macho eke podi aththak thiyanwa eth dan 15000 RPM Hard Drive thiyanwa.anika machan RAM eke mathakaya thawakalikai.eka nisa aniwa HDD ekak ona. poddak hoyala balapan macho... 






machan youn can't use the RAM as a hard disk.There is something like this is
true.
There is a paging mechanism that is used in the x86 and x64 computers.
So some part of the memory will be paged when they are not used recently
by the operating system automatically. we say this page file in windows and
swapping in linux world.This is to increase the performance.It's a good thing.
And that doesn't imply that you don't need High Speed hard disks as the
same speed or likely equally to the RAM speeds. Currently there are hard
disks on the markert which met that requirments. These hard disks are
called SSD disks , Solid State Drives.
http://en.wikipedia.org/wiki/Solid-state_drive
In SSD disks there are no moving parts, instead of moving electrons

or moving parts due to the terminal expansion.
But the technology is too costy still now.
I asked a 3 months ago 40GB costs around 35,000/- Rs , so can you
spend 35,000/- for just 40GB ? hell nooooooooo !
technology is there if you have the money,the problem is the money.
and integrating the CPU and GPU is a good thing. But intel cards are less
features, they are not experts in graphics chips. Even the ATI cards are
better than that intel chips.I think intel have to integrate with Nvdia ,in the
AMD fusion technology AMD is working closely with the ATI ,actually ATI is
owned by AMD now.
Typically when we talk about the graphics world , according to the book of
GPU gems book 1 , the performance does not really depend on the CPU
and GPU communication barriars. There are very small data to communicated
or shared. like vertices ,display lists and textures... very small when comparing the speed of the modern computer bus. Actually most of performance factors depends on the last rendering stage , last stages of the graphics pipeline.so I don't think this be a better idea.
But intel uses the 32nm or 45nm technologies which are really cheap and
cost effective.Nvdia chip technologies can't afford that 32nm technology due to their techinical constrains , so there will be low heat generation because they are 32nm transistors!
And when there are no intermediate chips for interfacing to the busses it will be again cheaper too. However if the GPU unit burns on a chip then whole the CPU have to be thrown away.