Hi,
I am working on a cluster of workstations to which I have remote access. The cluster is having 3 dual core nodes. It uses LINUX. The question is that if we accidentally delete a file with the command "rm <filename>", is it possible to recover the file? If yes how to do this?
Second, I use LAM/MPI on this cluster. When I use "top" command it shows only 4 cpus. If there are 6 cpus why is it showing only 4? It always shows the action of 4 cpus only. A snapshot of the result of top is given below:
--------------------------------------------------------------------------------------------------------------------------------------
12:04:44 up 26 days, 2:06, 2 users, load average: 0.00, 0.00, 0.00
84 processes: 83 sleeping, 1 running, 0 zombie, 0 stopped
CPU states: cpu user nice system irq softirq iowait idle
total 0.0% 0.0% 0.2% 0.0% 0.0% 0.0% 99.7%
cpu00 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 100.0%
cpu01 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 100.0%
cpu02 0.0% 0.0% 0.9% 0.0% 0.0% 0.0% 99.0%
cpu03 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 100.0%
Mem: 2061092k av, 982556k used, 1078536k free, 0k shrd, 261492k buff
426256k active, 146976k inactive
Swap: 4096564k av, 0k used, 4096564k free 264008k cached
----------------------------------------------------------------------------------------------------------------------------------------
I do parallel programming in C with MPI. Currently I can use at the most 45 processes with "mpirun -np <n> <filename>" command. Now I have following questions:
1) Is it possible to use more parallel processes in anyway? If yes, how?
2) When I use "MPI_Get_processor_name()" function, it always displays the name of one node only.
How can we see which nodes are working and how can we control the number of nodes to be used?
3) I found somewhere the concept of machine file that can be used for the purpose but I don't know how to use it.
Also some sources show PBS node file. But it is also not working. When I use PBS node file and select the number of nodes it still works the same. Actually I want to see how the processes are distributed among nodes and processors. How can I do this?
Any help will be greatly appreciated..
Thanks in advance :)
Madhuri
|
