From nrcb ＠ streamline-computing.com  Sun Feb  2 03:17:25 2003
From: nrcb ＠ streamline-computing.com (Nick Birkett)
Date: Sat, 1 Feb 2003 18:17:25 +0000
Subject: [SCore-users-jp] [SCore-users] Copper Myrinet pm  problems
Message-ID: <200302011817.h11IHPK10399@zeralda.streamline.com>

Hi, we have just upgraded one of our older clusters to SCore 5.0.1 from 4.x
(I think it was the first SCore to support Myrinet 2000 from 18 months ago - 
RedHat 6.2 dist).

The cluster was working more or less ok under the old SCore system.

The entire system has been re-installed as RedHat 7.2 + SCore 5.0.1.

Hardware: Copper based Myrinet2k (May 2001) and Pentium III dual 866Mhz 
SuperMicro 1U Superservers.

I have run rpmtest and the scstest -network myrinet2k for many hours over all 
compute nodes without problems.

Have run gm1.6.3 codes (e.g PMB) and they work fine.

SCore PM codes are having problems over Myrinet - e.g  running PMB:

<6:0> SCORE:WARNING MPICH/SCore    [buffer=0x8951498, type=1025, from=11, 
size=262144, offset=189520]
<6:0> SCORE:WARNING MPICH/SCore: receive-message-queue:
<6:0> SCORE:WARNING MPICH/SCore    (empty)
<6:0> SCORE:WARNING MPICH/SCore: received-fragment:
<6:0> SCORE:WARNING MPICH/SCore    [buffer=0x40066180, type=1025, from=11, 
size=262144, fragment_size=8240, offset=189521]
<6:0> SCORE:WARNING MPICH/SCore: queued-message:
<6:0> SCORE:WARNING MPICH/SCore    [buffer=0x8951498, type=1025, from=11, 
size=262144, offset=189520]
<6:0> SCORE:WARNING MPICH/SCore: received an invalid fragment (mismatched 
offset)
<6:0> SCORE:PANIC MPICH/SCore: critical error on message transfer
<6:0> Trying to attach GDB (DISPLAY=localhost:10.0): PANIC
SCORE: Program aborted.
SCOUT: Session done.


Lots of buffer mismatch errors. The same binary runs fine over ethernet or 
gigabit on the same hardware (i.e if add the -network=ethernet option then
all ok so it is a Myrinet problem).

We would like to keep SCore as the cluster has some new Xeon Gigabit nodes
but will have to convert to GM if we cannot resolve this.

Looks like a hardware problem (same code runs fine over Score 5.0.1 and fibre
optic Myrinet 2k on Intel Xeon systems).

Thanks,

Nick









 
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From s-sumi ＠ bd6.so-net.ne.jp  Sun Feb  2 09:30:18 2003
From: s-sumi ＠ bd6.so-net.ne.jp (Shinji Sumimoto)
Date: Sun, 02 Feb 2003 09:30:18 +0900 (JST)
Subject: [SCore-users-jp] Re: [SCore-users] Copper Myrinet pm problems
In-Reply-To: <200302011817.h11IHPK10399@zeralda.streamline.com>
References: <200302011817.h11IHPK10399@zeralda.streamline.com>
Message-ID: <20030202.093018.730554912.s-sumi@bd6.so-net.ne.jp>

Hi.

Could you check CRC errors of the cluster nodes?

Shinji.

From: Nick Birkett <nrcb ＠ streamline-computing.com>
Subject: [SCore-users] Copper Myrinet pm problems
Date: Sat, 1 Feb 2003 18:17:25 +0000
Message-ID: <200302011817.h11IHPK10399 ＠ zeralda.streamline.com>

nrcb> Hi, we have just upgraded one of our older clusters to SCore 5.0.1 from 4.x
nrcb> (I think it was the first SCore to support Myrinet 2000 from 18 months ago - 
nrcb> RedHat 6.2 dist).
nrcb> 
nrcb> The cluster was working more or less ok under the old SCore system.
nrcb> 
nrcb> The entire system has been re-installed as RedHat 7.2 + SCore 5.0.1.
nrcb> 
nrcb> Hardware: Copper based Myrinet2k (May 2001) and Pentium III dual 866Mhz 
nrcb> SuperMicro 1U Superservers.
nrcb> 
nrcb> I have run rpmtest and the scstest -network myrinet2k for many hours over all 
nrcb> compute nodes without problems.
nrcb> 
nrcb> Have run gm1.6.3 codes (e.g PMB) and they work fine.
nrcb> 
nrcb> SCore PM codes are having problems over Myrinet - e.g  running PMB:
nrcb> 
nrcb> <6:0> SCORE:WARNING MPICH/SCore    [buffer=0x8951498, type=1025, from=11, 
nrcb> size=262144, offset=189520]
nrcb> <6:0> SCORE:WARNING MPICH/SCore: receive-message-queue:
nrcb> <6:0> SCORE:WARNING MPICH/SCore    (empty)
nrcb> <6:0> SCORE:WARNING MPICH/SCore: received-fragment:
nrcb> <6:0> SCORE:WARNING MPICH/SCore    [buffer=0x40066180, type=1025, from=11, 
nrcb> size=262144, fragment_size=8240, offset=189521]
nrcb> <6:0> SCORE:WARNING MPICH/SCore: queued-message:
nrcb> <6:0> SCORE:WARNING MPICH/SCore    [buffer=0x8951498, type=1025, from=11, 
nrcb> size=262144, offset=189520]
nrcb> <6:0> SCORE:WARNING MPICH/SCore: received an invalid fragment (mismatched 
nrcb> offset)
nrcb> <6:0> SCORE:PANIC MPICH/SCore: critical error on message transfer
nrcb> <6:0> Trying to attach GDB (DISPLAY=localhost:10.0): PANIC
nrcb> SCORE: Program aborted.
nrcb> SCOUT: Session done.
nrcb> 
nrcb> 
nrcb> Lots of buffer mismatch errors. The same binary runs fine over ethernet or 
nrcb> gigabit on the same hardware (i.e if add the -network=ethernet option then
nrcb> all ok so it is a Myrinet problem).
nrcb> 
nrcb> We would like to keep SCore as the cluster has some new Xeon Gigabit nodes
nrcb> but will have to convert to GM if we cannot resolve this.
nrcb> 
nrcb> Looks like a hardware problem (same code runs fine over Score 5.0.1 and fibre
nrcb> optic Myrinet 2k on Intel Xeon systems).
nrcb> 
nrcb> Thanks,
nrcb> 
nrcb> Nick
nrcb> 
nrcb> 
nrcb> 
nrcb> 
nrcb> 
nrcb> 
nrcb> 
nrcb> 
nrcb> 
nrcb>  
nrcb> _______________________________________________
nrcb> SCore-users mailing list
nrcb> SCore-users ＠ pccluster.org
nrcb> http://www.pccluster.org/mailman/listinfo/score-users
nrcb> 
-----
Shinji Sumimoto    E-Mail: s-sumi ＠ bd6.so-net.ne.jp
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From ishikawa ＠ is.s.u-tokyo.ac.jp  Mon Feb  3 09:53:15 2003
From: ishikawa ＠ is.s.u-tokyo.ac.jp (Yutaka Ishikawa)
Date: Mon, 03 Feb 2003 09:53:15 +0900 (JST)
Subject: [SCore-users-jp] 第2 回ＰＣクラスタシンポジウム開催のお知らせ
Message-ID: <20030203.095315.884014234.ishikawa@is.s.u-tokyo.ac.jp>

SCore Usersの皆様、

シンポジウム事前申し込みの締め切りが一週間後となりました。
参加を予定されている方で、まだ登録されていない方は、今すぐWEB登録のほど
お願い致します。
シンポジウムでは、
	SCoreの最新情報（SCore 5.4の配布）
	企業によるPCクラスタの最新の取り組み
などが無料で聴講できます。

石川＠PCクラスタコンソーシアム会長
＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝
 
　　　　　　 第2回PCクラスタシンポジウムのご案内
 

   構造解析、流体力学、分子動力学、ゲノム情報処理など、大規模計算を
   必要とする計算科学・計算化学分野では、研究機関を中心にPCクラスタ
   が利用されてきました。最近では、PCクラスタは産業界にも浸透し、例え
   ば、自動車産業界では、車体設計、衝突シミュレーションなどにPCクラス
   タが利用されています。
   PCクラスタコンソーシアム主催の第2回PCクラスタシンポジウムでは、
   会員企業による産業界におけるPCクラスタ導入事例の講演や企業展示
   など、PCクラスタに関する最新の動向をお伝えします。皆様の御参加を
   お待ちしております。
 
 ■開催日時 2003年2月20日(木)〜2003年2月21日(金)
   10:00〜17:00
 
 ■開催場所 日本科学未来館7階
 
   2003年2月20日(木):ワークショップ・・・みらいCANホール
   2003年2月21日(金):シンポジウム・・・・みらいCANホール
   併設企業展示：イノベーションホール
   なお、2月20日(木)のワークショップの参加対象は会員のみとなります。
 
 ■参加費 無料
 
 ■申込み方法  http://www.pccluster.org/ <http://www.pccluster.org/> 　より
お申込み下さい。
 
 ■定員  300名
 
 ■主催  PCクラスタコンソーシアム
 
   TEL:03-3263-6474, FAX:03-3263-7537
   e-mail:sec ＠ pccluster.org
   U R L :http://www.pccluster.org/
 
 ＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝


From terasawa ＠ nssnet.co.jp  Mon Feb  3 18:06:42 2003
From: terasawa ＠ nssnet.co.jp (Terasawa)
Date: Mon, 03 Feb 2003 18:06:42 +0900
Subject: [SCore-users-jp] PETScのインストール方法
In-Reply-To: <20021219042516.A994920046@neal.il.is.s.u-tokyo.ac.jp>
References: <JOELIPBDDCMFGIHFCNCNEELFCCAA.e973220@eve.u-ryukyu.ac.jp>
	<20021219042516.A994920046@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <200302030914.SAA08680@nss-ntsv4.nssnet.co.jp>

御世話に成ります。
寺沢＠ＮＳＳと申します。

以前のメーリングリストで、下記の記述がありました。

>SCore の library は static link 版のみしか提供していないため, SCore の
>関数/変数を使用している library が shared library になっていると
>うまく link できないようです...

因みに私共では、ＭＰＩで通信をする親のモジュールにリンクさせて使用する
計算モジュールを動的に切換えたいために、動的にロードされるライブラリ
(zzz.so)にしたいと考えます。
計算モジュールのライブラリ(zzz.so)の内部ではSCoreの関数もMPIも使ってい
ません。

SCoreの環境と矛盾するでしょうか？

システムの設計にかかわることですので心配しております。
何方かお教えいただければ幸です。

よろしくお願いします。
以上

-- 
Terasawa mailto:terasawa ＠ nssnet.co.jp


From kameyama ＠ pccluster.org  Tue Feb  4 09:07:00 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Tue, 04 Feb 2003 09:07:00 +0900
Subject: [SCore-users-jp] PETScのインストール方法
In-Reply-To: Your message of "Mon, 03 Feb 2003 18:06:42 JST."
             <200302030914.SAA08680@nss-ntsv4.nssnet.co.jp>
Message-ID: <20030204000700.2A17F2004F@neal.il.is.s.u-tokyo.ac.jp>

亀山です.

In article <200302030914.SAA08680 ＠ nss-ntsv4.nssnet.co.jp> Terasawa <terasawa ＠ nssnet.co.jp> wrotes:
> 因みに私共では、ＭＰＩで通信をする親のモジュールにリンクさせて使用する
> 計算モジュールを動的に切換えたいために、動的にロードされるライブラリ
> (zzz.so)にしたいと考えます。
> 計算モジュールのライブラリ(zzz.so)の内部ではSCoreの関数もMPIも使ってい
> ません。
> 
> SCoreの環境と矛盾するでしょうか？

(link できれば) 動くとは思いますが, 以下のことに注意してください.

1. SCore が提供する compile driver (mpicc など) の default は static link
   になっています.
   dynamic link を使用するときは, -nostatic オプションをつけてください.

2. dynamic link されている実行ファイルは checkpoint/restart ができません.

3. SCore プログラムは compute host で実行されるため, zzz.so は compute host
   にも install されている必要があります.
   実行ファイル自体は scrun/mpirun がコピーするので, compute host から
   見えている必要は無いのですが...

                       from Kameyama Toyohisa


From k-egg ＠ gmx.de  Tue Feb  4 23:45:17 2003
From: k-egg ＠ gmx.de (Andreas Vitz)
Date: Tue, 4 Feb 2003 15:45:17 +0100
Subject: [SCore-users-jp] [SCore-users] pooma
Message-ID: <200302041437.h14EbUS09209@pccluster.org>

Hello everybody,

Did anybody ever tried out pooma ( www.codesourcery.com ) with score ??

Is it possible at all ??


Yours,

Andreas Vitz
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From M.Newiger ＠ deltacomputer.de  Fri Feb  7 04:08:07 2003
From: M.Newiger ＠ deltacomputer.de (Martin Newiger)
Date: Thu, 6 Feb 2003 20:08:07 +0100
Subject: [SCore-users-jp] [SCore-users] Changing IPs after SCore-Installation
Message-ID: <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030206190807Z-1411@abakus.delnet>

Hi,

is it possible to change the IPs of the master and the nodes after a
successful SCore-Installation. I need to change them from routeable
(194.x.x.x) into non-routeable (192.168.x.x) Addresses. If it is
possible what files must I edit exactly?

>Regards
>Martin Newiger
>
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Fri Feb  7 09:49:09 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Fri, 07 Feb 2003 09:49:09 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Changing IPs after SCore-Installation
In-Reply-To: Your message of "Thu, 06 Feb 2003 20:08:07 JST."
             <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030206190807Z-1411@abakus.delnet>
Message-ID: <20030207004909.1A37820050@neal.il.is.s.u-tokyo.ac.jp>

In article <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030206190807Z-1411 ＠ abakus.delnet> Martin Newiger <M.Newiger ＠ deltacomputer.de> wrotes:
> is it possible to change the IPs of the master and the nodes after a
> successful SCore-Installation. I need to change them from routeable
> (194.x.x.x) into non-routeable (192.168.x.x) Addresses. If it is
> possible what files must I edit exactly?

SCore configuration files is not include IP address.
(Except PM/UDP and PM/Agent/UDP, but this is obsolated.)
So if you want to change IP address only, 
you do't need change SCoe configuration files.
(Probably you must change /etc/sysconfig/network and
/etc/sysconfig/network-scripts/ifcfg-* to change IP address.)

If you want to change hostname, you must change follwing SCore
configuration files on server host:
    /opt/score/etc/scorehosts.db
    /opt/score/etc/scorehosts.defects
    /opt/score/etc/pm-ethernet.conf
	 If you use PM/Ethernet
    /opt/score/etc/pm-myrinet.conf
	 If you use PM/Myrinet
   (If you use more PM network (for example, PM/ethernet trunking),
   you must change more files.)
    /etc/profile.d/score.*
	To change SCBDSERV environment variable
And you must change /etc/hosts.equiv and /root/.rhosts
on server host and all compute hosts.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From linux ＠ jcsn.co.jp  Fri Feb  7 15:59:02 2003
From: linux ＠ jcsn.co.jp (linux)
Date: Fri, 07 Feb 2003 15:59:02 +0900
Subject: [SCore-users-jp] SCBD
In-Reply-To: <20030127105356.160F.SUGAWARA@mlab.jks.ynu.ac.jp>
References: <20030127013533.38A3220050@neal.il.is.s.u-tokyo.ac.jp> <20030127105356.160F.SUGAWARA@mlab.jks.ynu.ac.jp>
Message-ID: <20030207154917.AEF9.LINUX@jcsn.co.jp>

はじめましてJCSの柴田と申します。

> 念のためもう一度Serverのインストール作業をしてみましたが、
> やはりうまく動きません。
> また、pm-ethernet.confの内容も1行目の「unit 0」のみになってしまい、
> 前回書き込んだものが消えてしまいました。
私も先日来クラスタを組んでいるのですが、一度もうまく行きません。
違う場所で組んでいるので細かいところを忘れていますが、
nd...の場所のファイルがうまく生成されてきません。
最後に追加した機械を忘れているようでサーバー機を計算用クラスタとして
組み込んでいるのですが必要なファイルが出来上がらないのでmesg系が立ち
上がらず時間がかかってしまうようです。

マシンはインテル製 2U Xeon 2.8GHz x2
メモリ 512MB
SCSI HDD 18G
Onboard Gigabit
RedHat 7.3 + SCore5.2
です。
以前違うシステムですが5.0.1を組んでいたときには問題なかったのですが
このシステムで組始めたらだめになってしまったようです。
ほかのシステムがないのでどれが悪いというのがわかりませんが、何かわか
りましたらまたmailします。

-- shiabta ＠ jcsn.co.jp --
-- linux ＠ jcsn.co.jp --


From emile.carcamo ＠ nec.fr  Mon Feb 10 23:17:15 2003
From: emile.carcamo ＠ nec.fr (Emile CARCAMO)
Date: Mon, 10 Feb 2003 15:17:15 +0100
Subject: [SCore-users-jp] [SCore-users] Intel 7.0 compiler support with mpich-1.2.4
Message-ID: <200302101417.h1AEHFmq009995@emilepc.ess.nec.fr>

Dear List Members,

	As far as I know, a new version of Score is gonna be released
	by end of this month. I was just wondering if PC-Cluster Con-
	sortium was going to provide us with a set of MPI libraries
	compiled with Intel ICC/IFC release 7.0 compilers ?? And what
	about Intel 7.0 compilers support ?? Do we need to recompile
	SCore or not ? Many thanks for your help and support. Regards,

-- 
Emile_CARCAMO         NEC High Performance            http://www.hpce.nec.com
System Engineer         Computing Europe         mailto:ecarcamo ＠ hpce.nec.com
(+33)6-8063-7003   GSM
(+33)1-3930-6601   FAX   / Your mouse has moved. Windows NT must be restarted \
(+33)1-3930-6613  PHONE  \ for the change to take effect. Reboot now?  [ OK ] /



_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From bogdan.costescu ＠ iwr.uni-heidelberg.de  Mon Feb 10 23:51:49 2003
From: bogdan.costescu ＠ iwr.uni-heidelberg.de (Bogdan Costescu)
Date: Mon, 10 Feb 2003 15:51:49 +0100 (CET)
Subject: [SCore-users-jp] Re: [SCore-users] Intel 7.0 compiler support with mpich-1.2.4
In-Reply-To: <200302101417.h1AEHFmq009995@emilepc.ess.nec.fr>
Message-ID: <Pine.LNX.4.44.0302101549570.5956-100000@kenzo.iwr.uni-heidelberg.de>

On Mon, 10 Feb 2003, Emile CARCAMO wrote:

> Do we need to recompile SCore or not ?

Well, I don't care about recompilation, provided that it actually works 
without the large amount of fixes/patches that were needed for the present 
version...

-- 
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 M.Newiger ＠ deltacomputer.de  Tue Feb 11 00:16:50 2003
From: M.Newiger ＠ deltacomputer.de (Martin Newiger)
Date: Mon, 10 Feb 2003 16:16:50 +0100
Subject: [SCore-users-jp] [SCore-users] No start
Message-ID: <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030210151650Z-44@abakus.delnet>

Hi,

when I try to start SCore with scout -g pcc it doesn't start. Any ideas
how I can solve this?

Regards 
M.Newiger
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From M.Newiger ＠ deltacomputer.de  Tue Feb 11 01:36:03 2003
From: M.Newiger ＠ deltacomputer.de (Martin Newiger)
Date: Mon, 10 Feb 2003 17:36:03 +0100
Subject: [SCore-users-jp] [SCore-users] (no subject)
Message-ID: <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030210163603Z-60@abakus.delnet>

Hi,

when I try to start SCore with scout -g pcc it doesn't start. After a
while a message saying permission denied appears. Any ideas how I can
solve this?

Regards 
M.Newiger


_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From s-sumi ＠ bd6.so-net.ne.jp  Tue Feb 11 09:50:16 2003
From: s-sumi ＠ bd6.so-net.ne.jp (Shinji Sumimoto)
Date: Tue, 11 Feb 2003 09:50:16 +0900 (JST)
Subject: [SCore-users-jp] [SCore-users] Intel 7.0 compiler support with
 mpich-1.2.4
In-Reply-To: <200302101417.h1AEHFmq009995@emilepc.ess.nec.fr>
References: <200302101417.h1AEHFmq009995@emilepc.ess.nec.fr>
Message-ID: <20030211.095016.730552515.s-sumi@bd6.so-net.ne.jp>

Hi.

Next SCore version 5.4 will also provide MPICH/SCore liblrary object
of intel compiler version 7 in addition to intel compiler version 6.

Shinji.

From: Emile CARCAMO <emile.carcamo ＠ nec.fr>
Subject: [SCore-users-jp] [SCore-users] Intel 7.0 compiler support with mpich-1.2.4
Date: Mon, 10 Feb 2003 15:17:15 +0100
Message-ID: <200302101417.h1AEHFmq009995 ＠ emilepc.ess.nec.fr>

emile.carcamo> 
emile.carcamo> Dear List Members,
emile.carcamo> 
emile.carcamo> 	As far as I know, a new version of Score is gonna be released
emile.carcamo> 	by end of this month. I was just wondering if PC-Cluster Con-
emile.carcamo> 	sortium was going to provide us with a set of MPI libraries
emile.carcamo> 	compiled with Intel ICC/IFC release 7.0 compilers ?? And what
emile.carcamo> 	about Intel 7.0 compilers support ?? Do we need to recompile
emile.carcamo> 	SCore or not ? Many thanks for your help and support. Regards,
emile.carcamo> 
emile.carcamo> -- 
emile.carcamo> Emile_CARCAMO         NEC High Performance            http://www.hpce.nec.com
emile.carcamo> System Engineer         Computing Europe         mailto:ecarcamo ＠ hpce.nec.com
emile.carcamo> (+33)6-8063-7003   GSM
emile.carcamo> (+33)1-3930-6601   FAX   / Your mouse has moved. Windows NT must be restarted \
emile.carcamo> (+33)1-3930-6613  PHONE  \ for the change to take effect. Reboot now?  [ OK ] /
emile.carcamo> 
emile.carcamo> 
emile.carcamo> 
emile.carcamo> _______________________________________________
emile.carcamo> SCore-users mailing list
emile.carcamo> SCore-users ＠ pccluster.org
emile.carcamo> http://www.pccluster.org/mailman/listinfo/score-users
emile.carcamo> _______________________________________________
emile.carcamo> SCore-users-jp mailing list
emile.carcamo> SCore-users-jp ＠ pccluster.org
emile.carcamo> http://www.pccluster.org/mailman/listinfo/score-users-jp
emile.carcamo> 
-----
Shinji Sumimoto    E-Mail: s-sumi ＠ bd6.so-net.ne.jp

_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From s-sumi ＠ bd6.so-net.ne.jp  Tue Feb 11 09:53:25 2003
From: s-sumi ＠ bd6.so-net.ne.jp (Shinji Sumimoto)
Date: Tue, 11 Feb 2003 09:53:25 +0900 (JST)
Subject: [SCore-users-jp] [SCore-users] No start
In-Reply-To: <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030210151650Z-44@abakus.delnet>
References: <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030210151650Z-44@abakus.delnet>
Message-ID: <20030211.095325.846938538.s-sumi@bd6.so-net.ne.jp>

Hi.

How did you install SCore, EIT or RPM? 

Or which hosts are you trying to run scout?

Shinji.

From: Martin Newiger <M.Newiger ＠ deltacomputer.de>
Subject: [SCore-users-jp] [SCore-users] No start
Date: Mon, 10 Feb 2003 16:16:50 +0100
Message-ID: <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030210151650Z-44 ＠ abakus.delnet>

M.Newiger> Hi,
M.Newiger> 
M.Newiger> when I try to start SCore with scout -g pcc it doesn't start. Any ideas
M.Newiger> how I can solve this?
M.Newiger> 
M.Newiger> Regards 
M.Newiger> M.Newiger
M.Newiger> _______________________________________________
M.Newiger> SCore-users mailing list
M.Newiger> SCore-users ＠ pccluster.org
M.Newiger> http://www.pccluster.org/mailman/listinfo/score-users
M.Newiger> _______________________________________________
M.Newiger> SCore-users-jp mailing list
M.Newiger> SCore-users-jp ＠ pccluster.org
M.Newiger> http://www.pccluster.org/mailman/listinfo/score-users-jp
M.Newiger> 
-----
Shinji Sumimoto    E-Mail: s-sumi ＠ bd6.so-net.ne.jp

_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From francois.courteille ＠ nec.fr  Tue Feb 11 18:41:47 2003
From: francois.courteille ＠ nec.fr (Francois Courteille)
Date: Tue, 11 Feb 2003 10:41:47 +0100
Subject: [SCore-users-jp] [SCore-users] Intel 7.0 compiler support with mpich-1.2.4
References: <200302101417.h1AEHFmq009995@emilepc.ess.nec.fr> <20030211.095016.730552515.s-sumi@bd6.so-net.ne.jp>
Message-ID: <004201c2d1b1$cfce5e20$03e874c1@VERSAFC>

Dear Shinji san,

thank you for the information.

I have 3other questions :

1/ Do Score 5.4 support Itanium2 (IA-64) and Linux IA-64 ?
How far Score has been tested on IA-64 plate-form ?

2/Is Quadrics supported by Score ?
   What about Infiniband ?

3/I would like to have your advice about the following:
    we were trying to read hdw performance counters.
 For that we tried to install the kernel patch perfctr-2.4.5.tar.gz  which
is required to use lperfex but we got an error message when trying to
generate a new kernel.
(See appendix below - apparently their is an irq usage conflict with score)

Have you an idea how to fix this problem ?


Thank you for your help,

With best regards,

Francois Courteille

APPENDIX

make -C perfctr
make[2]: Entering directory `/work/linux-2.4.18score/drivers/perfctr'
make all_targets
make[3]: Entering directory `/work/linux-2.4.18score/drivers/perfctr'
gcc -D__KERNEL__ -I/work/linux-2.4.18score/include  -Wall -Wstrict-prototype
s
-Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pi
pe
-mpreferred-stack-boundary=2 -march=i686   -DKBUILD_BASENAME=x86_setup
-DEXPORT_SYMTAB -c x86_setup.c
x86_setup.c: In function `set_cpus_allowed':
x86_setup.c:32: warning: implicit declaration of function `BUG_ON'
gcc -D__KERNEL__ -I/work/linux-2.4.18score/include  -Wall -Wstrict-prototype
s
-Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pi
pe
-mpreferred-stack-boundary=2 -march=i686   -DKBUILD_BASENAME=init  -c -o
init.o init.c
In file included from /work/linux-2.4.18score/include/asm/perfctr.h:147,
                 from /work/linux-2.4.18score/include/linux/perfctr.h:9,
                 from init.c:12:
/work/linux-2.4.18score/include/asm/hw_irq.h:228: parse error before
`irq_desc'
/work/linux-2.4.18score/include/asm/hw_irq.h:228: warning: type defaults to
`int' in declaration of `irq_desc'
/work/linux-2.4.18score/include/asm/hw_irq.h:228: warning: data definition
has
no type or storage class
make[3]: *** [init.o] Error 1
make[3]: Leaving directory `/work/linux-2.4.18score/drivers/perfctr'
make[2]: *** [first_rule] Error 2
make[2]: Leaving directory `/work/linux-2.4.18score/drivers/perfctr'
make[1]: *** [_subdir_perfctr] Error 2
make[1]: Leaving directory `/work/linux-2.4.18score/drivers'
make: *** [_dir_drivers] Error 2


_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From s-sumi ＠ bd6.so-net.ne.jp  Tue Feb 11 20:46:16 2003
From: s-sumi ＠ bd6.so-net.ne.jp (Shinji Sumimoto)
Date: Tue, 11 Feb 2003 20:46:16 +0900 (JST)
Subject: [SCore-users-jp] [SCore-users] Intel 7.0 compiler support with
 mpich-1.2.4
In-Reply-To: <004201c2d1b1$cfce5e20$03e874c1@VERSAFC>
References: <200302101417.h1AEHFmq009995@emilepc.ess.nec.fr>
	<20030211.095016.730552515.s-sumi@bd6.so-net.ne.jp>
	<004201c2d1b1$cfce5e20$03e874c1@VERSAFC>
Message-ID: <20030211.204616.607958486.s-sumi@bd6.so-net.ne.jp>

Hi.

From: "Francois Courteille" <francois.courteille ＠ nec.fr>
Subject: Re: [SCore-users-jp] [SCore-users] Intel 7.0 compiler support with mpich-1.2.4
Date: Tue, 11 Feb 2003 10:41:47 +0100
Message-ID: <004201c2d1b1$cfce5e20$03e874c1 ＠ VERSAFC>

francois.courteille> I have 3other questions :
francois.courteille> 
francois.courteille> 1/ Do Score 5.4 support Itanium2 (IA-64) and Linux IA-64 ?
francois.courteille> How far Score has been tested on IA-64 plate-form ?

Yes, SCore 5.4 supports Itanium2, however, we have only few nodes for
testing SCore on IA-64, so the test is limited.  We hope SCore users
try to use SCore 5.4 on Itanium2 and give us feedback about bugs. 

francois.courteille> 2/Is Quadrics supported by Score ?
francois.courteille>    What about Infiniband ?

We have no schedule of Quadrics support. 

Infiniband is scheduled, however we do not when it will be included
now.

francois.courteille> 3/I would like to have your advice about the following:
francois.courteille>     we were trying to read hdw performance counters.
francois.courteille>  For that we tried to install the kernel patch perfctr-2.4.5.tar.gz  which
francois.courteille> is required to use lperfex but we got an error message when trying to
francois.courteille> generate a new kernel.
francois.courteille> (See appendix below - apparently their is an irq usage conflict with score)
francois.courteille> 
francois.courteille> Have you an idea how to fix this problem ?

Is the patch is for original linux kernel, or some distibution? SCore
5.2 kernel is besed on 2.4.18, however some function such as ACPI is
added.

If the patch is for original linux kernel, we can provide minimum
patch for 2.4.18 kernel.

Please give him some advices > Kameyama-san.

Shinji.

francois.courteille> Thank you for your help,
francois.courteille> 
francois.courteille> With best regards,
francois.courteille> 
francois.courteille> Francois Courteille
francois.courteille> 
francois.courteille> APPENDIX
francois.courteille> 
francois.courteille> make -C perfctr
francois.courteille> make[2]: Entering directory `/work/linux-2.4.18score/drivers/perfctr'
francois.courteille> make all_targets
francois.courteille> make[3]: Entering directory `/work/linux-2.4.18score/drivers/perfctr'
francois.courteille> gcc -D__KERNEL__ -I/work/linux-2.4.18score/include  -Wall -Wstrict-prototype
francois.courteille> s
francois.courteille> -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pi
francois.courteille> pe
francois.courteille> -mpreferred-stack-boundary=2 -march=i686   -DKBUILD_BASENAME=x86_setup
francois.courteille> -DEXPORT_SYMTAB -c x86_setup.c
francois.courteille> x86_setup.c: In function `set_cpus_allowed':
francois.courteille> x86_setup.c:32: warning: implicit declaration of function `BUG_ON'
francois.courteille> gcc -D__KERNEL__ -I/work/linux-2.4.18score/include  -Wall -Wstrict-prototype
francois.courteille> s
francois.courteille> -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pi
francois.courteille> pe
francois.courteille> -mpreferred-stack-boundary=2 -march=i686   -DKBUILD_BASENAME=init  -c -o
francois.courteille> init.o init.c
francois.courteille> In file included from /work/linux-2.4.18score/include/asm/perfctr.h:147,
francois.courteille>                  from /work/linux-2.4.18score/include/linux/perfctr.h:9,
francois.courteille>                  from init.c:12:
francois.courteille> /work/linux-2.4.18score/include/asm/hw_irq.h:228: parse error before
francois.courteille> `irq_desc'
francois.courteille> /work/linux-2.4.18score/include/asm/hw_irq.h:228: warning: type defaults to
francois.courteille> `int' in declaration of `irq_desc'
francois.courteille> /work/linux-2.4.18score/include/asm/hw_irq.h:228: warning: data definition
francois.courteille> has
francois.courteille> no type or storage class
francois.courteille> make[3]: *** [init.o] Error 1
francois.courteille> make[3]: Leaving directory `/work/linux-2.4.18score/drivers/perfctr'
francois.courteille> make[2]: *** [first_rule] Error 2
francois.courteille> make[2]: Leaving directory `/work/linux-2.4.18score/drivers/perfctr'
francois.courteille> make[1]: *** [_subdir_perfctr] Error 2
francois.courteille> make[1]: Leaving directory `/work/linux-2.4.18score/drivers'
francois.courteille> make: *** [_dir_drivers] Error 2
francois.courteille> 

-----
Shinji Sumimoto    E-Mail: s-sumi ＠ bd6.so-net.ne.jp
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Wed Feb 12 09:24:59 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Wed, 12 Feb 2003 09:24:59 +0900
Subject: [SCore-users-jp] [SCore-users] Intel 7.0 compiler support with mpich-1.2.4
In-Reply-To: Your message of "Tue, 11 Feb 2003 20:46:16 JST."
             <20030211.204616.607958486.s-sumi@bd6.so-net.ne.jp>
Message-ID: <20030212002459.C3DA120050@neal.il.is.s.u-tokyo.ac.jp>

In article <20030211.204616.607958486.s-sumi ＠ bd6.so-net.ne.jp> Shinji Sumimoto <s-sumi ＠ bd6.so-net.ne.jp> wrotes:
> francois.courteille> 3/I would like to have your advice about the following:
> francois.courteille>     we were trying to read hdw performance counters.
> francois.courteille>  For that we tried to install the kernel patch perfctr-2
> .4.5.tar.gz  which
> francois.courteille> is required to use lperfex but we got an error message w
> hen trying to
> francois.courteille> generate a new kernel.
> francois.courteille> (See appendix below - apparently their is an irq usage c
> onflict with score)
> francois.courteille> 
> francois.courteille> Have you an idea how to fix this problem ?
> 
> Is the patch is for original linux kernel, or some distibution? SCore
> 5.2 kernel is besed on 2.4.18, however some function such as ACPI is
> added.

SCore 5.2 kernel is based on 2.4.18 + IA64 patch.
(ACPI is added by IA 64 patch)
So some files are changed for original kernel by IA64 patch,
some patch for IA32 original kernel is failed.
If you want to use perfctr on SCore 5.4,
please apply original 2.4.18 kernel and linux2.4.18_minimal.patch
(score.rpm directory in SCore 5.2 CD-ROM).

On SCore 5.4, we separate IA32 kernel and IA64 kernel.
The kernel patch will be provide two files, for IA32 and for IA64.
IA32 kernel based on original 2.4.19.
IA64 kernel based on original 2.4.19 + IA64 patch.
IA64 kernel patch include IA64 patch.
So you will be install perfctr patch on SCore 5.4 (IA32) kernel.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hori ＠ swimmy-soft.com  Wed Feb 12 09:31:57 2003
From: hori ＠ swimmy-soft.com (Atsushi HORI)
Date: Wed, 12 Feb 2003 09:31:57 +0900
Subject: [SCore-users-jp] Re: [SCore-users] (no subject)
In-Reply-To: <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030210163603Z-60@abakus.delnet>
References: <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030210163603Z-60@abakus.delnet>
Message-ID: <3127887117.hori0000@swimmy-soft.com>

Hi,

>when I try to start SCore with scout -g pcc it doesn't start. After a
>while a message saying permission denied appears. Any ideas how I can
>solve this?

Try rsh command from the server host to one of the cluster host. I 
guess you will have the same phenomenon. 

Check /etc/hosts, /etc/hosts.equiv and /root/.rhosts files.

----
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  Wed Feb 12 09:49:49 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Wed, 12 Feb 2003 09:49:49 +0900
Subject: [SCore-users-jp] [SCore-users] (no subject)
In-Reply-To: Your message of "Mon, 10 Feb 2003 17:36:03 JST."
             <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030210163603Z-60@abakus.delnet>
Message-ID: <20030212004949.BCBA520050@neal.il.is.s.u-tokyo.ac.jp>

In article <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030210163603Z-60 ＠ abakus.delnet> Martin Newiger <M.Newiger ＠ deltacomputer.de> wrotes:
> Hi,
> 
> when I try to start SCore with scout -g pcc it doesn't start. After a
> while a message saying permission denied appears. Any ideas how I can
> solve this?

Please check /etc/hosts.equiv (and /root/.rhosts if you execute to root)
on compute hosts.
If you have 4 compute hosts (comp0, comp2, comp3),
and you try to execute scout on server,
scout access to following order:
    server -> comp0 -> comp1 -> comp2 -> comp3
This means compX must be access to comp(X-1).
If you want to scout to comp0 and comp2, comp2 must be access to comp0...
At last all compute hosts must access to server and all other compute hosts.
So /etc/hosts.equiv on all compute hosts must be include all compute hosts
and server host (and if you want to execute other hosts, please add the hosts).

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From m-kawaguchi ＠ pst.fujitsu.com  Thu Feb 13 14:45:14 2003
From: m-kawaguchi ＠ pst.fujitsu.com (川口貢)
Date: Thu, 13 Feb 2003 14:45:14 +0900
Subject: [SCore-users-jp] scrunコマンド
Message-ID: <20030213144514E.m-kawaguchi@pst.fujitsu.com>

富士通プライムソフトテクノロジの川口と申します。
いつもお世話になっております。
scrunコマンドで以下のようなことを行いたいのですが、
可能かどうか教えて頂けますでしょうか?


- グループ名 "PC1"に4台のノードが登録されている環境で
  以下の通りジョブ実行させる。

   % scrun -group=PC1,nodes=2x1 ./a.out

  上記状態で、ジョブ実行されていない2ノード上で別ジョブを実行させたい。
  但し、この時の2ノードの割り当てを動的に行いたい。

   % scrun ?????,nodes=2x1 ./a2.out

静的には、ジョブ実行されていないノードを自分で特定し
scrun実行時にオプション指定(ex. -group=PC1~node1~node2)すれば
実現できると思います。
ですが、空きノードがどれかということを意識せずに
ジョブ実行を行いたいのですが可能でしょうか?
  ＃ sc_qsubコマンドは不可とさせて下さい。

以上、宜しくお願い致します。

---
川口  mail => m-kawaguchi ＠ pst.fujitsu.com


From kameyama ＠ pccluster.org  Thu Feb 13 18:22:28 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Thu, 13 Feb 2003 18:22:28 +0900
Subject: [SCore-users-jp] scrunコマンド
In-Reply-To: Your message of "Thu, 13 Feb 2003 14:45:14 JST."
             <20030213144514E.m-kawaguchi@pst.fujitsu.com>
Message-ID: <20030213092228.9760B2004F@neal.il.is.s.u-tokyo.ac.jp>

亀山です.

In article <20030213144514E.m-kawaguchi ＠ pst.fujitsu.com> 川口貢 <m-kawaguchi ＠ pst.fujitsu.com> wrotes:
> - グループ名 "PC1"に4台のノードが登録されている環境で
>   以下の通りジョブ実行させる。
> 
>    % scrun -group=PC1,nodes=2x1 ./a.out
> 
>   上記状態で、ジョブ実行されていない2ノード上で別ジョブを実行させたい。
>   但し、この時の2ノードの割り当てを動的に行いたい。
> 
>    % scrun ?????,nodes=2x1 ./a2.out
> 
> 静的には、ジョブ実行されていないノードを自分で特定し
> scrun実行時にオプション指定(ex. -group=PC1~node1~node2)すれば
> 実現できると思います。

最初の scrun で PC1 group 全部が占有されてしまうので,
最初の scrun が group 指定である限り実現できないと思いますが...

> ですが、空きノードがどれかということを意識せずに
> ジョブ実行を行いたいのですが可能でしょうか?
>   ＃ sc_qsubコマンドは不可とさせて下さい。

multi user mode を使用すればできますが, single user mode では
できないと思います.

                       from Kameyama Toyohisa


From hori ＠ swimmy-soft.com  Thu Feb 13 18:46:56 2003
From: hori ＠ swimmy-soft.com (Atsushi HORI)
Date: Thu, 13 Feb 2003 18:46:56 +0900
Subject: [SCore-users-jp] scrunコマンド
In-Reply-To: <20030213092228.9760B2004F@neal.il.is.s.u-tokyo.ac.jp>
References: <20030213144514E.m-kawaguchi@pst.fujitsu.com>
Message-ID: <3128006816.hori0002@swimmy-soft.com>

堀です．
 
>> 静的には、ジョブ実行されていないノードを自分で特定し
>> scrun実行時にオプション指定(ex. -group=PC1~node1~node2)すれば
>> 実現できると思います。
>
>最初の scrun で PC1 group 全部が占有されてしまうので,
>最初の scrun が group 指定である限り実現できないと思いますが...

ちょっと時間がないので簡単に．

SCore 5.4 で新規導入された msgblock コマンドをうまく使うと可能ではあり
ますが，今考えるとちょっと問題があるようですね．ちょっとしたハックが必
要です．

----
堀 敦史（ほり あつし）
スイミー・ソフトウェア株式会社



From m-kawaguchi ＠ pst.fujitsu.com  Thu Feb 13 21:31:54 2003
From: m-kawaguchi ＠ pst.fujitsu.com (川口貢)
Date: Thu, 13 Feb 2003 21:31:54 +0900
Subject: [SCore-users-jp] scrunコマンド
In-Reply-To: <20030213092228.9760B2004F@neal.il.is.s.u-tokyo.ac.jp>
References: <20030213144514E.m-kawaguchi@pst.fujitsu.com>
	<20030213092228.9760B2004F@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030213213154T.m-kawaguchi@pst.fujitsu.com>

亀山殿、堀殿

川口＠富士通プライムソフトテクノロジです。
回答ありがとうございました。

From: kameyama ＠ pccluster.org
Subject: Re: [SCore-users-jp] scrunコマンド
Date: Thu, 13 Feb 2003 18:22:28 +0900
Message-ID: <20030213092228.9760B2004F ＠ neal.il.is.s.u-tokyo.ac.jp>

> 亀山です.
> 
> In article <20030213144514E.m-kawaguchi ＠ pst.fujitsu.com> 川口貢 <m-kawaguchi ＠ pst.fujitsu.com> wrotes:
> > - グループ名 "PC1"に4台のノードが登録されている環境で
> >   以下の通りジョブ実行させる。
> > 
> >    % scrun -group=PC1,nodes=2x1 ./a.out
> > 
> >   上記状態で、ジョブ実行されていない2ノード上で別ジョブを実行させたい。
> >   但し、この時の2ノードの割り当てを動的に行いたい。
> > 
> >    % scrun ?????,nodes=2x1 ./a2.out
> > 
> > 静的には、ジョブ実行されていないノードを自分で特定し
> > scrun実行時にオプション指定(ex. -group=PC1~node1~node2)すれば
> > 実現できると思います。
> 
> 最初の scrun で PC1 group 全部が占有されてしまうので,
> 最初の scrun が group 指定である限り実現できないと思いますが...

ちょっと間違えました。すみません。
先のscrun実行時にグループ内のノード数を指定し、
次のscrun実行時に上記以外のノードを指定すれば
静的には実現できると思います。

 % scrun -group=PC1:2 ./a.out

   この後、ジョブ実行されているノードを特定し、次のジョブを実行。

 % scrun -group=PC1~node1~node2 ./a.out

でも、やはり動的にノードを割り当てることはできなさそうですね。
ありがとうございました。

> > ですが、空きノードがどれかということを意識せずに
> > ジョブ実行を行いたいのですが可能でしょうか?
> >   ＃ sc_qsubコマンドは不可とさせて下さい。
> 
> multi user mode を使用すればできますが, single user mode では
> できないと思います.
> 
>                        from Kameyama Toyohisa
> _______________________________________________
> SCore-users-jp mailing list
> SCore-users-jp ＠ pccluster.org
> http://www.pccluster.org/mailman/listinfo/score-users-jp

---
川口  mail => m-kawaguchi ＠ pst.fujitsu.com


From s-sumi ＠ flab.fujitsu.co.jp  Sat Feb 15 23:20:18 2003
From: s-sumi ＠ flab.fujitsu.co.jp (Shinji Sumimoto)
Date: Sat, 15 Feb 2003 23:20:18 +0900 (JST)
Subject: [SCore-users-jp] 第2 回ＰＣクラスタシンポジウム開催のお知らせ
In-Reply-To: <20030203.095315.884014234.ishikawa@is.s.u-tokyo.ac.jp>
References: <20030203.095315.884014234.ishikawa@is.s.u-tokyo.ac.jp>
Message-ID: <20030215.232018.730558638.s-sumi@flab.fujitsu.co.jp>

SCoreユーザの皆様

今回のPCクラスタコンソーシアムの企業展示ではItanium 2クラスタやOpteron
マシンの展示など興味深い展示が目白押しです。企業展示は21日だけでなく、
20日もレジストレーション無しに参加可能ですので、興味のある方の参加をお
待ちしております。

http://www.pccluster.org/event/symp/2002/exhibits.html

From: Yutaka Ishikawa <ishikawa ＠ is.s.u-tokyo.ac.jp>
Subject: [SCore-users-jp] 第2 回ＰＣクラスタシンポジウム開催のお知らせ
Date: Mon, 03 Feb 2003 09:53:15 +0900 (JST)
Message-ID: <20030203.095315.884014234.ishikawa ＠ is.s.u-tokyo.ac.jp>

ishikawa> SCore Usersの皆様、
ishikawa> 
ishikawa> シンポジウム事前申し込みの締め切りが一週間後となりました。
ishikawa> 参加を予定されている方で、まだ登録されていない方は、今すぐWEB登録のほど
ishikawa> お願い致します。
ishikawa> シンポジウムでは、
ishikawa> 	SCoreの最新情報（SCore 5.4の配布）
ishikawa> 	企業によるPCクラスタの最新の取り組み
ishikawa> などが無料で聴講できます。
ishikawa> 
ishikawa> 石川＠PCクラスタコンソーシアム会長
ishikawa> ＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝
ishikawa>  
ishikawa> 　　　　　　 第2回PCクラスタシンポジウムのご案内
ishikawa>  
ishikawa> 
ishikawa>    構造解析、流体力学、分子動力学、ゲノム情報処理など、大規模計算を
ishikawa>    必要とする計算科学・計算化学分野では、研究機関を中心にPCクラスタ
ishikawa>    が利用されてきました。最近では、PCクラスタは産業界にも浸透し、例え
ishikawa>    ば、自動車産業界では、車体設計、衝突シミュレーションなどにPCクラス
ishikawa>    タが利用されています。
ishikawa>    PCクラスタコンソーシアム主催の第2回PCクラスタシンポジウムでは、
ishikawa>    会員企業による産業界におけるPCクラスタ導入事例の講演や企業展示
ishikawa>    など、PCクラスタに関する最新の動向をお伝えします。皆様の御参加を
ishikawa>    お待ちしております。
ishikawa>  
ishikawa>  ■開催日時 2003年2月20日(木)〜2003年2月21日(金)
ishikawa>    10:00〜17:00
ishikawa>  
ishikawa>  ■開催場所 日本科学未来館7階
ishikawa>  
ishikawa>    2003年2月20日(木):ワークショップ・・・みらいCANホール
ishikawa>    2003年2月21日(金):シンポジウム・・・・みらいCANホール
ishikawa>    併設企業展示：イノベーションホール
ishikawa>    なお、2月20日(木)のワークショップの参加対象は会員のみとなります。
ishikawa>  
ishikawa>  ■参加費 無料
ishikawa>  
ishikawa>  ■申込み方法  http://www.pccluster.org/ <http://www.pccluster.org/> 　より
ishikawa> お申込み下さい。
ishikawa>  
ishikawa>  ■定員  300名
ishikawa>  
ishikawa>  ■主催  PCクラスタコンソーシアム
ishikawa>  
ishikawa>    TEL:03-3263-6474, FAX:03-3263-7537
ishikawa>    e-mail:sec ＠ pccluster.org
ishikawa>    U R L :http://www.pccluster.org/
ishikawa>  
ishikawa>  ＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝
ishikawa> _______________________________________________
ishikawa> SCore-users-jp mailing list
ishikawa> SCore-users-jp ＠ pccluster.org
ishikawa> http://www.pccluster.org/mailman/listinfo/score-users-jp
ishikawa> 
------
Shinji Sumimoto, Fujitsu Labs


From salzmann ＠ mpch-mainz.mpg.de  Wed Feb 19 02:55:53 2003
From: salzmann ＠ mpch-mainz.mpg.de (Marc Salzmann)
Date: Tue, 18 Feb 2003 18:55:53 +0100 (CET)
Subject: [SCore-users-jp] [SCore-users] multiuser mode
Message-ID: <Pine.OSF.4.30.0302181640380.13995-100000@jupiter.mpch-mainz.mpg.de>

Hi everybody,

When I was using SCore-D (5.2.0) in multiuser mode I could
run a 'hello world' test program if I logged in as root but not
if I logged in as another user.
The error message was: FEP:ERROR Unable to open secure port.
The problem also occured after having started scored with the
-nosecure option.
So I decided to re-compile scrun with #define SCORE_NOT_SECURE
and this solved the problem.
Is there any better (and maybe saver?) way to solve the problem
than this?
I'm a newbee and would be grateful for any advice.


best regards, Marc

_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/

Marc Salzmann
Max Planck Institute for Chemistry            tel: +49/(0)6131/305311
Department of Atmospheric Chemistry / NWG     fax: +49/(0)6131/305577
Postfach 3060
55020 Mainz, Germany
e-mail: salzmann ＠ mpch-mainz.mpg.de

_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/


_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Wed Feb 19 09:14:31 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Wed, 19 Feb 2003 09:14:31 +0900
Subject: [SCore-users-jp] Re: [SCore-users] multiuser mode
In-Reply-To: Your message of "Tue, 18 Feb 2003 18:55:53 JST."
             <Pine.OSF.4.30.0302181640380.13995-100000@jupiter.mpch-mainz.mpg.de>
Message-ID: <20030219001431.9245720050@neal.il.is.s.u-tokyo.ac.jp>

In article <Pine.OSF.4.30.0302181640380.13995-100000 ＠ jupiter.mpch-mainz.mpg.de> Marc Salzmann <salzmann ＠ mpch-mainz.mpg.de> wrotes:
> When I was using SCore-D (5.2.0) in multiuser mode I could
> run a 'hello world' test program if I logged in as root but not
> if I logged in as another user.
> The error message was: FEP:ERROR Unable to open secure port.

Please check permition of scrun.
    % ls -l /opt/score/bin/bin.*/scrun.exe
    -rwsr-xr-x    1 root     root      1194841 Oct 24 11:04 /opt/score/bin/bin.i386-redhat7-linux2_4/scrun.exe

scrun.exe must be owned by root, and must set set-uid bit.

Note that scored must be run by root.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From vegeta ＠ nerv-center.de  Thu Feb 20 19:36:48 2003
From: vegeta ＠ nerv-center.de (Henryk Feider)
Date: 20 Feb 2003 11:36:48 +0100
Subject: [SCore-users-jp] [SCore-users] Root password after EIT installation on nodes
Message-ID: <1045737414.1135.2.camel@belldandy>

Hi,

I installed a SCore cluster with EIT 2.0. But I can not lock into the
nodes because I do not know the root password. What password do EIT set
or where I can find this Information?

Greetings
Henryk Feider
-- 
Fight for love and justice!

Public key ID: 8065655F
example command: gpg --recv-keys --keyserver www.keyserver.net 8065655F 

-------------- next part --------------
テキスト形式以外の添付ファイルを保管しました...
ファイル名: signature.asc
型:         application/pgp-signature
サイズ:     189 バイト
説明:       This is a digitally signed message part
URL:        <http://new1.pccluster.org/pipermail/score-users-jp/attachments/20030220/a537cc4a/attachment.sig>

From vegeta ＠ nerv-center.de  Thu Feb 20 21:17:19 2003
From: vegeta ＠ nerv-center.de (Henryk Feider)
Date: 20 Feb 2003 13:17:19 +0100
Subject: [SCore-users-jp] Re: [SCore-users] Root password after EIT installation on nodes
In-Reply-To: <200302201042.h1KAgKs06088@emilepc.ess.nec.fr>
References: <200302201042.h1KAgKs06088@emilepc.ess.nec.fr>
Message-ID: <1045743440.1135.6.camel@belldandy>

I already tried this, but it does not work. Problem with EIT?

Greetings
Henryk Feider

On Thu, 2003-02-20 at 11:42, Emile CARCAMO wrote:
> 
> 
> vegeta ＠ nerv-center.de said:
> > I installed a SCore cluster with EIT 2.0. But I can not lock into the nodes
> > because I do not know the root password. What password do EIT set or where I
> > can find this Information? 
> 
> 
> Hello Henryk,
> 
> 	The password is the same that the one you have for
> 	root on the master machine (where you've run EIT!)
> 	HTH, and best regards.
> 
> -- 
> Emile_CARCAMO         NEC High Performance            http://www.hpce.nec.com
> System Engineer         Computing Europe         mailto:ecarcamo ＠ hpce.nec.com
> (+33)6-8063-7003   GSM
> (+33)1-3930-6601   FAX   / Your mouse has moved. Windows NT must be restarted \
> (+33)1-3930-6613  PHONE  \ for the change to take effect. Reboot now?  [ OK ] /
> 
> 
> 
> 
-- 
Fight for love and justice!

Public key ID: 8065655F
example command: gpg --recv-keys --keyserver www.keyserver.net 8065655F 

-------------- next part --------------
テキスト形式以外の添付ファイルを保管しました...
ファイル名: signature.asc
型:         application/pgp-signature
サイズ:     189 バイト
説明:       This is a digitally signed message part
URL:        <http://new1.pccluster.org/pipermail/score-users-jp/attachments/20030220/210e8b74/attachment.sig>

From nrcb ＠ streamline-computing.com  Thu Feb 20 21:32:24 2003
From: nrcb ＠ streamline-computing.com (Nick Birkett)
Date: Thu, 20 Feb 2003 12:32:24 +0000
Subject: [SCore-users-jp] Re: [SCore-users] Root password after EIT installation on nodes
In-Reply-To: <1045743440.1135.6.camel@belldandy>
References: <200302201042.h1KAgKs06088@emilepc.ess.nec.fr> <1045743440.1135.6.camel@belldandy>
Message-ID: <200302201232.h1KCWOY05889@zeralda.streamline.com>

On Thursday 20 February 2003 12:17 pm, you wrote:
> I already tried this, but it does not work. Problem with EIT?
>

Score is set up so you can rsh into the compute nodes as an ordinary user but 
not as root.

However you can still rsh a command or rcp a file as root.

To enable rsh logins as root please rcp a new /etc/pam.d/rlogin to each 
compute node. I use this pam.d/rlogin   

(ie auth       sufficient   /lib/security/pam_rhosts_auth.so is first line)


#%PAM-1.0
# For root login to succeed here with pam_securetty, "rlogin" must be
# listed in /etc/securetty.
auth       sufficient   /lib/security/pam_rhosts_auth.so
auth       required     /lib/security/pam_nologin.so
auth       required     /lib/security/pam_securetty.so
auth       required     /lib/security/pam_env.so
auth       required     /lib/security/pam_stack.so service=system-auth
account    required     /lib/security/pam_stack.so service=system-auth
password   required     /lib/security/pam_stack.so service=system-auth
session    required     /lib/security/pam_stack.so service=system-auth



eg rcp rlogin comp00:/etc/pamd.d/rlogin

With this you will then be able to rsh into comp00 etc.

Nick
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Thu Feb 20 23:52:23 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Thu, 20 Feb 2003 23:52:23 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Root password after EIT installation on nodes
In-Reply-To: Your message of "Thu, 20 Feb 2003 12:32:24 JST."
             <200302201232.h1KCWOY05889@zeralda.streamline.com>
Message-ID: <20030220145223.9652620050@neal.il.is.s.u-tokyo.ac.jp>

In article <200302201232.h1KCWOY05889 ＠ zeralda.streamline.com> Nick Birkett <nrcb ＠ streamline-computing.com> wrotes:
> On Thursday 20 February 2003 12:17 pm, you wrote:
> > I already tried this, but it does not work. Problem with EIT?

EIT get password from /etc/shadow or /etc/passwd on server host.
If your root password is not here, probably root password does not set.

> Score is set up so you can rsh into the compute nodes as an ordinary user but
>  
> not as root.
> 
> However you can still rsh a command or rcp a file as root.

If you can rsh and rcp as root,
you can scout command as root, too.

> To enable rsh logins as root please rcp a new /etc/pam.d/rlogin to each 
> compute node. I use this pam.d/rlogin   
> 
> (ie auth       sufficient   /lib/security/pam_rhosts_auth.so is first line)
> 
> 
> #%PAM-1.0
> # For root login to succeed here with pam_securetty, "rlogin" must be
> # listed in /etc/securetty.

In this comment, you don7t need this file.
If you want to rlogin as user, please issue follwing command:
    # scout -g pcc
    # scout "echo rlogin >> /etc/securetty"
    # exit
                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From vegeta ＠ nerv-center.de  Fri Feb 21 01:11:04 2003
From: vegeta ＠ nerv-center.de (Henryk Feider)
Date: 20 Feb 2003 17:11:04 +0100
Subject: [SCore-users-jp] Re: [SCore-users] Root password after EIT installation on nodes
In-Reply-To: <20030220145223.9652620050@neal.il.is.s.u-tokyo.ac.jp>
References: <20030220145223.9652620050@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <1045757464.1135.20.camel@belldandy>

Hi,
The password is set in /etc/passwd on the nodes (I boot it with 'init
/bin/sh' to find out). But even using the keyboard on the node, I can
not log into it. The 'ssh' fails too.

Thanks & Greetings 
Henryk Feider

On Thu, 2003-02-20 at 15:52, kameyama ＠ pccluster.org wrote:
> In article <200302201232.h1KCWOY05889 ＠ zeralda.streamline.com> Nick Birkett <nrcb ＠ streamline-computing.com> wrotes:
> > On Thursday 20 February 2003 12:17 pm, you wrote:
> > > I already tried this, but it does not work. Problem with EIT?
> 
> EIT get password from /etc/shadow or /etc/passwd on server host.
> If your root password is not here, probably root password does not set.
> 
> > Score is set up so you can rsh into the compute nodes as an ordinary user but
> >  
> > not as root.
> > 
> > However you can still rsh a command or rcp a file as root.
> 
> If you can rsh and rcp as root,
> you can scout command as root, too.
> 
> > To enable rsh logins as root please rcp a new /etc/pam.d/rlogin to each 
> > compute node. I use this pam.d/rlogin   
> > 
> > (ie auth       sufficient   /lib/security/pam_rhosts_auth.so is first line)
> > 
> > 
> > #%PAM-1.0
> > # For root login to succeed here with pam_securetty, "rlogin" must be
> > # listed in /etc/securetty.
> 
> In this comment, you don7t need this file.
> If you want to rlogin as user, please issue follwing command:
>     # scout -g pcc
>     # scout "echo rlogin >> /etc/securetty"
>     # exit
>                        from Kameyama Toyohisa
> _______________________________________________
> SCore-users mailing list
> SCore-users ＠ pccluster.org
> http://www.pccluster.org/mailman/listinfo/score-users
-- 
Fight for love and justice!

Public key ID: 8065655F
example command: gpg --recv-keys --keyserver www.keyserver.net 8065655F 

-------------- next part --------------
テキスト形式以外の添付ファイルを保管しました...
ファイル名: signature.asc
型:         application/pgp-signature
サイズ:     189 バイト
説明:       This is a digitally signed message part
URL:        <http://new1.pccluster.org/pipermail/score-users-jp/attachments/20030220/4b7e3b9a/attachment.sig>

From jure.jerman ＠ rzs-hm.si  Fri Feb 21 15:00:12 2003
From: jure.jerman ＠ rzs-hm.si (Jure Jerman)
Date: Fri, 21 Feb 2003 07:00:12 +0100
Subject: [SCore-users-jp] [SCore-users] problems with jobs locking cluster
Message-ID: <3E55C06C.1040306@rzs-hm.si>

Dear all,

sometimes I have annoying problem with score: when for instance
the deadlock is detected, score tries to attach gdb debugger
to the process, but in my case this fails due to the DISPLAY
problems (Can not open DISPLAY - even if I set the DISPLAY variable explicitly
and I am able to run other X applications it
does not work). The problem is that such a process lock whole
machine: it just hangs there and no other no jobs are running,
no jobs can be submitted and even no jobs can be aborted or killed.
The only solution is to restart sc_watch.

Does anyone has an idea how to:

1. get the gdb working with score
2. prevent score of attaching debugger to the process when not desired
    (cron submitted job for example)
3. generaly handle the situation where some jobs can not be killed via
    sc_console.

I am using Score 5.2.0.

Many thanks in advance, Jure Jerman

_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hori ＠ swimmy-soft.com  Fri Feb 21 20:14:09 2003
From: hori ＠ swimmy-soft.com (Atsushi HORI)
Date: Fri, 21 Feb 2003 20:14:09 +0900
Subject: [SCore-users-jp] Re: [SCore-users] problems with jobs locking cluster
In-Reply-To: <3E55C06C.1040306@rzs-hm.si>
References: <3E55C06C.1040306@rzs-hm.si>
Message-ID: <3128703249.hori0001@swimmy-soft.com>

Hi.

>Does anyone has an idea how to:
>
>1. get the gdb working with score

I suspect that your X server is not accesible from cluster compute 
hosts. you must do "xhost +" to allow access from any hosts.

>2. prevent score of attaching debugger to the process when not desired
>    (cron submitted job for example)

SCore tries to attach debugger only with "debug" optiom. Further, 
jobs scheduled in bacth scheduling mode, attaching debugger feature 
is disabled.

>3. generaly handle the situation where some jobs can not be killed via
>    sc_console.

This could be a SCore bug. Could you tell me the situation in more 
deitail ?

----
Atsushi HORI
Swimmy Software, Inc.

_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From okamoto ＠ gsport.co.jp  Fri Feb 21 23:15:22 2003
From: okamoto ＠ gsport.co.jp (Okamoto Masafumi)
Date: Fri, 21 Feb 2003 23:15:22 +0900
Subject: [SCore-users-jp] mpi++における共有ライブラリのリンクについて
Message-ID: <005101c2d9b3$abeeea10$6e0010ac@tommy>

お世話になっております、ジースポートの岡本です。

SCore5.0.1をredhat7.2上で使用しております。

開発しているプログラムでXMLを使用しているため、
Xerces(http://xml.apache.org/xerces-c/)を利用しています。

Xercesを共有ライブラリとして使っているのですが
ここでMakefileのg++をmpic++に置き換えてビルドすると 以下のエラーが起きてしま
います。
/usr/bin/ld; cannot find -lxerces-c
collect2: ld returned 1 exit status

g++でmakeした場合には問題なくビルドできます。

Makefileの中身は
g++ → mpic++
gcc → mpicc
を変更しただけです。

リンクしている部分は以下の様です。

/opt/score/bin/mpic++ -L. -o hoge1.o hoge2.o
hoge3.o -L../xerces-c-src2_1_0/lib -lxerces-c


何かアドバイスをいただけると幸いです。





From kameyama ＠ pccluster.org  Mon Feb 24 09:28:49 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Mon, 24 Feb 2003 09:28:49 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Root password after EIT installation on nodes
In-Reply-To: Your message of "20 Feb 2003 17:11:04 JST."
             <1045757464.1135.20.camel@belldandy>
Message-ID: <20030224002849.8059C20054@neal.il.is.s.u-tokyo.ac.jp>

In article <1045757464.1135.20.camel ＠ belldandy> Henryk Feider <vegeta ＠ nerv-center.de> wrotes:
> The password is set in /etc/passwd on the nodes (I boot it with 'init
> /bin/sh' to find out). But even using the keyboard on the node, I can
> not log into it. The 'ssh' fails too.

Please check /etc/passwd and /etc/shadow on the nodes.
If /etc/passwd is following:
    root:x:0:0:root:/root:/bin/bash
(password field is x), root password is stored in /etc/shadow.
I think password field is same as server host.
If this is diffirent, please change to password using passwd command (boot with
"init /bin/sh").

Can you use "su" to normal user?
    % su
    Password:
    #
If you can become root using su, please check /etc/securetty.


                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Mon Feb 24 09:33:49 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Mon, 24 Feb 2003 09:33:49 +0900
Subject: [SCore-users-jp] mpi++における共有ライブラリのリンクについて
In-Reply-To: Your message of "Fri, 21 Feb 2003 23:15:22 JST."
             <005101c2d9b3$abeeea10$6e0010ac@tommy>
Message-ID: <20030224003349.7B18C20054@neal.il.is.s.u-tokyo.ac.jp>

亀山です.

In article <005101c2d9b3$abeeea10$6e0010ac ＠ tommy> "Okamoto Masafumi" <okamoto ＠ gsport.co.jp> wrotes:
> お世話になっております、ジースポートの岡本です。
> 
> SCore5.0.1をredhat7.2上で使用しております。
> 
> 開発しているプログラムでXMLを使用しているため、
> Xerces(http://xml.apache.org/xerces-c/)を利用しています。
> 
> Xercesを共有ライブラリとして使っているのですが
> ここでMakefileのg++をmpic++に置き換えてビルドすると 以下のエラーが起きてしま
> います。
> /usr/bin/ld; cannot find -lxerces-c
> collect2: ld returned 1 exit status
> 
> g++でmakeした場合には問題なくビルドできます。

SCore では checkpoint が static link された binary を前提としているため,
default で static link するようにしています.
共有 library を使用する場合は -nostatic オプションをつけてください.

> /opt/score/bin/mpic++ -L. -o hoge1.o hoge2.o
> hoge3.o -L../xerces-c-src2_1_0/lib -lxerces-c

    /opt/score/bin/mpic++ -nostatic -L. -o hoge1.o hoge2.o hoge3.o -L../xerces-c-src2_1_0/lib -lxerces-c

のようになると思います.

                       from Kameyama Toyohisa


From johannes.werhahn ＠ imk.fzk.de  Tue Feb 25 00:37:15 2003
From: johannes.werhahn ＠ imk.fzk.de (Johannes Werhahn)
Date: Mon, 24 Feb 2003 16:37:15 +0100
Subject: [SCore-users-jp] [SCore-users] Adding new hosts with eit
Message-ID: <3E5A3C2B.9060900@imk.fzk.de>

Hi there,
we got a ready configured system with 10 hosts (1 server, 9 compute 
hosts) from our distributer and now want to add 2 other compute hosts 
using eit.
Is this the correct procedure:
after pressing the Load botton I go through the Network Configuration 
etc. leaving them unchanged until
Host information
# of Hosts  12   instead of 10
Name Prefix star (unchanged)
digit 0 Figur 1 (unchanged)
If I then press Add, I get
Duplicate host name star1.cluster.domain
So where should  I inluce star11.cluster.domain to get it ready for the 
automatic installation procedure?
And is it possible to configure two groups, one with Ethernet and 1 with 
Myrinet2000 (our new hosts have Myrinet-cards, the others not)?

Thanks for any hint and best regards


Johannes Werhahn

--------------------------------------------------------
Institut für Meteorologie und Klimaforschung
Bereich Atmosphärische Umweltforschung
Forschungszentrum Karlsruhe GmbH
Kreuzeckbahnstr. 19
D-82467 Garmisch-Partenkirchen
Germany
Phone  +49-8821-183-244
Fax    +49-8821-183-243
Email  johannes.werhahn ＠ imk.fzk.de
WWW    http://www.fzk.de
--------------------------------------------------------

_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Tue Feb 25 09:15:55 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Tue, 25 Feb 2003 09:15:55 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Adding new hosts with eit
In-Reply-To: Your message of "Mon, 24 Feb 2003 16:37:15 JST."
             <3E5A3C2B.9060900@imk.fzk.de>
Message-ID: <20030225001555.E367820054@neal.il.is.s.u-tokyo.ac.jp>

In article <3E5A3C2B.9060900 ＠ imk.fzk.de> Johannes Werhahn <johannes.werhahn ＠ imk.fzk.de> wrotes:
> after pressing the Load botton I go through the Network Configuration 
> etc. leaving them unchanged until
> Host information
> # of Hosts  12   instead of 10
> Name Prefix star (unchanged)
> digit 0 Figur 1 (unchanged)
> If I then press Add, I get
> Duplicate host name star1.cluster.domain
> So where should  I inluce star11.cluster.domain to get it ready for the 
> automatic installation procedure?

Please specify ONLY new hostname.
Probably, # of hostname is 2, Name prefix is unchanged, start is 10.

> And is it possible to configure two groups, one with Ethernet and 1 with 
> Myrinet2000 (our new hosts have Myrinet-cards, the others not)?

It is possible.
You can add new group and modify group on group list (next window).
But if you use SCore 5.0.1 or older, EIT generate wrong pm-myrinet.conf.
Please check /opt/score/etc/pm-myrinet.conf.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From okamoto ＠ gsport.co.jp  Tue Feb 25 11:16:09 2003
From: okamoto ＠ gsport.co.jp (Okamoto Masafumi)
Date: Tue, 25 Feb 2003 11:16:09 +0900
Subject: [SCore-users-jp] mpi++における共有ライブラリのリンクについて
In-Reply-To: <20030224003349.7B18C20054@neal.il.is.s.u-tokyo.ac.jp>
References: <005101c2d9b3$abeeea10$6e0010ac@tommy> <20030224003349.7B18C20054@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030225111444.5FC6.OKAMOTO@gsport.co.jp>

岡本です、-nostaticオプションをつけたところ
問題なくビルドできました。

どうも有難うございました。

kameyama> 亀山です.
kameyama> 
kameyama> In article <005101c2d9b3$abeeea10$6e0010ac ＠ tommy> "Okamoto Masafumi" <okamoto ＠ gsport.co.jp> wrotes:
kameyama> > お世話になっております、ジースポートの岡本です。
kameyama> > 
kameyama> > SCore5.0.1をredhat7.2上で使用しております。
kameyama> > 
kameyama> > 開発しているプログラムでXMLを使用しているため、
kameyama> > Xerces(http://xml.apache.org/xerces-c/)を利用しています。
kameyama> > 
kameyama> > Xercesを共有ライブラリとして使っているのですが
kameyama> > ここでMakefileのg++をmpic++に置き換えてビルドすると 以下のエラーが起きてしま
kameyama> > います。
kameyama> > /usr/bin/ld; cannot find -lxerces-c
kameyama> > collect2: ld returned 1 exit status
kameyama> > 
kameyama> > g++でmakeした場合には問題なくビルドできます。
kameyama> 
kameyama> SCore では checkpoint が static link された binary を前提としているため,
kameyama> default で static link するようにしています.
kameyama> 共有 library を使用する場合は -nostatic オプションをつけてください.
kameyama> 
kameyama> > /opt/score/bin/mpic++ -L. -o hoge1.o hoge2.o
kameyama> > hoge3.o -L../xerces-c-src2_1_0/lib -lxerces-c
kameyama> 
kameyama>     /opt/score/bin/mpic++ -nostatic -L. -o hoge1.o hoge2.o hoge3.o -L../xerces-c-src2_1_0/lib -lxerces-c
kameyama> 
kameyama> のようになると思います.
kameyama> 
kameyama>                        from Kameyama Toyohisa
kameyama> _______________________________________________
kameyama> SCore-users-jp mailing list
kameyama> SCore-users-jp ＠ pccluster.org
kameyama> http://www.pccluster.org/mailman/listinfo/score-users-jp




From nrcb ＠ streamline-computing.com  Wed Feb 26 20:49:05 2003
From: nrcb ＠ streamline-computing.com (Nick Birkett)
Date: Wed, 26 Feb 2003 11:49:05 +0000
Subject: [SCore-users-jp] [SCore-users] score 5.0.1 large memory jobs
Message-ID: <200302261149.h1QBn5L04090@zeralda.streamline.com>

Hi I got a report from one of our users runnning very big 
Pallas benchmark tests (128 cpus).



#----------------------------------------------------------------
# Benchmarking Alltoall 
# ( #processes = 64 ) 
# ( 64 additional processes waiting in MPI_Barrier)
#----------------------------------------------------------------
       #bytes #repetitions  t_min[usec]  t_max[usec]  t_avg[usec]
            0         1000       784.39       784.86       784.79
            1         1000       792.01       792.18       792.10
            2         1000       785.42       785.78       785.68
            4         1000       796.99       797.23       797.14
            8         1000       800.98       801.11       801.06
           16         1000       778.84       779.19       779.11
           32         1000       787.78       788.14       788.03
           64         1000       821.54       821.79       821.66
          128         1000       881.18       881.38       881.30
          256         1000       952.46       952.64       952.56
          512         1000      1158.49      1159.00      1158.88
         1024         1000      1640.78      1644.24      1641.00
         2048         1000      3454.18      3454.95      3454.62
         4096         1000      6882.82      6884.97      6883.97
         8192         1000     16088.81     16094.80     16091.81
        16384         1000     33715.59     33732.60     33727.56
        32768         1000     65014.80     65027.62     65023.50
        65536          640    129590.04    129636.99    129623.44
       131072          320    263434.38    263628.56    263587.57
       262144          160    531708.42    532274.39    532124.75
       524288           80   1069253.25   1071251.60   1070571.90
      1048576           40   2173875.02   2187574.55   2184477.23
      2097152           20   4228944.70   4270372.05   4258162.98
      4194304           10   8398147.40   8512784.40   8478838.18
<8> SCore-D:PANIC Network freezing timed out !!
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hori ＠ swimmy-soft.com  Wed Feb 26 23:13:27 2003
From: hori ＠ swimmy-soft.com (Atsushi HORI)
Date: Wed, 26 Feb 2003 23:13:27 +0900
Subject: [SCore-users-jp] Re: [SCore-users] score 5.0.1 large memory jobs
In-Reply-To: <200302261149.h1QBn5L04090@zeralda.streamline.com>
References: <200302261149.h1QBn5L04090@zeralda.streamline.com>
Message-ID: <3129146007.hori0002@swimmy-soft.com>

Hi.

>Hi I got a report from one of our users runnning very big 
>Pallas benchmark tests (128 cpus).

This looks like the timeout value is too small. BTW, what network are 
you (is he or she) using ? Mriynet or Ethernet (I bet on ethernet) ? 
The next question is how big is the maximum messages ?

>#----------------------------------------------------------------
># Benchmarking Alltoall 
># ( #processes = 64 ) 
># ( 64 additional processes waiting in MPI_Barrier)
>#----------------------------------------------------------------
>       #bytes #repetitions  t_min[usec]  t_max[usec]  t_avg[usec]
>            0         1000       784.39       784.86       784.79
>            1         1000       792.01       792.18       792.10
>            2         1000       785.42       785.78       785.68
>            4         1000       796.99       797.23       797.14
>            8         1000       800.98       801.11       801.06
>           16         1000       778.84       779.19       779.11
>           32         1000       787.78       788.14       788.03
>           64         1000       821.54       821.79       821.66
>          128         1000       881.18       881.38       881.30
>          256         1000       952.46       952.64       952.56
>          512         1000      1158.49      1159.00      1158.88
>         1024         1000      1640.78      1644.24      1641.00
>         2048         1000      3454.18      3454.95      3454.62
>         4096         1000      6882.82      6884.97      6883.97
>         8192         1000     16088.81     16094.80     16091.81
>        16384         1000     33715.59     33732.60     33727.56
>        32768         1000     65014.80     65027.62     65023.50
>        65536          640    129590.04    129636.99    129623.44
>       131072          320    263434.38    263628.56    263587.57
>       262144          160    531708.42    532274.39    532124.75
>       524288           80   1069253.25   1071251.60   1070571.90
>      1048576           40   2173875.02   2187574.55   2184477.23
>      2097152           20   4228944.70   4270372.05   4258162.98
>      4194304           10   8398147.40   8512784.40   8478838.18
><8> SCore-D:PANIC Network freezing timed out !!
>_______________________________________________
>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 suga ＠ sse.co.jp  Thu Feb 27 15:50:38 2003
From: suga ＠ sse.co.jp (Sugano, Mitsukuni)
Date: Thu, 27 Feb 2003 15:50:38 +0900
Subject: [SCore-users-jp] SCore 5.4.0 - Itanium2
Message-ID: <3E5DB53E.4A85E7AC@sse.co.jp>

SSE　菅野と申します。
お世話になります。

現在、Itanium2上で、SCore5.4をソースから構築しているのですが、
以下のような不具合および疑問点があります。

１．SMP

実行時に、SHMEMデバイスが無いので、smpの数を１にリセットする
旨のWarningが表示される。現在、２ｘ２環境で実行していますが、
nodesが２であればWarnigだけで実行課のですが、４の場合、SMP
が使えないので、当然のように実行できません。

なお、scorehosts.dbのSHMEM設定は、確認済みです。

また、5.4のBugFix情報で、
−−−
Bug Fixes 
SCore does not work more than 128 hosts.
The mpich 1.2.4 logviewer command does not have an execution right.
The mpich 1.2.4 Fortran 90 library does not exist.
PM/shmem does not work in IA64.
−−−
と、いう記述があり、「PM/shmem does not work in IA64.」の不具合
は解消されているようなのですが、通常の5.4.0のソースではなく異なる
ソースが存在するのでしょうか？

２．mandel
mandelのデモが、以下のエラーで実行できません。
<0:0> SCORE: 2 nodes (2x1) ready.
Worker die, display exit

３．オプションコンパイラ

siteファイルを変更しただけではだめなようなので、
mpi等をSCoreソースから再コンパイルする必要があると思うのですが、
その際に、複数のコンパイラを指定可能なのですよね？
また、Cは、gccのみ。 Fortranは、g77とifc(v6)とifc(v7)
といったような混在環境で且つ、ifcはあるがiccはないと
いった場合でも構築可能なのでしょうか？
その手順を含めご教示いただけませんでしょうか？


From kameyama ＠ pccluster.org  Thu Feb 27 16:54:46 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Thu, 27 Feb 2003 16:54:46 +0900
Subject: [SCore-users-jp] SCore 5.4.0 - Itanium2
In-Reply-To: Your message of "Thu, 27 Feb 2003 15:50:38 JST."
             <3E5DB53E.4A85E7AC@sse.co.jp>
Message-ID: <20030227075446.4C90220054@neal.il.is.s.u-tokyo.ac.jp>

亀山です.

In article <3E5DB53E.4A85E7AC ＠ sse.co.jp> "Sugano, Mitsukuni" <suga ＠ sse.co.jp> wrotes:
> 実行時に、SHMEMデバイスが無いので、smpの数を１にリセットする
> 旨のWarningが表示される。現在、２ｘ２環境で実行していますが、
> nodesが２であればWarnigだけで実行課のですが、４の場合、SMP
> が使えないので、当然のように実行できません。

compute host で dmesg を行い,
    pmshmem: version = $Id: pm_shmem.c,v ...
    pmshmem_init: register pm_shmem as major(124)
のような message がでているかどうか確認してください.

また,
    /dev/pmshmem/0
などの file があるかどうかを確認してください.

> また、5.4のBugFix情報で、
> −−−
> Bug Fixes 
> SCore does not work more than 128 hosts.
> The mpich 1.2.4 logviewer command does not have an execution right.
> The mpich 1.2.4 Fortran 90 library does not exist.
> PM/shmem does not work in IA64.
> −−−
> と、いう記述があり、「PM/shmem does not work in IA64.」の不具合
> は解消されているようなのですが、通常の5.4.0のソースではなく異なる
> ソースが存在するのでしょうか？

SCore の IA64 のサポートは SCore 5.2.0 から入っています.
これは 5.2.0 から 5.4.0 で fix された bug で,
しかも PM/shmem 自体は認識されていてエラーが出るというものです.

> ２．mandel
> mandelのデモが、以下のエラーで実行できません。
> <0:0> SCORE: 2 nodes (2x1) ready.
> Worker die, display exit

再現しました.
調査します.

> ３．オプションコンパイラ
> 
> siteファイルを変更しただけではだめなようなので、
> mpi等をSCoreソースから再コンパイルする必要があると思うのですが、
> その際に、複数のコンパイラを指定可能なのですよね？
> また、Cは、gccのみ。 Fortranは、g77とifc(v6)とifc(v7)
> といったような混在環境で且つ、ifcはあるがiccはないと
> いった場合でも構築可能なのでしょうか？
> その手順を含めご教示いただけませんでしょうか？

ifc ということは IA32 の場合でしょうか?
(IA64 の Intel compiler は efc になります.)
IA32 の場合は SCore 5.2.0 より Intel compiler version 6 の binary が
入っています.
SCore 5.4.0 では EIT のときに Intel compiler version 6, version 7 ともに
入っていますので, site ファイルを変更しただけで動くと思います.

この binary は icc, ifc ともにある場合を想定しているため,
うまく link できないかもしれません.
(済みません, 試していないのでわかりません.)
その場合は MPI を再コンパイルする必要があると思います.
その場合, 以下の手順になります.
(IA64 の場合も同様の手順でインストールできます.)

1. /opt/score/etc/compiler/site ファイルを修正します.
2. mpi のソースプログラムを /opt/score のしたに展開します.
      # cd /opt/score
      # tar vxzf /mnt/cdrom/score.ssource/score-5.4.0.mpi.tar.gz
3. MPI プログラムをコンパイル, インストールします.
      # cd /opt/score/score-src/runtime/mpi
      # smake
      # smake install
4. これでエラーがなければ使用できると思います.

                       from Kameyama Toyohisa


From kameyama ＠ pccluster.org  Thu Feb 27 18:19:39 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Thu, 27 Feb 2003 18:19:39 +0900
Subject: [SCore-users-jp] SCore 5.4.0 - Itanium2
In-Reply-To: Your message of "Thu, 27 Feb 2003 16:54:46 JST."
             <20030227075446.4C90220054@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030227091939.BA5BE20054@neal.il.is.s.u-tokyo.ac.jp>

亀山です.

In article <20030227075446.4C90220054 ＠ neal.il.is.s.u-tokyo.ac.jp> kameyama ＠ pccluster.org wrotes:
> > ２．mandel
> > mandelのデモが、以下のエラーで実行できません。
> > <0:0> SCORE: 2 nodes (2x1) ready.
> > Worker die, display exit
> 
> 再現しました.
> 調査します.

IA64 だと, SCore 環境で fork() が無視されるようになっていました.
とりあえず,
      score-src/program/demo/mandel/Makefile
の LDFLAGS に
    -nockpt
を追加して checkpoint 無しで compile してください.
(それでもまだ表示が変ですけど...)

                       from Kameyama Toyohisa


From suga ＠ sse.co.jp  Thu Feb 27 19:13:21 2003
From: suga ＠ sse.co.jp (Sugano, Mitsukuni)
Date: Thu, 27 Feb 2003 19:13:21 +0900
Subject: [SCore-users-jp] SCore 5.4.0 - Itanium2
References: <20030227091939.BA5BE20054@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <3E5DE4C1.C1274784@sse.co.jp>

亀山様：

菅野です。
お世話になります。

おかげさまで、先ほどお伺いした問題すべてが解決いたしました。

今後とも、よろしくお願いいたします。

kameyama ＠ pccluster.org wrote:
> 
> 亀山です.
> 
> In article <20030227075446.4C90220054 ＠ neal.il.is.s.u-tokyo.ac.jp> kameyama ＠ pccluster.org wrotes:
> > > ２．mandel
> > > mandelのデモが、以下のエラーで実行できません。
> > > <0:0> SCORE: 2 nodes (2x1) ready.
> > > Worker die, display exit
> >
> > 再現しました.
> > 調査します.
> 
> IA64 だと, SCore 環境で fork() が無視されるようになっていました.
> とりあえず,
>       score-src/program/demo/mandel/Makefile
> の LDFLAGS に
>     -nockpt
> を追加して checkpoint 無しで compile してください.
> (それでもまだ表示が変ですけど...)
> 
>                        from Kameyama Toyohisa
> _______________________________________________
> SCore-users-jp mailing list
> SCore-users-jp ＠ pccluster.org
> http://www.pccluster.org/mailman/listinfo/score-users-jp


From nrcb ＠ streamline-computing.com  Thu Feb 27 19:29:11 2003
From: nrcb ＠ streamline-computing.com (Nick Birkett)
Date: Thu, 27 Feb 2003 10:29:11 +0000
Subject: [SCore-users-jp] Re: [SCore-users] score 5.0.1 large memory jobs
Message-ID: <200302271029.h1RATB801837@zeralda.streamline.com>

Sorry, the message I sent was truncated and therefore confusing.

Here it is again:

Score 5.0.1, Myrinet 2000 system.

---------message from user -------------------------------


Following the addition of swap on all Snowdon compute nodes, I reran the
PALLAS benchmark tests (on 64 nodes running 2 processes per node). The
following output was recorded towards the end of the run:

#----------------------------------------------------------------
# Benchmarking Alltoall 
# ( #processes = 64 ) 
# ( 64 additional processes waiting in MPI_Barrier)
#----------------------------------------------------------------
       #bytes #repetitions  t_min[usec]  t_max[usec]  t_avg[usec]
            0         1000       784.39       784.86       784.79
            1         1000       792.01       792.18       792.10
            2         1000       785.42       785.78       785.68
            4         1000       796.99       797.23       797.14
            8         1000       800.98       801.11       801.06
           16         1000       778.84       779.19       779.11
           32         1000       787.78       788.14       788.03
           64         1000       821.54       821.79       821.66
          128         1000       881.18       881.38       881.30
          256         1000       952.46       952.64       952.56
          512         1000      1158.49      1159.00      1158.88
         1024         1000      1640.78      1644.24      1641.00
         2048         1000      3454.18      3454.95      3454.62
         4096         1000      6882.82      6884.97      6883.97
         8192         1000     16088.81     16094.80     16091.81
        16384         1000     33715.59     33732.60     33727.56
        32768         1000     65014.80     65027.62     65023.50
        65536          640    129590.04    129636.99    129623.44
       131072          320    263434.38    263628.56    263587.57
       262144          160    531708.42    532274.39    532124.75
       524288           80   1069253.25   1071251.60   1070571.90
      1048576           40   2173875.02   2187574.55   2184477.23
      2097152           20   4228944.70   4270372.05   4258162.98
      4194304           10   8398147.40   8512784.40   8478838.18
<8> SCore-D:PANIC Network freezing timed out !!

And the .e file states:

<0:0> SCORE: 128 nodes (64x2) ready.
<56:1> SCORE:WARNING MPICH/SCore: pmGetSendBuffer(pmc=0x8541db8, dest=37,
len=8256) failed, errno=22
<56:1> SCORE:PANIC MPICH/SCore: critical error on message transfer
<56:1> Trying to attach GDB (DISPLAY=snowdon.leeds.ac.uk:18.0): PANIC
SCOUT: Session done.

It looks like now the memory allocation is working fine, but the benchmark
is unable to undertake the next test in the benchmark.

The next test is an all-to-all zero length message to 128 processors (on 64
nodes). Extrapolating the results, this should take about 1.6 ms.

It appears as if the communications grind to a halt when we try to
communicate between 128 processes (and over) when running 2 processes per
node.
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From chen ＠ mdl.ipc.pku.edu.cn  Fri Feb 28 01:40:42 2003
From: chen ＠ mdl.ipc.pku.edu.cn (Chen Hao)
Date: Fri, 28 Feb 2003 00:40:42 +0800
Subject: [SCore-users-jp] [SCore-users] A very strange problem
Message-ID: <008201c2de7e$f9b94770$9101a8c0@billgates>

Dear All:
I have a very strange problem when I install GROMACS to my cluster(score 5.4.0). First the configure program reports that mpicc cannot handle assembly files, so I disable assembly loops and then configure program produces the Makefile file successfully. And then I execute make, after a few minutes I get the execute files. But when I try to run it, system reports that "No such file or directory". what's the matter?

BTW, the execute file can be opened in vi and I run it with the correct path.

-------------- next part --------------
HTMLの添付ファイルを保管しました...
URL: <http://new1.pccluster.org/pipermail/score-users-jp/attachments/20030228/a23f47f6/attachment.html>

From s-sumi ＠ flab.fujitsu.co.jp  Fri Feb 28 12:59:18 2003
From: s-sumi ＠ flab.fujitsu.co.jp (Shinji Sumimoto)
Date: Fri, 28 Feb 2003 12:59:18 +0900 (JST)
Subject: [SCore-users-jp] Re: [SCore-users] score 5.0.1 large memory jobs
In-Reply-To: <200302271029.h1RATB801837@zeralda.streamline.com>
References: <200302271029.h1RATB801837@zeralda.streamline.com>
Message-ID: <20030228.125918.294709392.s-sumi@flab.fujitsu.co.jp>

Hi Nick.

Could you run the benchmark with PM_DEBUG=1?
If there are PM/Myrinet problems, some messages are output.

Ex: sh
export PM_DEBUG=1  

Shinji.

From: Nick Birkett <nrcb ＠ streamline-computing.com>
Subject: Re: [SCore-users] score 5.0.1 large memory jobs
Date: Thu, 27 Feb 2003 10:29:11 +0000
Message-ID: <200302271029.h1RATB801837 ＠ zeralda.streamline.com>

nrcb> Sorry, the message I sent was truncated and therefore confusing.
nrcb> 
nrcb> Here it is again:
nrcb> 
nrcb> Score 5.0.1, Myrinet 2000 system.
nrcb> 
nrcb> ---------message from user -------------------------------
nrcb> 
nrcb> 
nrcb> Following the addition of swap on all Snowdon compute nodes, I reran the
nrcb> PALLAS benchmark tests (on 64 nodes running 2 processes per node). The
nrcb> following output was recorded towards the end of the run:
nrcb> 
nrcb> #----------------------------------------------------------------
nrcb> # Benchmarking Alltoall 
nrcb> # ( #processes = 64 ) 
nrcb> # ( 64 additional processes waiting in MPI_Barrier)
nrcb> #----------------------------------------------------------------
nrcb>        #bytes #repetitions  t_min[usec]  t_max[usec]  t_avg[usec]
nrcb>             0         1000       784.39       784.86       784.79
nrcb>             1         1000       792.01       792.18       792.10
nrcb>             2         1000       785.42       785.78       785.68
nrcb>             4         1000       796.99       797.23       797.14
nrcb>             8         1000       800.98       801.11       801.06
nrcb>            16         1000       778.84       779.19       779.11
nrcb>            32         1000       787.78       788.14       788.03
nrcb>            64         1000       821.54       821.79       821.66
nrcb>           128         1000       881.18       881.38       881.30
nrcb>           256         1000       952.46       952.64       952.56
nrcb>           512         1000      1158.49      1159.00      1158.88
nrcb>          1024         1000      1640.78      1644.24      1641.00
nrcb>          2048         1000      3454.18      3454.95      3454.62
nrcb>          4096         1000      6882.82      6884.97      6883.97
nrcb>          8192         1000     16088.81     16094.80     16091.81
nrcb>         16384         1000     33715.59     33732.60     33727.56
nrcb>         32768         1000     65014.80     65027.62     65023.50
nrcb>         65536          640    129590.04    129636.99    129623.44
nrcb>        131072          320    263434.38    263628.56    263587.57
nrcb>        262144          160    531708.42    532274.39    532124.75
nrcb>        524288           80   1069253.25   1071251.60   1070571.90
nrcb>       1048576           40   2173875.02   2187574.55   2184477.23
nrcb>       2097152           20   4228944.70   4270372.05   4258162.98
nrcb>       4194304           10   8398147.40   8512784.40   8478838.18
nrcb> <8> SCore-D:PANIC Network freezing timed out !!
nrcb> 
nrcb> And the .e file states:
nrcb> 
nrcb> <0:0> SCORE: 128 nodes (64x2) ready.
nrcb> <56:1> SCORE:WARNING MPICH/SCore: pmGetSendBuffer(pmc=0x8541db8, dest=37,
nrcb> len=8256) failed, errno=22
nrcb> <56:1> SCORE:PANIC MPICH/SCore: critical error on message transfer
nrcb> <56:1> Trying to attach GDB (DISPLAY=snowdon.leeds.ac.uk:18.0): PANIC
nrcb> SCOUT: Session done.
nrcb> 
nrcb> It looks like now the memory allocation is working fine, but the benchmark
nrcb> is unable to undertake the next test in the benchmark.
nrcb> 
nrcb> The next test is an all-to-all zero length message to 128 processors (on 64
nrcb> nodes). Extrapolating the results, this should take about 1.6 ms.
nrcb> 
nrcb> It appears as if the communications grind to a halt when we try to
nrcb> communicate between 128 processes (and over) when running 2 processes per
nrcb> node.
nrcb> _______________________________________________
nrcb> SCore-users mailing list
nrcb> SCore-users ＠ pccluster.org
nrcb> http://www.pccluster.org/mailman/listinfo/score-users
nrcb> 
nrcb> 
------
Shinji Sumimoto, Fujitsu Labs
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From tyokoi ＠ jodco.co.jp  Fri Feb 28 14:57:40 2003
From: tyokoi ＠ jodco.co.jp (Takeshi Yokoi)
Date: Fri, 28 Feb 2003 14:57:40 +0900
Subject: [SCore-users-jp] MPICH-Score version problem
Message-ID: <3E5EFA54.A4A32194@jodco.co.jp>

みなさん、こんにちは

　　　現在SCoreの導入を検討しております。
MPICH-SCoreについて、

　社内既存のSimulation Softのにカスタマイズされた
　MPICHがはいっているため、SCoreのMPICHでは
　なく既存のMPICHに見に行かせることは可能でしょうか。
　（それとも単純に上書きすれば良いとか）


  初歩的な質問ですみません。


　　　　　　　　　　　　　　　　　　　　　　　　　　　　　T.Yokoi
　　　



From kameyama ＠ pccluster.org  Fri Feb 28 16:40:06 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Fri, 28 Feb 2003 16:40:06 +0900
Subject: [SCore-users-jp] MPICH-Score version problem
In-Reply-To: Your message of "Fri, 28 Feb 2003 14:57:40 JST."
             <3E5EFA54.A4A32194@jodco.co.jp>
Message-ID: <20030228074006.16C8820055@neal.il.is.s.u-tokyo.ac.jp>

亀山です.

In article <3E5EFA54.A4A32194 ＠ jodco.co.jp> Takeshi Yokoi <tyokoi ＠ jodco.co.jp> wrotes:
> 　社内既存のSimulation Softのにカスタマイズされた
> 　MPICHがはいっているため、SCoreのMPICHでは
> 　なく既存のMPICHに見に行かせることは可能でしょうか。
> 　（それとも単純に上書きすれば良いとか）

やりたいことがはっきりしませんが...
  1. MPICH/SCore とその他の MPI を使い分けたい場合
  2. MPICH/SCore でカスタマイズされた機能を使いたい場合
が考えられ, どちらがやりたいのか分かりませんので, 両方書きます.

1. MPICH/SCore とその他の MPI を使い分けたい場合

SCore 自体は /opt/score に install されます.
それと異なった場所に別の mpi が存在しているかどうかは何も調べません.
(たとえば, Redhat 7.3 を full i0nstal すると LAM が install されています.
それと独立に SCore が install されます.)
RPM を使用して SCore を install すると, /etc/profile.d/score.* という
ファイルが install され, このファイルの中で SCore 関連の PATH を
いれています.

つまり, 同一 host に SCore 以外の MPI を共存させることは可能で,
user の PATH によってどちらの MPI を使用するかを選択することができます.
たとえば, redhat 7.3 を full install した場合, user のサーチパスが
/opt/score/bin のほうが先にあれば SCore の MPI が /usr/bin のほうが
先にあれば LAM の MPI が使用されることになります.
(コマンドを full path で指定すれば, 明示的にどちらかを指定することが
できます.)

注意しなければならないのは, コンパイルと実行環境を会わせる必要があることです.
MPICH/SCore の mpicc などでコンパイルした実行ファイルは MPICH/SCore の
mpirun で実行する必要がありますし, 他の MPI のコンパイルコマンドで
作成した実行ファイルはその MPI の mpirun で実行する必要があります.

2. MPICH/SCore でカスタマイズされた機能を使いたい場合

MPICH のライブラリにデバイス依存の部分も組み込まれてしまっているため,
カスタマイズ部分がデバイスに依存していない部分であっても, binary
のままでは使用できません.
もし, そのカスタマイズされた部分のソースがあるのでしたら,
その部分をマージすることができるかも知れません.

基本的には MPICH/SCore は MPICH のソースに SCore 依存部分をいれたものです.
(独自の bug fix, 複数 compiler や RPM に対応するための修正が入っていますが...)
そのため, MPICH/SCore のソースファイルを展開し, カスタマイズした部分を
マージすれば使えると思います.

                       from Kameyama Toyohisa


From kameyama ＠ pccluster.org  Fri Feb 28 19:24:40 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Fri, 28 Feb 2003 19:24:40 +0900
Subject: [SCore-users-jp] Re: [SCore-users] A very strange problem
In-Reply-To: Your message of "Fri, 28 Feb 2003 00:40:42 JST."
             <008201c2de7e$f9b94770$9101a8c0@billgates>
Message-ID: <20030228102440.2577D20055@neal.il.is.s.u-tokyo.ac.jp>

In article <008201c2de7e$f9b94770$9101a8c0 ＠ billgates> "Chen Hao" <chen ＠ mdl.ipc.p
ku.edu.cn> wrotes:
> I have a very strange problem when I install GROMACS to
> my cluster(score 5.4.0).
> First the configure program reports that mpicc cannot handle assembly files,
> so I disable assembly loops and then configure program produces the Makefile
> file successfully.

Currently, mpicc cannot handle assembly files,
This probrem is cause mpich 1.2.4, too, but fixed on mpich 1.2.5.

> And then I execute make, after a few minutes I get the execute files.
> But when I try to run it, system reports that "No such file or directory".

Please add configure options to:
    --disable-shared --without-motif-libraries

If you want to run mdrun, please use scrun:
    % scrun mdrun -np ...

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users