maxsc2 problem
 

Hi,

I am trying to generate a simple hex mesh to investigate shear flow. In the shear layer I need a lot of cells. Generating a refined mesh gives me the standard errors to increase maxvrt and maxcel.

However, I am now hitting the maximum for maxsc2.

Using prosize to increase it to what is required results in "use a large machine" error - my 2Gb of RAM is not enough.

What methods are there to optimize memory usage within prosize?

Is it safe to reduce some of the parameters (the ones I may not use eq splines) to zero?

Thanks,

Tristan

Re: maxsc2 problem
 

You might delete the unused vertices and cell numbers after each refinement step to save memory.


cset all
vset news cset
vset inve
vdel vset
ccom all $ y
vcom all $ y

Re: maxsc2 problem
 

You can try setting unused items (splines, droplets, ...) to small non-0 numbers but you are not likely to save much. Try to get maxcel, maxvrt, maxnbu to as tight a number as you possibly can. Also set them if at all possible to the max number you have from the start instead of making them bigger while you are in pro-STAR. Make sure you have enough swap space (more than 2Gb) so you can borrow from that if needed. Swapping may be slow but it beats not being able to do a job.

Re: maxsc2 problem
 

Cheers Steve,

Just a quick question again: How do I increase the swap space that StarCD will use? I have 90Gig of HD space to play with and 2Gig of RAM. Is there a similar MEMORY command to increase this?

See my newer post about how I am trying to get around this memory issue by using a better mesh with grading and other refinements created in Gambit.

Tristan