Hey Jed Model precaching is very important considering the model first exists on the hard drive, and then it needs to be broken up once loaded.
The model is loaded (FROM THE HARD DRIVE), all the required data is laid out in an easy to retrieve manner in memory, the video card memory probably gets a vertex buffer sent to it plus more. The biggest thing to keep in mind especially when developing games is to never load from the hard drive at run time, the bus between the hard drive and your memory is ridiculously slow. Again, caching is very important mostly in that the model exists on the hard drive and shouldn't be constantly loaded from the hard drive every time it is needed. Also, the memory usage is insignificant compared to the performance gains. I hope this somewhat explains why Valve throw up red warnings whenever you precache something late. Ryan On Wed, Sep 14, 2011 at 8:02 AM, AndreaZzZ <andrea_s...@mail.ru> wrote: > If there were levels of detail, I think things would be a lot easier. > > > 2011/9/14 Garry Newman <garrynew...@gmail.com> > >> I stopped models precaching in GMod - and you can get up to a second >> loading delay when spawning a model - especially ragdolls. >> >> So I think it's safe to assume it loads the whole thing, model, physics, >> materials, textures. >> >> garry >> >> >> On Wed, Sep 14, 2011 at 2:57 PM, Jed <j...@wunderboy.org> wrote: >> >>> I've been wondering about model pre-caching lately and what exactly it >>> achieves. >>> >>> I'm noticing that the potential number of player models for our mod is >>> growing exponentially and along with that there are various play model >>> attachments which are needed. I could just put them all into an array >>> and pre-cache them in one go but I keep wondering if that in fact >>> could have a negative effect? >>> >>> "aching" to me implies that the model is loaded into memory and held >>> there for quick access but in our case while there are say maybe 32 >>> separate models, only 6 would say ever be used at one time in a single >>> map. So my thought is that while the caching will stop stutters >>> waiting for the model to load, will it in fact use more memory as not >>> all the cached models are being used. >>> >>> So does model pre-caching actually store the whole model in RAM or >>> does is simply let the system know that this model might be needed and >>> do some sort of preparation for it? >>> >>> - Jed >>> >>> _______________________________________________ >>> To unsubscribe, edit your list preferences, or view the list archives, >>> please visit: >>> http://list.valvesoftware.com/mailman/listinfo/hlcoders >>> >>> >> >> _______________________________________________ >> To unsubscribe, edit your list preferences, or view the list archives, >> please visit: >> http://list.valvesoftware.com/mailman/listinfo/hlcoders >> >> >> > > _______________________________________________ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlcoders > > > -- ~Ryan ( skidz )
_______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
