DSXchange

Does DS uses 2 processors on the same machine ?

I have a win 2000 server with 2 processors on borad. Was just wondering when I run the jobs does DS server uses the second processor or not ?

Quite probably.
DataStage generates a process to run the job, and possibly (probably) other processes to execute active stages (such as Transformer stages). The operating system is likely to allocate multiple processes to multiple CPUs. Of course, on a Windows machine, there will be many other demands for CPU resources (look in Task Manager).
DataStage itself can not control the allocation of CPUs in version 5.x server jobs.

(The story changes in your future, when the DataStage licensing model changes so that it is licensed for a maximum number of CPUs rather than for a maximum number of projects. It remains true that DataStage does not control the allocation of particular CPUs, but it will impose an upper limit on the number of CPUs used based on the number specified in your licensing.)

You are using DataStage 5.x. This means that the jobs are single threaded, and have no inter-process stages. Because of this, a job will use only a single cpu. Run two jobs at the same time, you can use two cpus. If a job spends 1/4 the time on cpu #1 and 3/4 of the time on cpu #2, that's up to the operating system, but it will not use more than 4/4 of a cpu. If a job only needs 1/2 of a cpu, and you run a second job at the same time that only needs 1/4 of a cpu, it's still up to the OS to decide which cpu does which particular slice of work. Under later releases of DataStage, you do have the opportunity to have multiple processors actively helping in the work that a job does. There's no configuration for DataStage to limit or force cpu usage, that's completely up to the discretion of the OS.

Got that .. now we are in the process of upgrading the DS to DS 7.0.

What will be the scenario in this case? Will DS allocate the tasks to 2 CPU's or still leave it upto OS to decide it ?

Still pretty much up to the O/S and the new licensing model will mean that while your Project counts are now "unlimited", your CPU usage is now limited to your license count.

You'll be able to spread a single job across multiple CPUs by leveraging IPC at the job level or via the use of the new IPC stage. Check out the docs for the low-down on that new feature.