Smaller cores = more money for chip makers. They can produce much more.
Printable View
Smaller cores = more money for chip makers. They can produce much more.
This is because of silicon impurities and cost. Silicon wafers have impurities which effectively destroy a working CPU if the impurity is within the CPU area. The larger a CPU die is made, there is an exponentially greater chance that the CPU die will have some impurity that causes the CPU to not work. One you get above something like 500mm^2, it becomes near cost infeasible.Quote:
Originally Posted by epion2985
The art of growing a silicon tower from which wafers are cut is rather expensive. Refer to impurities above. A CPU the size of a playing card would mean that they could only fit perhaps 6 or 8 CPU's on a manufacturing wafer, each of which probably has something like a 80% chance that they won't work, meaning that you're only getting perhaps 1 or 2 working CPU's per wafer, as opposed to around 50-100 CPU's as they achieve nowadays. This is what is called yield.
In short, it's all down to costs and economics. They don't make individual CPU's larger because no one could afford to buy them.
Not at the track, but here's a recent one.Quote:
PS:
hey Cathar, do you have any more videos of you on your bike at the track? :D
http://www.employees.org/~slf/ictohe-q25.mov
Here's a couple more:
http://video.google.com.au/videoplay...47393071011479
http://video.google.com.au/videoplay...21118904104194
HAHAHAH the blacked out SPEEDO :banana: :banana: :banana: :banana: yeahQuote:
Originally Posted by Cathar
we've got a nice little ripper of a road in Royal National Park south of Sydney..........lovely at 4-5am when i'm going fishing :D i'll say no more.....looks familar to black forest
Nice videos :) Looks like fun.
So a good waterblock (for example the storm) on a naked die + liquid metal TIM will result in the best possible temps with current blocks?
I dont think that direct IHS cooling with a unmodded IHS could be any better. Sure you dont have to use TIM, but you still have the crappy paste between the die and the IHS + a flat IHS, which doesnt sound it will be that good.
Direct die cooling seemed to be the only thing I could think of to get even better watercooled temperatures, but with what you posted on the previous page seems to make it harder to realize. I didnt think that water would find its way trough the die. I was more worried about the PCB, which can be made waterproof using silicon, nail polish, glue, or whatever.
Do you have personal experience with that kind of cooling or was your statement based on a theoretical view at it?
Nonetheless I will give it a shot with one of my semprons, but your posts made me think about the whole thing once more.
Thank you for the insights :)
And drive carefully, we want even better waterblocks! ;)