performance issue

Dear All,
I have an LP problem which has 678,870 constraints and 375,630 variables. Gams has been running for 5 hours. I use Cplex solver version 10.2.0. The computer has 2.2GHz Dual Core CPU and 1 GB ram.
Is this a normal situation such a big problem? Indeed, it is not a big problem. I need to solve bigger problems having 10 million constraints. What can I do other than reducing the number of constraints?

Thanks in advance.


To post to this group, send email to gamsworld@googlegroups.com.
To unsubscribe from this group, send email to gamsworld+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/gamsworld?hl=en.

Hakan,

You don’t say how many nonzeros your model has. With the row and
column counts you report it could easily be “pretty big”, but that
depends on the NZ/column.

I suggest you monitor the memory usage via the Windows task manager or
Linux’s top and see if that shows your model has a virtual memory
footprint larger than the 1GB of physical RAM you have. If so, you
will need to get more physical memory.

At commercial prices, the software you are using is far more costly
than the hardware you describe, and the 1GB of RAM you have will be a
limiting factor in what you can do. For large models, you should run
64-bit software on a machine with at least 4GB of RAM. It is easy to
find reasonably-priced machines with 8GB RAM, and colleagues of mine
report 16GB is nothing to consider exotic. Memory is cheap enough
that you don’t want to starve your model of it.

-Steve

On Sat, Jun 25, 2011 at 3:31 PM, Hakan Kutucu wrote:

Dear All,
I have an LP problem which has 678,870 constraints and 375,630 variables.
Gams has been running for 5 hours. I use Cplex solver version 10.2.0. The
computer has 2.2GHz Dual Core CPU and 1 GB ram.
Is this a normal situation such a big problem? Indeed, it is not a big
problem. I need to solve bigger problems having 10 million constraints. What
can I do other than reducing the number of constraints?

Thanks in advance.


“gamsworld” group.
To post to this group, send email to gamsworld@googlegroups.com.
To unsubscribe from this group, send email to
gamsworld+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/gamsworld?hl=en.




– Steven Dirkse, Ph.D. GAMS Development Corp., Washington DC Voice: (202)342-0180 Fax: (202)342-0181 sdirkse@gams.com http://www.gams.com

Steven,
thank you so much. You always give me real helpfull answers.
The number of nonzeros is 2.217.234. I found a pc which has 4 cores and 8 GB of Ram. I run the code. But Gams didn’t stop after 3 hours. Then I decided to use 4 threads. When I run the code with 4 threads gams produced an output less than 3 minutes. I would not expect such a big time diference. Could you explain why?

Best regards.

Date: Mon, 27 Jun 2011 16:24:40 -0400
Subject: Re: performance issue
From: sdirkse@gams.com
To: gamsworld@googlegroups.com

Hakan,

You don’t say how many nonzeros your model has. With the row and
column counts you report it could easily be “pretty big”, but that
depends on the NZ/column.

I suggest you monitor the memory usage via the Windows task manager or
Linux’s top and see if that shows your model has a virtual memory
footprint larger than the 1GB of physical RAM you have. If so, you
will need to get more physical memory.

At commercial prices, the software you are using is far more costly
than the hardware you describe, and the 1GB of RAM you have will be a
limiting factor in what you can do. For large models, you should run
64-bit software on a machine with at least 4GB of RAM. It is easy to
find reasonably-priced machines with 8GB RAM, and colleagues of mine
report 16GB is nothing to consider exotic. Memory is cheap enough
that you don’t want to starve your model of it.

-Steve

On Sat, Jun 25, 2011 at 3:31 PM, Hakan Kutucu wrote:

Dear All,
I have an LP problem which has 678,870 constraints and 375,630 variables.
Gams has been running for 5 hours. I use Cplex solver version 10.2.0. The
computer has 2.2GHz Dual Core CPU and 1 GB ram.
Is this a normal situation such a big problem? Indeed, it is not a big
problem. I need to solve bigger problems having 10 million constraints. What
can I do other than reducing the number of constraints?

Thanks in advance.


“gamsworld” group.
To post to this group, send email to gamsworld@googlegroups.com.
To unsubscribe from this group, send email to
gamsworld+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/gamsworld?hl=en.


Steven Dirkse, Ph.D.
GAMS Development Corp., Washington DC
Voice: (202)342-0180 Fax: (202)342-0181
sdirkse@gams.com
http://www.gams.com


To post to this group, send email to gamsworld@googlegroups.com.
To unsubscribe from this group, send email to gamsworld+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/gamsworld?hl=en.


To post to this group, send email to gamsworld@googlegroups.com.
To unsubscribe from this group, send email to gamsworld+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/gamsworld?hl=en.