STL, in large systems sims/high computing performance

[neurocomp2003]

(hehe here's another one for you dduardo :biggrin: )
Does anyone know if STL is used or is capable of being used in large system simulations? Or is it best to write your own datastructtures?
Looking to do some computational astrophysics or large neural nets and am not sure whether to use STL or not... I want to if it can be used for the large scale that i'm envisioning, but i don't now if there would be overhead that i wouldn't now about.

[dduardo]

You have to remember that STL is generalized. Depending on what you are doing you may be able to optimize the structures and performance of certain operations.

What type of structures do you plan on using?

[Hurkyl]

"More computing sins are committed in the name of efficiency (without necessarily achieving it) than for any other single reason - including blind stupidity."
"Premature optimization is the root of all evil."
It would probably be best to use the STL data structures where convenient, to more quickly produce working code, and then figure out what and how to optimize.

(reference: Wikipedia (http://en.wikipedia.org/wiki/Optimization_(computer_science)))

[dduardo]

"[...] the Linux philosophy is 'laugh in the face of danger'. Oops. Wrong one. 'Do it yourself'. That's it."

If you can reinvent the wheel, do it!

(reference: wikipedia (http://en.wikiquote.org/wiki/Linus_Torvalds))

j/k :tongue2:

[neurocomp2003]

vector stack, PQueue and list. are the ones i'm looking at. I'm just running a simple particle collision(n=1000) looking to go to n=10000 if i can find a system better than mind. Trying to land a research job. =]