[OMPI devel] 答复: 答复: 答复: 答复: doubt on latency result with OpenMPI library

2014-03-28 Thread Wang,Yanfei(SYS)
复: 答复: 答复: doubt on latency result with OpenMPI library On Mar 27, 2014, at 11:45 PM, "Wang,Yanfei(SYS)" wrote: > 1. In the RoCE, we cannot use OOB(via tcp socket) for RDMA connection. More specifically, RoCE QPs can only be made using the RDMA connection manager. > However,

[OMPI devel] 答复: 答复: 答复: doubt on latency result with OpenMPI library

2014-03-27 Thread Wang,Yanfei(SYS)
oubt on latency result with OpenMPI library On Mar 27, 2014, at 11:15 AM, "Wang,Yanfei(SYS)" wrote: > Normally we use rdma-cm to build rdma connection ,then create Qpairs to do > rdma data transmit ion, so what is the consideration for separating rdma-cm > connection built and

Re: [OMPI devel] 答复: 答复: doubt on latency result with OpenMPI library

2014-03-27 Thread Wang,Yanfei(SYS)
to use RDMA for communication. What it does is tell the openib BTL to use RDMA to establish the point-to-point connection between the two processes. The actual messaging may or may not use RDMA to move the bytes - that's a totally separate code path. On Thu, Mar 27, 2014 at 6:2

Re: [OMPI devel] 答复: 答复: doubt on latency result with OpenMPI library

2014-03-27 Thread Wang,Yanfei(SYS)
e point-to-point connection between the two processes. The actual messaging may or may not use RDMA to move the bytes - that's a totally separate code path. On Thu, Mar 27, 2014 at 6:21 AM, Wang,Yanfei(SYS) mailto:wangyanfe...@baidu.com>> wrote: it seem that all confusions have already been

Re: [OMPI devel] 答复: 答复: doubt on latency result with OpenMPI library

2014-03-27 Thread Wang,Yanfei(SYS)
off-node > communication. > > Also, remember that you can "mpirun ... hostname" (i.e., the Linux "hostname" > command) to verify what servers you are actually running on. > > I see that the ompi_info(1) man page is not super-detailed about the --level >

[OMPI devel] 答复: 答复: doubt on latency result with OpenMPI library

2014-03-27 Thread Wang,Yanfei(SYS)
btl_tcp_if_include. [root@bb-nsi-ib04 pt2pt]# Hope to have a deep understand on it~ Thanks --Yanfei 发件人: devel [mailto:devel-boun...@open-mpi.org] 代表 Wang,Yanfei(SYS) 发送时间: 2014年3月27日 18:17 收件人: Open MPI Developers 主题: [OMPI devel] 答复: doubt on latency result with OpenMPI library HI,

[OMPI devel] 答复: doubt on latency result with OpenMPI library

2014-03-27 Thread Wang,Yanfei(SYS)
间: 2014年3月27日 18:05 收件人: Open MPI Developers 主题: Re: [OMPI devel] doubt on latency result with OpenMPI library Try adding "--map-by node" to your command line to ensure the procs really are running on separate nodes. On Thu, Mar 27, 2014 at 1:40 AM, Wang,Yanfei(SYS) mailto:wangyanfe...

[OMPI devel] doubt on latency result with OpenMPI library

2014-03-27 Thread Wang,Yanfei(SYS)
Hi, HW Test Topology: Ip:192.168.72.4/24 �C192.168.72.4/24, enable vlan and RoCE IB03 server 40G port-- - 40G Ethernet switch IB04 server 40G port: configure it as RoCE link IP: 192.168.71.3/24 ---192.168.71.4/24 IB03 server 10G port �C 10G Ethernet switch �C IB04 server 10G port: configure

[OMPI devel] 答复: 答复: example/Hello_c.c : mpirun run failed on two physical nodes.

2014-03-27 Thread Wang,Yanfei(SYS)
. Thanks Ralph On Tue, Mar 25, 2014 at 9:51 PM, Wang,Yanfei(SYS) mailto:wangyanfe...@baidu.com>> wrote: Hi, Thanks jeff, and I have not figured out what happened yet with this FAQ. 1. Ssh remote login OK: [root@bb-nsi-ib04 examples]# ssh ib03 hostname bb-nsi-ib03.bb01.*.com [root@bb-nsi-ib04 ex

[OMPI devel] 答复: example/Hello_c.c : mpirun run failed on two physical nodes.

2014-03-26 Thread Wang,Yanfei(SYS)
pi.org/faq/?category=running#diagnose-multi-host-problems On Mar 25, 2014, at 6:54 AM, "Wang,Yanfei(SYS)" wrote: > Hi, > > I am a fresh learner of OpenMPI programmer, and have some troubles on > building mpi programming, hope some helps.. > > The example/helloc_c ca

[OMPI devel] example/Hello_c.c : mpirun run failed on two physical nodes.

2014-03-25 Thread Wang,Yanfei(SYS)
Hi, I am a fresh learner of OpenMPI programmer, and have some troubles on building mpi programming, hope some helps.. The example/helloc_c can works successfully with 2 process on local machine, however, do not work on two separate physical node. Physical two nodes: Eg: [root@bb-nsi-ib04 examp