memory/RAM-question

[Helmut Sonnweber]

Dear fds users,

 

I would like to run a fds 5.5.1 calculation for a 50MW fire in a tunnel, the problem is forrtl: severe (41): “insufficient virtual memory”

after about one hour of CPU time.

 

The model size is 20*3072*12 = 737280 cells (number of cells should meet fds mesh requirement 2-3-5) for the dimensions of 9.3 * 1425 * 5.6 meter.

The computer is equipped with 3GB RAM. (The rough number is 1GB RAM for 10^6 cells – depending on the problem)

it is no parallel FDS run

no other tasks are done by the computer

there are no fans in the tunnel

the fire is represented by a lorry with an area of 192m² and is ramped

 

I have seen the posts about virtual memory and does not exactly know, why it is used and if more RAM would be the solution or if it is related to something other (thermal boundary conditions,…)

 

Thanks for any hint, Helmut Sonnweber

[dr_jfloyd]

You have a large domain surface area compared to your volume (i.e. your domain is more like a plate than a cube).  This gives you a very large number of surface cells and surface cells consume memory as well.   You may just be running out of memory.  Have you tried reducing the number of grid cells (either resolution or change in domain size) to see if that at least gets FDS to run?  If it does, then your issue is probably memory.

[Helmut Sonnweber]

Dear dr_jfloyd,

 

yes I tried a smaller domain size = a piece of 50 meters (9.3 * 50 * 5.6 meter with 10*200*10 = 20000 cells), that calculation was ok so far as I can evaluate.

Thanks for your answer,

 

Helmut Sonnweber

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To view this discussion on the web visit https://groups.google.com/d/msg/fds-smv/-/s1lEhefclHUJ.
To post to this group, send email to fds...@googlegroups.com.
To unsubscribe from this group, send email to fds-smv+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/fds-smv?hl=en.

[martia shahsavan]

Hi dear Helmut,

You can use dimensionless analysis for your domain.

HRR : Q1/Q2 = (L1/L2)^(5/2)

Velocity : V1/V2 = (L1/L2)^(1/2)

Time : t1/t2 = (L1/L2)^(1/2)

Temperature : T1/T2 = 1

Energy : E1/E2 = (L1/L2)^3

Hope you be successful

On Jul 26, 3:05 pm, "Helmut Sonnweber" <h.sonnwe...@ids.or.at> wrote:
> Dear fds users,
>
> I would like to run a fds 5.5.1 calculation for a 50MW fire in a tunnel, the problem is forrtl: severe (41): "insufficient virtual memory"
>
> after about one hour of CPU time.
>
> The model size is 20*3072*12 = 737280 cells (number of cells should meet fds mesh requirement 2-3-5) for the dimensions of 9.3 * 1425 * 5.6 meter.
>

> The computer is equipped with 3GB RAM. (The rough number is 1GB RAM for 10^6 cells - depending on the problem)

>
> it is no parallel FDS run
>
> no other tasks are done by the computer
>
> there are no fans in the tunnel
>
> the fire is represented by a lorry with an area of 192m² and is ramped
>

> I have seen the posts about virtual memory and does not exactly know, why it is used and if more RAM would be the solution or if it is related to something other (thermal boundary conditions,...)

[Helmut Sonnweber]

Hi Martia,

Thanks for the post; how do I have to read this ?

If I have two domains, 1 and 2, and just one differing dimension between those two objects, this are the scaling factors to obtain what ?
Nearly the same memory usage for the 'problem' ?

Thanks a lot, Helmut Sonnweber

-----Ursprüngliche Nachricht-----
Von: fds...@googlegroups.com [mailto:fds...@googlegroups.com] Im Auftrag von martia shahsavan
Gesendet: Mittwoch, 27. Juli 2011 08:49
An: FDS and Smokeview Discussions
Betreff: [fds-smv] Re: memory/RAM-question

Hi dear Helmut,

HRR : Q1/Q2 = (L1/L2)^(5/2)

Velocity : V1/V2 = (L1/L2)^(1/2)

Time : t1/t2 = (L1/L2)^(1/2)

Temperature : T1/T2 = 1

Energy : E1/E2 = (L1/L2)^3

Hope you be successful

--

You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.