Is it worth to upgrade?

[daraeill]

i'd say you'd experience a very small increase in performance depending on what you are trying to do.......which is a rather ridiculous way of asking what you do with maya.....

rendering times would be decreased, certainly...but the ui, physics calculations (which i must say it's rather ridiculous that modern software hasn't been programmed to use multiple threads for physics calcs...that is something that does need to be changed...anyone with a good understanding of calculus or physics could tell you that all of the complicated calculations can be multi-calculated with accuracy ---i'm talking common calcs that cg artists would encounter...force, inertia, caustics...etc)anyhoo...physics calculations which includes hair, fur, dynamics, etc....may slow down going with a slightly slower processor....

this is postulated with the idea that you wouldn't be overclocking of course...if you overclock you can get the best of both worlds (single and multi threaded)

[Julius]

My main concern for a cpu upgrade is rendertime. I have been reading people are raving about them but havent found anything comparing to what i have atm... Just wondering if anyone else has made an upgrade from a high single core to a dual core and if its more then a couple seconds faster?

[daraeill]

you know i really should read the whole question first...

you're comparing a 3500+ (2.2GHz) vs a 4800+ X2 (2.4GHz)...

ummm...well if you have the money...yes (ie if you are deciding between paying rent and buying this processor--NO...if the decision is between buying this processor or a glass pack for your car...get the 4800+ :)

wth...im a know-it-all??

[lots]

Not all algorithms lend thems selves well to multi threading. Some cant even be broken down in such a way with current methods. This is why things like phyics still are single threaded. There's not much that can be done about those types of calculations.

Overall having the second core will make renders faster in general. Granted some renders will benifit more than others, but different types of renders benifit from different types of hardware setups. It's just how things are.

EDIT:
Taking the advice of an earlier post, I went and asked a Math professor about this. I work in the department of mathematics at the university in my area, so this is a rather simple thing for me to do (aside from the fact that both my parents are math professors, my dad with a PH.d from Berkley :P) Anyway, the answer I got was things like projectile motion, collisions, etc. do not really lend them selves well to multi threading. The reasons being that finding the path of a projectile object requires knowledge of previously calculated steps, if you only know the origin and velocity of the object. It is possible to multithread some of these calculations, for example finding the paths of an object's peices when shattered, however this requires constant communication between all threads ensuring that each fragment does not further collide with other fragments on thier own separate paths, which would further alter thier velocity. So while there can be some paralellizm in physics computations, it is not possible to further break down the calculation of each individual path.

Granted if you knew the beginning end, and velocity, it would be easier to calculate the object at any point in its path.. But alot of times this is not the case in physics calculations in CG.

[Julius]

Thankyou so much for the replys, i think i'll just stick with what i have for now :)

[daraeill]

ahh thanks for that explanation lots...my calculus is just a tad rusty ;) i guess i do see why noone bothers to add multi-threaded capability in the physics engines given that it would really only help on a very limited basis...

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.