If GM supports multiple interfaces in the same host,how does the mapper work? For example,node A with two NICs named A1 and A2 communicate nodeB with two NICs named B1 and B2.A message sent to node B from A1,how do I know that B1 receive the message or B2 receive the message?
By default, the driver names each interface as hostname:1, hostname:2 etc. The mapper uses these names in the routing table and the name table, it distributes to all nodes at the end of the mapping process. You would need to use "interface names", such as A:1 and B:2, in your call to gm_host_name_to_node_id() to get the remote interface id. This id is then used for your regular send and receive. Vijay -- Post-Doctoral Research Associate, Architecture & Real-time Systems (ARTS) Lab, 310 Knowles Engineering Bldg, University of Massachusetts, Amherst MA 01003 Tel: (413)-545-0715 (O) Fax: (413)-545-1993 On Thu, 5 Dec 2002 wangjsa <at> Legend.com wrote: > If GM supports multiple interfaces in the same host,how does the mapper > work? > For example,node A with two NICs named A1 and A2 communicate nodeB with two > NICs named B1 and B2.A message sent to node B from A1,how do I know that B1 > receive the message or B2 receive the message? > > _______________________________________________(Continue reading)
Hi! I've got a problem running HPL (High Performance Linpack), compiled with Intel MKL, under MPICH-GM. I am trying to run it on four 2-way Xeon nodes, starting 1 process per node. I let MKL to parallelize on each node (i.e. MKL creates 2 threads per node and both CPUs are loaded). This trick works with MPICH from Argonne, which utilizes P4 device, but not MPICH-GM. Under MPICH-GM I've got some messages about signal 11 after program start and it hangs. Has someone experience with MPICH-GM and MKL? P.S. When I limiting MKL with 1 thread per process, and running 2 processes per node, everything work fine. But I have information that MKL performs better with automatic parallelization. P.P.S. Configuration: - 4 2-way Xeon nodes with 1 Gig of RAM. - 1 Myrinet M3S-PCI64B adapter per node. - RedHat Linux 7.3 (kernel 2.4.16 from MSC.Linux distro) - GM 1.6.3 - MPICH for GM 1.2.4..8a - MPICH from Argonne 1.2.4 -------------------------------------------------------------- | Birukov Kirill birk <at> rdi-kvant.ru | | Senior System Administrator | | R&D Institute Kvant Moscow, Russia, 125438 | | Phone: +7 (095) 154-4362 4th Likhachevsky lane, 15 | --------------------------------------------------------------(Continue reading)
On Wed, Dec 11, 2002 at 04:17:37PM +0300, Birukov Kirill wrote: > Hi! > > I've got a problem running HPL (High Performance Linpack), compiled > with Intel MKL, under MPICH-GM. I am trying to run it on four 2-way > Xeon nodes, starting 1 process per node. I let MKL to parallelize on > each node (i.e. MKL creates 2 threads per node and both CPUs are > loaded). This trick works with MPICH from Argonne, which utilizes P4 > device, but not MPICH-GM. Under MPICH-GM I've got some messages about > signal 11 after program start and it hangs. Has someone experience > with MPICH-GM and MKL? I have run mpich-gm and mkl, but not with multiple threads per process. Other tests I have tried using multiple threads per process were never as fast as run running multiple processes (mkl or autoparallelization of a code). When I tried mkl 5.0, I found that Altas produced much faster results than mkl when running HPL. Have you turned hyperthreading off? If not, are you starting 4 processes or threads per node? AT SC2002, Dell presented a paper in their booth where running 4 processes on 2 proc box with HT on gave a boost of 10% or so. Craig > > P.S. When I limiting MKL with 1 thread per process, and running 2 > processes per node, everything work fine. But I have information that(Continue reading)
I get hordes of the following warnings anytime I compile C++ mpich code using mpich version 1.2.4: /usr/local/mpich-gm-1.2.4..8a/include/mpi2c++/functions_inln.h: In function `void PMPI::Pcontrol (int, ...)': /usr/local/mpich-gm-1.2.4..8a/include/mpi2c++/functions_inln.h:145: warning: cannot pass objects of non-POD type `struct va_list' through `...' Even the simplest mpi program yields this warning. Does anyome know what this means/how to fix it/how to make it go away? -Josh ----------------------------------------------- Josh England Sandia National Laboratory, Livermore, CA Distributed Information Systems email: jjengla <at> sandia.gov phone: (925) 294-2076
I am having problems installing GM v1.6.3 drivers on HP zx6000 boxes w/ Itanium2 processors. I've heard that there may be a patch to the drivers. If someone can point me in the right direction that would help. Thanks, -- John --
Hello, John Laroco, John wrote: > I am having problems installing GM v1.6.3 drivers on HP zx6000 boxes w/ Itanium2 processors. I've heard that there may be a patch to the drivers. > > If someone can point me in the right direction that would help. http://www.myri.com/scs/faq/faq-mapper.html#mapper1a Please reference above site. We installed 24 zx2000 at last week according some patch file. Best regards, - Hiroyuki Kobayashi
I've installed gm-1.6.3 on my Itanium2 machines and get an error message when I try and run the mapper. I've been pointed to the faq URL below for guidance. The faq states that there is patch to the strncpy function in glibc-2.2.4 to help fix the error message. The details into how to actually apply the patch and replace the glibc are very vague to me. http://www.myri.com/scs/faq/faq-mapper.html#mapper1a I was wondering if someone had a more detailed explanation for patching the glibc with respect to the faq at the above URL. Thanks for your help. -- John --
RSS Feed3 | |
|---|---|
1 | |
5 | |
3 | |
3 | |
8 | |
1 | |
2 | |
5 | |
4 | |
1 | |
2 | |
14 | |
8 | |
1 | |
3 | |
22 | |
5 | |
1 | |
6 | |
4 | |
2 | |
3 | |
4 | |
8 | |
4 | |
17 | |
9 | |
20 | |
7 | |
31 | |
3 | |
4 | |
3 | |
10 | |
2 | |
13 | |
5 | |
9 | |
14 | |
29 | |
48 | |
11 | |
8 | |
25 | |
16 | |
19 | |
5 | |
6 | |
50 | |
11 | |
9 | |
11 | |
35 | |
6 | |
47 | |
35 | |
12 | |
41 | |
12 | |
23 | |
6 | |
11 | |
7 | |
16 | |
35 | |
6 | |
11 | |
14 | |
19 | |
14 | |
10 | |
15 | |
12 | |
14 | |
1 | |
9 | |
7 | |
18 |
