Charm++ parallel programming system

[Gengbin Zheng <zhenggb AT gmail.com>]

That is what I thought. The production release of charm does not
support heterogeneous clusters.

Gengbin


On Wed, Jun 30, 2010 at 1:47 AM, Chiara Orsini
<chiara.orsini AT iet.unipi.it>
 wrote:
> The first machine is a MacBook with MacOSX 10.6.4
> and charm-6.2.0-net-darwin-x86_64.
> The second machine is a PC with Ubuntu 9.04 ( 2.6.28.19-generic )
> and charm-6.2.0-net-linux-x86_64.
> Chiara
>
> 2010/6/29 Gengbin Zheng 
> <zhenggb AT gmail.com>
>>
>> How different are these two machines?
>> It looks like the program starts ok on two machines, but they crashed
>> at very early stage.
>>
>> Gengbin
>>
>> On Tue, Jun 29, 2010 at 9:10 AM, Chiara Orsini
>> <chiara.orsini AT iet.unipi.it>
>>  wrote:
>> > Dear charm++ users,
>> >
>> >
>> > I tried to run SimpleArrayHello program on multiple PCs, but the program
>> > ends with this error:
>> >
>> > Charmrun> node programs all connected
>> > Charmrun: error on request socket--
>> > Socket closed before recv.
>> >
>> > I read the FAQ, but linking with -memory paranoid did not give me any
>> > useful
>> > information.
>> > Consider two host A and B. I configured host A and B in order to be
>> > accessible with each other through ssh (without password).
>> > I can run BasicHelloWorld on host A (launching it from host A)
>> > successfully.
>> > I can run BasicHelloWorldon host B (launching it from host A)
>> > successfully.
>> > But, I cannot run BasicHelloWorld on host A and B (launching it from
>> > host
>> > A). This is my nodelist file:
>> > group main ++shell ssh
>> > host IP_B
>> > host IP_A
>> > This is the output that I obtain running this command:  ./charmrun
>> > PROGRAM_NAME +p2 ++verbose
>> > Charmrun> charmrun started...
>> > Charmrun> using ./nodelist as nodesfile
>> > Charmrun> adding client 0: "IP_B", IP:IP_B
>> > Charmrun> adding client 1: "IP_A", IP:IP_A
>> > Charmrun> Charmrun = IP_A, port = 50501
>> > Charmrun> Sending "0 IP_A 50501 2547 0" to client 0.
>> > Charmrun> find the node program "/PATH/PROGRAM_NAME/prefix" at
>> > "/PATH/PROGRAM_NAME" for 0.
>> > Charmrun> node 0: xterm is xterm
>> > Charmrun> Starting ssh IP_B -l user /bin/sh -f
>> > Charmrun> remote shell (IP_B:0) started
>> > Charmrun> Sending "1 IP_A 50501 2547 0" to client 1.
>> > Charmrun> find the node program "/PATH/PROGRAM_NAME/prefix" at
>> > "/PATH/PROGRAM_NAME" for 1.
>> > Charmrun> node 1: xterm is xterm
>> > Charmrun> Starting ssh IP_A -l user /bin/sh -f
>> > Charmrun> remote shell (IP_A:1) started
>> > Charmrun> node programs all started
>> > Charmrun remote shell(IP_A.1)> remote responding...
>> > Charmrun remote shell(IP_A.1)> using xterm /usr/X11R6/bin/xterm
>> > Charmrun remote shell(IP_A.1)> starting node-program...
>> > Charmrun remote shell(IP_A.1)> rsh phase successful.
>> > Charmrun remote shell(IP_B.0)> remote responding...
>> > Charmrun remote shell(IP_B.0)> using xterm /usr/X11R6/bin/xterm
>> > Charmrun remote shell(IP_B.0)> starting node-program...
>> > Charmrun remote shell(IP_B.0)> rsh phase successful.
>> > Charmrun> Waiting for 0-th client to connect.
>> > Charmrun> Waiting for 1-th client to connect.
>> > Charmrun> client 1 connected (IP=IP_A data_port=54659)
>> > Charmrun> client 0 connected (IP=IP_B data_port=34817)
>> > Charmrun> All clients connected.
>> > Charmrun> IP tables sent.
>> > Charmrun> node programs all connected
>> > Charmrun: error on request socket--
>> > Socket closed before recv.
>> >
>> >
>> > Could anyone explain how to solve this problem?
>> > Any information you'll give me will be surely appreciated.
>> > Thank you in advance.
>> > Best regards,
>> > Chiara
>> >
>> >
>> >
>> >
>> >
>> > _______________________________________________
>> > charm mailing list
>> > charm AT cs.uiuc.edu
>> > http://lists.cs.uiuc.edu/mailman/listinfo/charm
>> >
>> >
>>
>
>