.d/pm_ethernet start from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From domon_ay_8 ＠ yahoo.co.jp Fri Sep 12 16:27:41 2003 From: domon_ay_8 ＠ yahoo.co.jp (安田　　 陽) Date: Fri, 12 Sep 2003 16:27:41 +0900 (JST) Subject: [SCore-users-jp] OS サポート Message-ID: <20030912072741.71052.qmail@web2303.mail.yahoo.co.jp> --0-1219155293-1063351661=:70737 Content-Type: text/plain; charset=iso-2022-jp SCoreについてお聞きしたいことがあり、メールをさせて頂きました。 SCore5.4は下記ディストリビューションをサポートしている旨がHPに 書かれてありました。(Linux OS) 　・Red Hat 7.3 ・Red Hat 8 ・Red Hat Advanced Workstation 2.1 (IA64) Red Hat 9はサポートしてないのでしょうか？ 知識不足で説明不十分なところがあるかもしれませんが ご返答よろしくお願いします。 --------------------------------- Do You Yahoo!? Yahoo! BB is Broadband by Yahoo! http://bb.yahoo.co.jp/ --0-1219155293-1063351661=:70737 Content-Type: text/html; charset=iso-2022-jp
SCoreについてお聞きしたいことがあり、メールをさせて頂きました。
SCore5.4は下記ディストリビューションをサポートしている旨がHPに
書かれてありました。(Linux OS)
　・Red Hat 7.3
   ・Red Hat 8
   ・Red Hat Advanced Workstation 2.1 (IA64)
Red Hat 9はサポートしてないのでしょうか？
知識不足で説明不十分なところがあるかもしれませんが
ご返答よろしくお願いします。


Do You Yahoo!?
Yahoo! BB is Broadband by Yahoo!
http://bb.yahoo.co.jp/ --0-1219155293-1063351661=:70737-- From kameyama ＠ pccluster.org Fri Sep 12 16:49:53 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Fri, 12 Sep 2003 16:49:53 +0900 Subject: [SCore-users-jp] OS サポート In-Reply-To: Your message of "Fri, 12 Sep 2003 16:27:41 JST." <20030912072741.71052.qmail@web2303.mail.yahoo.co.jp> Message-ID: <20030912074248.CE536128944@neal.il.is.s.u-tokyo.ac.jp> 亀山です. In article <20030912072741.71052.qmail ＠ web2303.mail.yahoo.co.jp> 安田 陽 wrotes: > SCore5.4は下記ディストリビューションをサポートしている旨がHPに > 書かれてありました。(Linux OS) > 　・Red Hat 7.3 > ・Red Hat 8 > ・Red Hat Advanced Workstation 2.1 (IA64) > Red Hat 9はサポートしてないのでしょうか？ サポートしていません. SCore 5.4 のリリース 2003/2/20 Red Hat 9 のリリース 2003/4/11 というリリース順序なんですが... from Kameyama Toyohisa From Motzkus ＠ ifr.de Fri Sep 12 19:02:07 2003 From: Motzkus ＠ ifr.de (=?iso-8859-1?Q?Andr=E9_Motzkus?=) Date: Fri, 12 Sep 2003 12:02:07 +0200 Subject: [SCore-users-jp] AW: AW: [SCore-users] Some probs when installing IBM HS20 Blades (Bla decenter) Linux Re dHat 7.3 engl. -Problem fixed!!! Message-ID: <76970123516DD611B30300105A66B2A80B584C@IFRSERVER01> Hi @ all, thanxx to anybody was trying to help fixing my HS20 Blade problem. The solution: add the following lines in the pcitable oft 1Gig-Ethernet Node Bootdisk: 0x14e4 0x16a7 "bcm5700" "Broadcom Corporation|NetXtreme BCM5703 Gigabit Ethernet" 0x14e4 0x16c7 "bcm5700" "Broadcom Corporation|NetXtreme BCM5703 Gigabit Ethernet" Thats the trick ! After change, it ´ll works fine! No new bcm5700.o Driver, no modification of the mod-info needed! So the EIT also works with the newer generation of broadcom´s 5703 nic´s Dear SCore team, I would ask you to add this lines in your floppy boot image, so it´s more compatible with braodcom´s nic´s.... But another prob, maybe it´s easy to fix: After installing all Node images on the HS20 Blade-clients, i was rebooting all clients and got the following error: ************************************************************************ ******************* -> MOUNTING NFS FILE System:Mount can´t get address for master01.ifr1.com [FAILED] ( e.g. :it´s the master server) -> Starting sshd Generating ssh1 RSA host key [FAILED] -> Starting console mouse service : keyboard timeout - AT-Keyboard not present?(f4) [FAILED] ************************************************************************ ******************* The screen shows "comp01 login:" But local keyboard doesn´t work, no input possible. Please don´t ask me using the right keyboard connector, of course i did ;o) Anybody an idea? Thanxx, André _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From igor_makunin ＠ nightmail.ru Sat Sep 13 20:46:57 2003 From: igor_makunin ＠ nightmail.ru (=?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?=) Date: Sat, 13 Sep 2003 15:46:57 +0400 Subject: [SCore-users-jp] [SCore-users] scali SCI Message-ID: <20030913154657.01602046.igor_makunin@nightmail.ru> Hi, I have installed ScaSCI.Linux2.i86pc-2.4.18.rpm ScaSCIadap.Linux2.i86pc-2.4.18.rpm ScaSCIddk.Linux2.i86pc-2.4.18.rpm and SCore 5.4.0 and when in the directory /opt/score/score-src/SCore/pm2/arch/sci/driver/linux I do 'make' I have the next error: gcc -D__KERNEL__ -I. -I../../include -I../../../../include -I/home/makunin/kernel-2.4.19-score/include -I/opt/scali/include -I/opt/scali/include/icm -I/opt/scali/kernel/rebuild/include -I/opt/scali/kernel/rebuild/Linux -Wall -Wstrict-prototypes -O2 -fomit-frame-pointer -fno-strict-aliasing -D__SMP__ -pipe -fno-strength-reduce -m486 -malign-loops=2 -malign-jumps=2 -malign-functions=2 -DCPU=686 -DMODULE -DEXPORT_SYMTAB -D__KERNEL__=1 -DLinux=1 -DUSE_ALLOC_FIX=1 -c pm_sci.c `-m486' is deprecated. Use `-march=i486' or `-mcpu=i486' instead. In file included from pm_sci.c:72: /home/makunin/kernel-2.4.19-score/include/linux/malloc.h:4:2: warning: #warning linux/malloc.h is deprecated, use linux/slab.h instead. pm_sci.c: In function `pm_sci_dev_getcsraddress': pm_sci.c:995: structure has no member named `csr_base' pm_sci.c:996: warning: control reaches end of non-void function pm_sci.c: At top level: pm_sci.c:29: warning: `rcsid' defined but not used make: *** [pm_sci.o] Error 1 what version of Scali SCI driver should I use? Thanks, Makunin Igor _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From joseph_antony ＠ yahoo.com Sat Sep 13 21:19:26 2003 From: joseph_antony ＠ yahoo.com (Joseph Antony) Date: Sat, 13 Sep 2003 05:19:26 -0700 (PDT) Subject: [SCore-users-jp] [SCore-users] SCash runtime error : Unsure how to correct Message-ID: <20030913121926.29943.qmail@web41602.mail.yahoo.com> Hello, I've setup SCore 5.4.0 on a Pentium III (dual) cluster with ethernet. Point to point tests for pm_ethernet and shmem work. MPI applications also run correctly. But I am unable to get SCash applications to work - [user ＠ master laplace.c]$ scrun -nodes=1x2 ./laplace_c SCore-D 5.4.0 connected. <0:0> SCORE: 2 nodes (1x2) ready. <0>scash_shmem_alloc:can not write map file: Operation not permitted[1] I have compiled the laplace_c application using --nocheckpoint. Hoping someone might know what the issue might be? Thanks, Joseph __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From kameyama ＠ pccluster.org Tue Sep 16 10:31:20 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Tue, 16 Sep 2003 10:31:20 +0900 Subject: [SCore-users-jp] Re: [SCore-users] SCash runtime error : Unsure how to correct In-Reply-To: Your message of "Sat, 13 Sep 2003 05:19:26 JST." <20030913121926.29943.qmail@web41602.mail.yahoo.com> Message-ID: <20030916012404.56A0612894C@neal.il.is.s.u-tokyo.ac.jp> In article <20030913121926.29943.qmail ＠ web41602.mail.yahoo.com> Joseph Antony wrotes: > But I am unable to get SCash applications to work - > [user ＠ master laplace.c]$ scrun -nodes=1x2 ./laplace_c > SCore-D 5.4.0 connected. > <0:0> SCORE: 2 nodes (1x2) ready. > <0>scash_shmem_alloc:can not write map file: Operation > not permitted[1] scash create temporary file under /var/scored for shared memory. But cannot resize this file enough. Please check your file size limiit. > from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From kameyama ＠ pccluster.org Tue Sep 16 11:02:59 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Tue, 16 Sep 2003 11:02:59 +0900 Subject: [SCore-users-jp] Re: [SCore-users] scali SCI In-Reply-To: Your message of "Sat, 13 Sep 2003 15:46:57 JST." <20030913154657.01602046.igor_makunin@nightmail.ru> Message-ID: <20030916015543.4D6BB12894C@neal.il.is.s.u-tokyo.ac.jp> In article <20030913154657.01602046.igor_makunin ＠ nightmail.ru> =?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?= wrotes: > I have installed > ScaSCI.Linux2.i86pc-2.4.18.rpm > ScaSCIadap.Linux2.i86pc-2.4.18.rpm > ScaSCIddk.Linux2.i86pc-2.4.18.rpm > and SCore 5.4.0 > > and when in the directory /opt/score/score-src/SCore/pm2/arch/sci/driver/linu > x > I do 'make' > I have the next error: > gcc -D__KERNEL__ -I. -I../../include -I../../../../include -I/home/maku > nin/kernel-2.4.19-score/include -I/opt/scali/include -I/opt/scali/include/icm > -I/opt/scali/kernel/rebuild/include -I/opt/scali/kernel/rebuild/Linux -Wall > -Wstrict-prototypes -O2 -fomit-frame-pointer -fno-strict-aliasing -D__SMP__ - > pipe -fno-strength-reduce -m486 -malign-loops=2 -malign-jumps=2 -malign-funct > ions=2 -DCPU=686 -DMODULE -DEXPORT_SYMTAB -D__KERNEL__=1 -DLinux=1 -DUSE_ALLO > C_FIX=1 -c pm_sci.c > `-m486' is deprecated. Use `-march=i486' or `-mcpu=i486' instead. > In file included from pm_sci.c:72: > /home/makunin/kernel-2.4.19-score/include/linux/malloc.h:4:2: warning: > #warning linux/malloc.h is deprecated, use linux/slab.h instead. > pm_sci.c: In function `pm_sci_dev_getcsraddress': > pm_sci.c:995: structure has no member named `csr_base' > pm_sci.c:996: warning: control reaches end of non-void function > pm_sci.c: At top level: > pm_sci.c:29: warning: `rcsid' defined but not used > make: *** [pm_sci.o] Error 1 > > what version of Scali SCI driver should I use? We compile with ScaSCI* 2.4.4. If you want to compile with 2.4.18, please change at 995 from: return (void *)(sc->padapter->card->csr_base); to return (void *)(&(sc->padapter->card->asadr[0])); Note that we cannot test PM/sci now because we don't have SCI board. from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From joseph_antony ＠ yahoo.com Tue Sep 16 16:34:40 2003 From: joseph_antony ＠ yahoo.com (Joseph Antony) Date: Tue, 16 Sep 2003 00:34:40 -0700 (PDT) Subject: [SCore-users-jp] [SCore-users] Omni OpenMP linker error : atomic_dec, atomic_inc Message-ID: <20030916073440.2943.qmail@web41613.mail.yahoo.com> Hello all, I've been trying to get the Omni OpenMP compiler to work on our cluster. I found the following compilation error and am wondering how it could be resolved? Thanks, Joseph PS I found a link to the Japanese section of the mailing list, but was unable to follow - http://www.pccluster.org/pipermail/score-users-jp/2002-December/001004.html ----------------- Problem - [user ＠ master laplace.omp]$ omcc -omniconfig=scash -o laplace_omp laplace_omp.c second.c -lm Compiling 'laplace_omp.c'... Compiling 'second.c'... /home/user/openmp_gcc_scash/lib/openmp/gcc/lib/openmp/lib/libompc_sm.a(ompsm_scash.o): In function `_ompsm_scash_lib_in': ompsm_scash.o(.text+0x129b): undefined reference to `atomic_inc' /home/user/openmp_gcc_scash/lib/openmp/gcc/lib/openmp/lib/libompc_sm.a(ompsm_scash.o): In function `_ompsm_scash_lib_out': ompsm_scash.o(.text+0x12c0): undefined reference to `atomic_dec' ompsm_scash.o(.text+0x1303): undefined reference to `atomic_inc' collect2: ld returned 1 exit status Error 1 __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From kameyama ＠ pccluster.org Tue Sep 16 16:47:52 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Tue, 16 Sep 2003 16:47:52 +0900 Subject: [SCore-users-jp] Re: [SCore-users] Omni OpenMP linker error : atomic_dec, atomic_inc In-Reply-To: Your message of "Tue, 16 Sep 2003 00:34:40 JST." <20030916073440.2943.qmail@web41613.mail.yahoo.com> Message-ID: <20030916074035.6BA7312894C@neal.il.is.s.u-tokyo.ac.jp> In article <20030916073440.2943.qmail ＠ web41613.mail.yahoo.com> Joseph Antony wrotes: > Hello all, > I've been trying to get the Omni OpenMP compiler to > work on our cluster. > > I found the following compilation error and am > wondering how it could be resolved? > PS I found a link to the Japanese section of the > mailing list, but was unable to follow - > http://www.pccluster.org/pipermail/score-users-jp/2002-December/001004.html atomic_inc() inline function defined on kernel header file before redhat 7.2: usr/include/asm/atomic.h On redhat 7.3, this function is not found in this file, (perhaerps redhat policy is changed.) if you want to compile SCore and OpenMP on redhat 7.3 or newer, you must install kernel source code on the host. If you installed kernel-source rpm, atomic_inc() is defined following file: usr/src/linux-2.4/include/asm/atomic.h And SCore 5.2.0 or later use the file. So please install kernel-source rpm on the host, and recompile SCore and OMNI again. from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From igor_makunin ＠ nightmail.ru Tue Sep 16 16:57:33 2003 From: igor_makunin ＠ nightmail.ru (=?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?=) Date: Tue, 16 Sep 2003 11:57:33 +0400 Subject: [SCore-users-jp] Re: [SCore-users] scali SCI In-Reply-To: <20030916015543.4D6BB12894C@neal.il.is.s.u-tokyo.ac.jp> References: <20030913154657.01602046.igor_makunin@nightmail.ru> <20030916015543.4D6BB12894C@neal.il.is.s.u-tokyo.ac.jp> Message-ID: <20030916115733.1ae7b532.igor_makunin@nightmail.ru> > In article <20030913154657.01602046.igor_makunin ＠ nightmail.ru> =?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?= wrotes: > > I have installed > > ScaSCI.Linux2.i86pc-2.4.18.rpm > > ScaSCIadap.Linux2.i86pc-2.4.18.rpm > > ScaSCIddk.Linux2.i86pc-2.4.18.rpm > > and SCore 5.4.0 > > > > and when in the directory /opt/score/score-src/SCore/pm2/arch/sci/driver/linu > > x > > I do 'make' > > I have the next error: > > gcc -D__KERNEL__ -I. -I../../include -I../../../../include -I/home/maku > > nin/kernel-2.4.19-score/include -I/opt/scali/include -I/opt/scali/include/icm > > -I/opt/scali/kernel/rebuild/include -I/opt/scali/kernel/rebuild/Linux -Wall > > -Wstrict-prototypes -O2 -fomit-frame-pointer -fno-strict-aliasing -D__SMP__ - > > pipe -fno-strength-reduce -m486 -malign-loops=2 -malign-jumps=2 -malign-funct > > ions=2 -DCPU=686 -DMODULE -DEXPORT_SYMTAB -D__KERNEL__=1 -DLinux=1 -DUSE_ALLO > > C_FIX=1 -c pm_sci.c > > `-m486' is deprecated. Use `-march=i486' or `-mcpu=i486' instead. > > In file included from pm_sci.c:72: > > /home/makunin/kernel-2.4.19-score/include/linux/malloc.h:4:2: warning: > > #warning linux/malloc.h is deprecated, use linux/slab.h instead. > > pm_sci.c: In function `pm_sci_dev_getcsraddress': > > pm_sci.c:995: structure has no member named `csr_base' > > pm_sci.c:996: warning: control reaches end of non-void function > > pm_sci.c: At top level: > > pm_sci.c:29: warning: `rcsid' defined but not used > > make: *** [pm_sci.o] Error 1 > > > > what version of Scali SCI driver should I use? > > We compile with ScaSCI* 2.4.4. > If you want to compile with 2.4.18, please change at 995 from: > return (void *)(sc->padapter->card->csr_base); > to > return (void *)(&(sc->padapter->card->asadr[0])); > > Note that we cannot test PM/sci now because we don't have SCI board. > > from Kameyama Toyohisa Hi, I've done this change, but now I have one more problem. I see the next message in /var/log/messages on compute nodes: PM/SCI[867]: this SCI Card is not support or test(unknown chip id(chip id = 0xa5d792a8)) I have Dolphine D335 SCI card. Thanks, Makunin Igor _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From kameyama ＠ pccluster.org Tue Sep 16 17:18:51 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Tue, 16 Sep 2003 17:18:51 +0900 Subject: [SCore-users-jp] Re: [SCore-users] scali SCI In-Reply-To: Your message of "Tue, 16 Sep 2003 11:57:33 JST." <20030916115733.1ae7b532.igor_makunin@nightmail.ru> Message-ID: <20030916081134.397F212894C@neal.il.is.s.u-tokyo.ac.jp> In article <20030916115733.1ae7b532.igor_makunin ＠ nightmail.ru> =?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?= wrotes: > Hi, > I've done this change, but now I have one more problem. > > I see the next message in /var/log/messages on compute nodes: > PM/SCI[867]: this SCI Card is not support or test(unknown chip id(chip > id = 0xa5d792a8)) > > I have Dolphine D335 SCI card. We tested Dolphine D330 SCI card only. This is message only. If PM/sci is worked, this is not probrem. from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From mika126 ＠ hotmail.com Thu Sep 18 21:20:23 2003 From: mika126 ＠ hotmail.com (chris kiam) Date: Thu, 18 Sep 2003 12:20:23 +0000 Subject: [SCore-users-jp] [SCore-users] SCORE 5.4 installation Message-ID: hi, im trying to install the Score 5.4,from the installation cd, on RedHat 7.3 and a cluster of one host and one combined server/computer host (both AMD Athlon pcs), but after configuring the host succesfuly, during server side setup i get the message: Cannot create /opt/score/etc/pm-ethernet.conf "/opt/score/etc/ndconf/pmethernetconf-skelton:1:13:0: No such file or directory" The same happened trying to install SCore 5.2 can anybody help? Best regards Kristian Kiamos dpt of Computer Architecture UPC, Barcelona, Spain _________________________________________________________________ The new MSN 8: advanced junk mail protection and 2 months FREE* http://join.msn.com/?page=features/junkmail _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From 086eood ＠ yahoo.com Fri Sep 19 15:11:20 2003 From: 086eood ＠ yahoo.com (Carlo Steward) Date: Fri, 19 Sep 03 06:11:20 GMT Subject: [SCore-users-jp] [SCore-users] how to have POWER erections Message-ID: HTMLの添付ファイルを保管しました... URL: From kameyama ＠ pccluster.org Fri Sep 19 09:42:09 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Fri, 19 Sep 2003 09:42:09 +0900 Subject: [SCore-users-jp] Re: [SCore-users] SCORE 5.4 installation In-Reply-To: Your message of "Thu, 18 Sep 2003 12:20:23 JST." Message-ID: <20030919003443.AAAC312894C@neal.il.is.s.u-tokyo.ac.jp> In article "chris kiam" wrotes: > hi, > im trying to install the Score 5.4,from the installation cd, on RedHat 7.3 > and a cluster of one host and one combined server/computer host (both AMD > Athlon pcs), but after configuring the host succesfuly, during server side > setup i get the message: > Cannot create /opt/score/etc/pm-ethernet.conf > "/opt/score/etc/ndconf/pmethernetconf-skelton:1:13:0: No such file or > directory" Please check disk space of /opt partition of the server. from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From ersoz ＠ cse.psu.edu Sat Sep 20 03:56:46 2003 From: ersoz ＠ cse.psu.edu (Deniz Ersoz) Date: Fri, 19 Sep 2003 14:56:46 -0400 Subject: [SCore-users-jp] [SCore-users] scrun wait option Message-ID: <20030919185646.GA26464@titan.cse.psu.edu> Is there a way to disable starting a job, when there ois another job witing for resources to become available. For example if there is 8 nodes available and a 16-node-job is waiting for available nodes, i don't want another 4-node job start running before the 16-node one. by default score let the 4-node job get ahead of the waiting line and runs it. is there a way to disable this? that is a strict FIFO order? Thanks, Deniz _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From igor_makunin ＠ nightmail.ru Sat Sep 20 06:08:51 2003 From: igor_makunin ＠ nightmail.ru (=?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?=) Date: Sat, 20 Sep 2003 01:08:51 +0400 Subject: [SCore-users-jp] [SCore-users] pm/sci test routines Message-ID: <20030920010851.07f45f4f.igor_makunin@nightmail.ru> Hi, How can I test PM/SCI ? There is no any record of PM/SCI tests on the page http://pccluster.org/faq/en/installation/pm-test.html. Thanks, Igor Makunin _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From joseph_antony ＠ yahoo.com Sat Sep 20 12:21:21 2003 From: joseph_antony ＠ yahoo.com (Joseph Antony) Date: Fri, 19 Sep 2003 20:21:21 -0700 (PDT) Subject: [SCore-users-jp] [SCore-users] Unable to run omp_microbenchmarks using Omni Message-ID: <20030920032121.4564.qmail@web41610.mail.yahoo.com> Hello All, I've been trying to run the omp_microbenchmarks (OpenMP benchmarks) from the EPCC. Has anyone been able to run these using SCore? The Laplace.omp example code that comes with SCore compiles and runs properly. This is what the error message I get - [user ＠ master omp_microbench_scash]$ scrun -nodes=4x2 ./syncbench SCore-D 5.4.0 connected. <0:0> SCORE: 8 nodes (4x2) ready. count=28, _ompsm_gdata_count=26 OMPSM FATAL: _ompsm_init_global_data failed I am not sure if this has to do with current limitations to the Omni OpenMP compiler. BTW Does anyone know what is the difference between -omniconfig=scash and -omniconfig=scash_smp? Cheers, Joseph __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From hori ＠ swimmy-soft.com Mon Sep 22 11:07:42 2003 From: hori ＠ swimmy-soft.com (Atsushi HORI) Date: Mon, 22 Sep 2003 11:07:42 +0900 Subject: [SCore-users-jp] Re: [SCore-users] scrun wait option In-Reply-To: <20030919185646.GA26464@titan.cse.psu.edu> References: <20030919185646.GA26464@titan.cse.psu.edu> Message-ID: <3147073662.hori0002@swimmy-soft.com> Hi, >Is there a way to disable starting a job, when there ois another job >witing for resources >to become available. > >For example if there is 8 nodes available and a 16-node-job is >waiting for available >nodes, i don't want another 4-node job start running before the 16-node one. > >by default score let the 4-node job get ahead of the waiting line >and runs it. is there a >way to disable this? that is a strict FIFO order? This sounds like that the 16-node job and the 4 node-job has a dependency in which the 16 node job must be executed before the 4-node job. In Unix and Linux, this kind of dependency problem is resolved by using the "make" command. How about invokeing those 2 jobs by using the make command ? ---- Atsushi HORI Swimmy Software, Inc. _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From kameyama ＠ pccluster.org Mon Sep 22 11:46:04 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Mon, 22 Sep 2003 11:46:04 +0900 Subject: [SCore-users-jp] Re: [SCore-users] pm/sci test routines In-Reply-To: Your message of "Sat, 20 Sep 2003 01:08:51 JST." <20030920010851.07f45f4f.igor_makunin@nightmail.ru> Message-ID: <20030922023828.4B16F12894C@neal.il.is.s.u-tokyo.ac.jp> In article <20030920010851.07f45f4f.igor_makunin ＠ nightmail.ru> =?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?= wrotes: > > Hi, > > How can I test PM/SCI ? Please try Omni OpenMP on SCASH Test in the following page: http://www.pccluster.org/score/dist/score/html/en/installation/pmsci.html If Omni OpenMP on SCASH Test is successfull, PM/SCI is OK, too. from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From a-hasegawa ＠ pb.jp.nec.com Mon Sep 22 16:25:03 2003 From: a-hasegawa ＠ pb.jp.nec.com (=?iso-2022-jp?b?YS1oYXNlZ2F3YSAbJEIhdxsoQiBwYi5qcC5uZWMuY29t?=) Date: Mon, 22 Sep 2003 16:25:03 +0900 Subject: [SCore-users-jp] [SCore-users] Unable to run omp_microbenchmarks using Omni In-Reply-To: <20030920032121.4564.qmail@web41610.mail.yahoo.com> References: <20030920032121.4564.qmail@web41610.mail.yahoo.com> Message-ID: <3F6EA3CF.6020302@pb.jp.nec.com> Hi Joseph > Hello All, > I've been trying to run the omp_microbenchmarks > (OpenMP benchmarks) > from the EPCC. Has anyone been able to run these using > SCore? > > The Laplace.omp example code that comes with SCore > compiles > and runs properly. > > This is what the error message I get - > [user ＠ master omp_microbench_scash]$ scrun -nodes=4x2 > ./syncbench > SCore-D 5.4.0 connected. > <0:0> SCORE: 8 nodes (4x2) ready. > count=28, _ompsm_gdata_count=26 > OMPSM FATAL: _ompsm_init_global_data failed > > I am not sure if this has to do with current > limitations to > the Omni OpenMP compiler. This message shows Omni/SCASH bug. Would you tell me a URL to get the benchmark source code ? I can't understard this problem from only this message. > BTW Does anyone know what is the difference between > -omniconfig=scash and -omniconfig=scash_smp? The scash_smp is obsolute. In this version, the scash_smp is mearged to the scash. ----- Atsushi Hasegawa NEC Infomatec systems -- ??? ?? E-Mail:a-hasega ＠ ats.nis.nec.co.jp ????NEC??????? ??????????? ???????? ??:03-3798-9991(Fax.03-3798-9198) / ??:8-115-2410(Fax.8-115-2419) _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From igor_makunin ＠ nightmail.ru Mon Sep 22 18:08:57 2003 From: igor_makunin ＠ nightmail.ru (=?KOI8-R?Q?=ED=C1=CB=D5=CE=C9=CE_=E9=C7=CF=D2=D8?=) Date: Mon, 22 Sep 2003 13:08:57 +0400 Subject: [SCore-users-jp] Re: [SCore-users] pm/sci test routines In-Reply-To: <20030922023828.4B16F12894C@neal.il.is.s.u-tokyo.ac.jp> References: <20030920010851.07f45f4f.igor_makunin@nightmail.ru> <20030922023828.4B16F12894C@neal.il.is.s.u-tokyo.ac.jp> Message-ID: <20030922130857.3150802e.igor_makunin@nightmail.ru> Hi, > > How can I test PM/SCI ? > > Please try Omni OpenMP on SCASH Test in the following page: > http://www.pccluster.org/score/dist/score/html/en/installation/pmsci.html > If Omni OpenMP on SCASH Test is successfull, PM/SCI is OK, too. I've tryed do it. But I had some problems, that's why I asked you about PM/SCI tests. If I try to run /opt/score/omni/lib/examples/scash-test-laplace I have the next error: [makunin ＠ cluster laplace]$ scrun -nodes=2x1,network=sci ./lap-omp SCore-D 5.4.0 connected. <0> SCORE-D:ERROR pmGetSendBuffer(dest=1,size=1388) timed out <0> ULT: Exception Signal (11) <0> SCORE: Program signaled (SIGTERM). [makunin ＠ cluster laplace]$ If necessary I can send you debug info (PM_DEBUG=5). Thanks, Igor Makunin. _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From augusto.gomes ＠ poli.usp.br Tue Sep 23 04:19:23 2003 From: augusto.gomes ＠ poli.usp.br (Augusto Mendes Gomes Junior) Date: Mon, 22 Sep 2003 16:19:23 -0300 Subject: [SCore-users-jp] [SCore-users] Problem with installation Score5.4 Message-ID: <77E4B59F0FABF443A0E3344564C0FC2E34EFC2@apl03.poli.usp.br> Hi, I'm having some problems with the installation of the Score 5.4 . I followed all the steps but when I start the Scoreboard daemon, the following message appears: root ＠ sol etc]# /etc/rc.d/init.d/scoreboard start SCOREBOARD: bind(OpenServerPort) Anyone can help me??? Best regards, Augusto Jr. _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From kameyama ＠ pccluster.org Wed Sep 24 09:32:37 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Wed, 24 Sep 2003 09:32:37 +0900 Subject: [SCore-users-jp] Re: [SCore-users] Problem with installation Score5.4 In-Reply-To: Your message of "Mon, 22 Sep 2003 16:19:23 JST." <77E4B59F0FABF443A0E3344564C0FC2E34EFC2@apl03.poli.usp.br> Message-ID: <20030924002455.3F45B12895C@neal.il.is.s.u-tokyo.ac.jp> In article <77E4B59F0FABF443A0E3344564C0FC2E34EFC2 ＠ apl03.poli.usp.br> "Augusto Mendes Gomes Junior" wrotes: > > root ＠ sol etc]# /etc/rc.d/init.d/scoreboard start > SCOREBOARD: bind(OpenServerPort) Please check scoreboard is running already: % /etc/rc.d/init.d/scoreboard status If the number is returned, scoreboard is running already. Otherwith probabry, other program ia used scoreboard port. scoreboard use tcp port 9950 by default. If yoy want to hange this port, please execute as following: 1. please edit /etc/rc.d/init.d/scoreboard scoreboard started function startsccoreboard() in this file. please add -port option to scoreboard. From su nobody -s /bin/sh -c "$INSTALL_ROOT/sbin/scoreboard $DEFECTS_ARG ... to su nobody -s /bin/sh -c "$INSTALL_ROOT/sbin/scoreboard -port XXXX $DEFECTS_ARG ... 2. please use SCBDSERV to host:port instead of host. from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From bhoward_10 ＠ gte.net Wed Sep 24 16:05:09 2003 From: bhoward_10 ＠ gte.net (guest) Date: Wed, 24 Sep 2003 10:05:09 +0300 Subject: [SCore-users-jp] [SCore-users] survey finds refinanced mortgages work! vcxkopzxzzzz Message-ID: <200309240704.h8O74gQ12455@pccluster.org> HTMLの添付ファイルを保管しました... URL: From hori ＠ swimmy-soft.com Wed Sep 24 16:31:07 2003 From: hori ＠ swimmy-soft.com (Atsushi HORI) Date: Wed, 24 Sep 2003 16:31:07 +0900 Subject: [SCore-users-jp] Re: [SCore-users] Problem with installation Score5.4 In-Reply-To: <77E4B59F0FABF443A0E3344564C0FC2E34EFC2@apl03.poli.usp.br> References: <77E4B59F0FABF443A0E3344564C0FC2E34EFC2@apl03.poli.usp.br> Message-ID: <3147265867.hori0003@swimmy-soft.com> Hi, >root ＠ sol etc]# /etc/rc.d/init.d/scoreboard start >SCOREBOARD: bind(OpenServerPort) Probably, another scoreboard server is already running. Or, the port 9950 is opened by some other daemon program. ---- Atsushi HORI Swimmy Software, Inc. _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From dsternkopf ＠ hpce.nec.com Wed Sep 24 19:23:34 2003 From: dsternkopf ＠ hpce.nec.com (Danny Sternkopf) Date: Wed, 24 Sep 2003 12:23:34 +0200 Subject: [SCore-users-jp] [SCore-users] PM device busy Message-ID: <20030924102334.GB8482@server.ess.nec.de> Hi, we have a 2-nodes SCore5.4 system based on RH7.3. We always have to restart pm_ethernet after a job completed. Otherwise the next job will fail. Here an example: [esi ＠ score0 esi]$ scrun -nodes=4,file=hf ./a.out SCOUT: Spawning done. SCore-D 5.4.0 connected. <0:0> SCORE: 4 nodes (2x2) ready. Process 0 of 4 on score0.cluster.org Process 2 of 4 on score1.cluster.org Process 3 of 4 on score1.cluster.org Process 1 of 4 on score0.cluster.org pi is approximately 3.1416009869231249, Error is 0.0000083333333318 wall clock time = 0.000866 SCOUT: Session done. [esi ＠ score0 esi]$ scrun -nodes=4,file=hf ./a.out SCOUT: Spawning done. <0> SCore-D:PM-PANIC(../network.c:331) pmBindChannel() (16:Device or resource busy) So it seems when a job has completed the PM device is still busy. But I don't know why. Do you have any ideas? Is there any configuration issue? Best regards, Danny Sternkopf _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From gchoi ＠ cse.psu.edu Thu Sep 25 05:36:44 2003 From: gchoi ＠ cse.psu.edu (Gyu Sang Choi) Date: Wed, 24 Sep 2003 16:36:44 -0400 Subject: [SCore-users-jp] [SCore-users] Backfilling in Score Message-ID: <3F72005C.7020809@cse.psu.edu> Hi, Now I am using SCORE in Linux Cluster and getting the results. I found one strange thing. Score looks like doing backfilling in SCORE. I didn't use PBS. So, doest Score support backfilling? Best Regards, -- --------------------------------------------------------------- | Gyu Sang Choi | | Office : 313 Pond Lab (863-3814) | | email : gchoi ＠ cse.psu.edu | | Tel : (814)861-6986 | | Address : 425 Waupelani Drive #327, State College, PA 16801 | --------------------------------------------------------------- _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From kameyama ＠ pccluster.org Thu Sep 25 09:29:02 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Thu, 25 Sep 2003 09:29:02 +0900 Subject: [SCore-users-jp] Re: [SCore-users] PM device busy In-Reply-To: Your message of "Wed, 24 Sep 2003 12:23:34 JST." <20030924102334.GB8482@server.ess.nec.de> Message-ID: <20030925002117.767B3128990@neal.il.is.s.u-tokyo.ac.jp> In article <20030924102334.GB8482 ＠ server.ess.nec.de> Danny Sternkopf wrotes: > [esi ＠ score0 esi]$ scrun -nodes=4,file=hf ./a.out > SCOUT: Spawning done. > SCore-D 5.4.0 connected. > <0:0> SCORE: 4 nodes (2x2) ready. > Process 0 of 4 on score0.cluster.org > Process 2 of 4 on score1.cluster.org > Process 3 of 4 on score1.cluster.org > Process 1 of 4 on score0.cluster.org > pi is approximately 3.1416009869231249, Error is 0.0000083333333318 > wall clock time = 0.000866 > SCOUT: Session done. > [esi ＠ score0 esi]$ scrun -nodes=4,file=hf ./a.out > SCOUT: Spawning done. > <0> SCore-D:PM-PANIC(../network.c:331) pmBindChannel() (16:Device or resource > busy) Please check privious program (a.out or scored) is still running on score0.cluster.org. If the job is running, please kill this program. from Kameyama Toyohisa _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From hori ＠ swimmy-soft.com Thu Sep 25 10:02:39 2003 From: hori ＠ swimmy-soft.com (Atsushi HORI) Date: Thu, 25 Sep 2003 10:02:39 +0900 Subject: [SCore-users-jp] Re: [SCore-users] Backfilling in Score In-Reply-To: <3F72005C.7020809@cse.psu.edu> References: <3F72005C.7020809@cse.psu.edu> Message-ID: <3147328959.hori0001@swimmy-soft.com> Hi, >Now I am using SCORE in Linux Cluster and getting the results. >I found one strange thing. >Score looks like doing backfilling in SCORE. >I didn't use PBS. >So, doest Score support backfilling? No, SCore does NOT support backfilling, but DOES support Gang Scheduling, which is a time sharing scheduling on a cluster. In backfilling and gang scheduling, the latter submitted job can be scheduled before the previous job terminates. But this can happen only when some conditions are met in backfilling. In gang scheduling, any jobs are scheduled when they are submitted, just because gang scheduling is a kinf of time scheduling. But of course, when multiple jobs are scheduled in gang scheduling, their execution speeds get slower according to the number of jobs simultaneously running. This scheduling behaviour is the same as in Unix or Linux. ---- Atsushi HORI Swimmy Software, Inc. _______________________________________________ SCore-users mailing list SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users From tsujita ＠ hiro.kindai.ac.jp Sat Sep 27 13:45:57 2003 From: tsujita ＠ hiro.kindai.ac.jp (Yuichi Tsujita) Date: Sat, 27 Sep 2003 13:45:57 +0900 (JST) Subject: [SCore-users-jp] SCore-Dのエラー Message-ID: <20030927.134557.730550109.tsujita@hiro.kindai.ac.jp> 近畿大学工学部の辻田といいます。 SCore-Dから出されるエラーの対処方で困っております。 現在、5台のDual XeonのPC（DELL PowerEdge1600SC）をギガビットの スイッチで繋ぎ、管理ノード1台+計算ノード4台の構成でSCore5.4を インストールしています。 EITによるインストールではどうもSCSI関連のトラブルでインストール できなかったため、PCクラスタコンソーシアムのWebサイトにある インストール方法に従い、各ノードにRedHat7.3をインストールした後に SCore5.4のrpmファイルを各ノードにダウンロードしてから インストールしました。 インストール作業を終え、必要な設定をした後、まずscoutが実行できる ことを確認しました。 次に、MPIのプログラム（既に他のSCoreクラスタで動作確認済み）を用いて scout -g half0 -e scrun -nodes=2x1 /home/tsujita/work/score/stampi-test/mpi-1-comm/mpi-1-comm -nm 1 -mode 0 のように動かしてみたところ、 SCOUT: Spawning done. <1> SCore-D:WARNING Unable to open a network configuration file (13): network='gigaethernet', attribute='-config:file' <1> SCore-D:ERROR No PM device opened. <0> SCore-D:WARNING Unable to open a network configuration file (13): network='gigaethernet', attribute='-config:file' <0> SCore-D:ERROR No PM device opened. のようなエラーメッセージが出てしまいました。 /opt/score/etc/にあるscorehosts.dbのgigaethernetの記述は gigaethernet type=ethernet \ -config:file=/opt/score/etc/pm-gigaethernet.conf となっており、指定したファイル /opt/score/etc/pm-gigaethernet.conf はきちんと存在していることを確認しました。 このファイルは、mkpmehernetconfにより作成しており、また各ノードの MACアドレスが正しいことも確認しております。 また/opt/score/etcにあるファイルは以下のようになっています。 drwxr-xr-x 2 root root 4096 Sep 25 11:48 compilers -rw-r--r-- 1 root root 203 Sep 26 17:34 hosts.equiv drwxr-xr-x 2 root root 4096 Sep 26 18:44 ndconf -rw-r--r-- 1 root root 25 Sep 26 17:30 pbs_server_name -rw-r--r-- 1 root root 454 Sep 27 11:05 pm-gigaethernet.conf -rw-r--r-- 1 root root 165 Sep 27 11:10 pm-udp.conf -rw-r--r-- 1 root root 2055 Sep 27 12:52 scorehosts.db -rw-r--r-- 1 root root 0 Feb 5 2003 scorehosts.defects -rw-r--r-- 1 root root 6 Feb 5 2003 version このような場合、どこを修正したらよろしいのでしょうか。 宜しくお願い致します。 辻田 祐一 (tsujita ＠ hiro.kindai.ac.jp) From kameyama ＠ pccluster.org Sun Sep 28 18:27:43 2003 From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=) Date: Sun, 28 Sep 2003 18:27:43 +0900 Subject: [SCore-users-jp] SCore-Dのエラー In-Reply-To: Your message of "Sat, 27 Sep 2003 13:45:57 JST." <20030927.134557.730550109.tsujita@hiro.kindai.ac.jp> Message-ID: <20030928091953.6959512894C@neal.il.is.s.u-tokyo.ac.jp> 亀山です. In article <20030927.134557.730550109.tsujita ＠ hiro.kindai.ac.jp> Yuichi Tsujita wrotes: > 次に、MPIのプログラム（既に他のSCoreクラスタで動作確認済み）を用いて > > scout -g half0 -e scrun -nodes=2x1 > /home/tsujita/work/score/stampi-test/mpi-1-comm/mpi-1-comm -nm 1 > -mode 0 > > のように動かしてみたところ、 > > SCOUT: Spawning done. > <1> SCore-D:WARNING Unable to open a network configuration file (13): > network='gigaethernet', attribute='-config:file' error number 13 (EACCES) ですので, config file のパーミッションの問題である可能性が高そうです. /opt/score/etc/pm-gigaethernet.conf のパーミッションを確認してみて ください. from Kameyama Toyohisa From x86lmgjpf ＠ hotmail.com Mon Sep 29 10:44:37 2003 From: x86lmgjpf ＠ hotmail.com (Alejandra Herrington) Date: Mon, 29 Sep 03 01:44:37 GMT Subject: [SCore-users-jp] [SCore-users] High Potenbt female attractant Message-ID: HTMLの添付ファイルを保管しました... URL: From bdass7 ＠ yahoo.com Mon Sep 29 00:54:03 2003 From: bdass7 ＠ yahoo.com (rje) Date: Sun, 28 Sep 2003 23:54:03 +0800 Subject: [SCore-users-jp] [SCore-users] Don't let him get away with cheating Message-ID: <200309281546.h8SFkvP14482@pccluster.org> HTMLの添付ファイルを保管しました... URL: From tsujita ＠ hiro.kindai.ac.jp Mon Sep 29 16:41:48 2003 From: tsujita ＠ hiro.kindai.ac.jp (Yuichi Tsujita) Date: Mon, 29 Sep 2003 16:41:48 +0900 (JST) Subject: [SCore-users-jp] SCore-Dのエラー In-Reply-To: <20030928091953.6959512894C@neal.il.is.s.u-tokyo.ac.jp> References: <20030927.134557.730550109.tsujita@hiro.kindai.ac.jp> <20030928091953.6959512894C@neal.il.is.s.u-tokyo.ac.jp> Message-ID: <20030929.164148.730558827.tsujita@hiro.kindai.ac.jp> 亀山様 近畿大学工学部の辻田です。 下記の件、御回答を頂きありがとうございました。 ご指摘のpm-gigaethernet.confを確認してみましたが、 所有者並びにパーミッションは、 -rw-r--r-- 1 root root 454 Sep 29 16:06 pm-gigaethernet.conf のようになっており、念の為、scoreboardを起動しなおして みましたが、以前と同じエラーが出ております。 ちなみにrootではエラーが出ることなく実行できます。 一般ユーザーでエラーが出るということは、上記の設定ではまずい のでしょうか。 宜しくお願い致します。 辻田 祐一 (tsujita ＠ hiro.kindai.ac.jp) kameyama> 亀山です. kameyama> kameyama> In article <20030927.134557.730550109.tsujita ＠ hiro.kindai.ac.jp> Yuichi Tsujita wrotes: kameyama> > 次に、MPIのプログラム（既に他のSCoreクラスタで動作確認済み）を用いて kameyama> > kameyama> > scout -g half0 -e scrun -nodes=2x1 kameyama> > /home/tsujita/work/score/stampi-test/mpi-1-comm/mpi-1-comm -nm 1 kameyama> > -mode 0 kameyama> > kameyama> > のように動かしてみたところ、 kameyama> > kameyama> > SCOUT: Spawning done. kameyama> > <1> SCore-D:WARNING Unable to open a network configuration file (13): kameyama> > network='gigaethernet', attribute='-config:file' kameyama> kameyama> error number 13 (EACCES) ですので, kameyama> config file のパーミッションの問題である可能性が高そうです. kameyama> kameyama> /opt/score/etc/pm-gigaethernet.conf のパーミッションを確認してみて kameyama> ください. kameyama> kameyama> kameyama> from Kameyama Toyohisa kameyama> From hori ＠ swimmy-soft.com Mon Sep 29 17:21:28 2003 From: hori ＠ swimmy-soft.com (Atsushi HORI) Date: Mon, 29 Sep 2003 17:21:28 +0900 Subject: [SCore-users-jp] SCore-Dのエラー In-Reply-To: <20030929.164148.730558827.tsujita@hiro.kindai.ac.jp> References: <20030927.134557.730550109.tsujita@hiro.kindai.ac.jp> Message-ID: <3147700888.hori0006@swimmy-soft.com> 堀@スイミー・ソフトウェアです． >ご指摘のpm-gigaethernet.confを確認してみましたが、 > >所有者並びにパーミッションは、 > >-rw-r--r-- 1 root root 454 Sep 29 16:06 pm-gigaethernet.conf > >のようになっており、念の為、scoreboardを起動しなおして >みましたが、以前と同じエラーが出ております。 >ちなみにrootではエラーが出ることなく実行できます。 >一般ユーザーでエラーが出るということは、上記の設定ではまずい >のでしょうか。 SCore ではサーバホストの /var/scored/etc/pm-gigaethernet.conf を計算ホ ストの /var/scored/scoreboard のディレクトリにキャッシュとしてコピーし ます．/var/scored/scoreboard のパーミッションをチェックしてみてくださ い． 注）/var/scored/scoreboard 以下のファイル名はオリジナルとは違う名前に なっていますので，驚かないでください． ---- 堀 敦史（ほり あつし） スイミー・ソフトウェア株式会社 From tsujita ＠ hiro.kindai.ac.jp Mon Sep 29 17:42:08 2003 From: tsujita ＠ hiro.kindai.ac.jp (Yuichi Tsujita) Date: Mon, 29 Sep 2003 17:42:08 +0900 (JST) Subject: [SCore-users-jp] SCore-Dのエラー In-Reply-To: <3147700888.hori0006@swimmy-soft.com> References: <20030927.134557.730550109.tsujita@hiro.kindai.ac.jp> <20030929.164148.730558827.tsujita@hiro.kindai.ac.jp> <3147700888.hori0006@swimmy-soft.com> Message-ID: <20030929.174208.730558873.tsujita@hiro.kindai.ac.jp> 堀様 近畿大学工学部の辻田です。 御指摘の通り、各計算ホストの /var/scoredのパーミッションを チェックしてみたところ、rootのみアクセス可能な状態に なっていた為、一般ユーザで利用できていないという状況でした。 御指摘の点は、PCクラスタコンソーシアムのSCoreのインストール 方法のページに記述がありましたが、私のミスでこの処理がなされて おりませんでした。申し訳ありませんでした。 一般ユーザからもアクセスできるように上記ディレクトリの パーミッションを変更したところ、エラーが出ることなく動作させる ことができました。 ありがとうございました。 辻田 祐一 (tsujita ＠ hiro.kindai.ac.jp) hori> 堀@スイミー・ソフトウェアです． hori> hori> >ご指摘のpm-gigaethernet.confを確認してみましたが、 hori> > hori> >所有者並びにパーミッションは、 hori> > hori> >-rw-r--r-- 1 root root 454 Sep 29 16:06 pm-gigaethernet.conf hori> > hori> >のようになっており、念の為、scoreboardを起動しなおして hori> >みましたが、以前と同じエラーが出ております。 hori> >ちなみにrootではエラーが出ることなく実行できます。 hori> >一般ユーザーでエラーが出るということは、上記の設定ではまずい hori> >のでしょうか。 hori> hori> SCore ではサーバホストの /var/scored/etc/pm-gigaethernet.conf を計算ホ hori> ストの /var/scored/scoreboard のディレクトリにキャッシュとしてコピーし hori> ます．/var/scored/scoreboard のパーミッションをチェックしてみてくださ hori> い． hori> hori> 注）/var/scored/scoreboard 以下のファイル名はオリジナルとは違う名前に hori> なっていますので，驚かないでください． hori> hori> ---- hori> 堀 敦史（ほり あつし） hori> スイミー・ソフトウェア株式会社 hori> hori> From vasgcqd ＠ msn.com Tue Sep 30 23:15:55 2003 From: vasgcqd ＠ msn.com (Jeanie Hansen) Date: Tue, 30 Sep 03 14:15:55 GMT Subject: [SCore-users-jp] [SCore-users] the best value penis pills Message-ID: <91m$$yw$t9-4wtxt60z10@f0c.3t> HTMLの添付ファイルを保管しました... URL: From 09hxhjx ＠ bigfoot.com Tue Sep 30 11:43:09 2003 From: 09hxhjx ＠ bigfoot.com (Brian Hyatt) Date: Tue, 30 Sep 03 02:43:09 GMT Subject: [SCore-users-jp] [SCore-users] I took these & my cock grew 3'!! juimuuuek Message-ID: HTMLの添付ファイルを保管しました... URL: