[torqueusers] Directly linked nodes via cross crossover cable

Aaron Greenwood agreenwo at uci.edu
Mon Feb 13 09:26:20 MST 2006


Consider the following hardware configuration:

NODE 1 (2 CPUS)
eth0 - Connected to cluster Ethernet switch.
eth1 - Directly linked via cross crossover cable to NODE 2

NODE 2 (2 CPUS)
eth0 - Connected to cluster Ethernet switch.
eth1 - Directly linked via cross crossover cable to NODE 1

Is it possible to configure PBS in such a way that a parallel job
submitted from the head node will use all CPUS on NODE 1 and NODE 2
running over the Ethernet cards that are directly linked?

The directly linked cards are on a private network listed in the local
hosts file.

I talked with a guy who does this. He said that in the script that he
submits his jobs he modifies the machine_file as in lamboot -s
machine_file. When I do that the jobs run using the Ethernet cards
connected to the cluster switch. I checked this by logging on to both of
the nodes and checking traffic with tcpdump and running lamnodes.

The purpose for this scheme is high throughput for certain jobs that run
for may days and weeks.

I am new to clusters and parallel computing so if this has been
discussed before I would appreciate a pointer to available information
on this topic.

Best regards,

Aaron



More information about the torqueusers mailing list