|
Cores and effect
|
04/03/2005
|
| |
Never before have processor vendors faced such a precarious challenge, where the demand for increased processing power is finely balanced with the restrictions of power density; a move too far either way will result in lost equilibrium. Doing nothing isn’t an option either, as it would essentially result in stagnation. The only solution is for both sides of the scale to ‘grow’ at the same rate. Adding more features to a processor has, in the past, been the way forward and this would probably remain a viable solution if it weren’t for the fact that it needs to be met with further scaling on the other side of equation. The trend here, of course, has suddenly reversed and the problems with power density at the nanometre nodes are well known.
Simply making bigger processor cores isn’t the way forward anymore. Today, increased processing power is buoyed by the move towards multiprocessor devices, where multiple processor cores are integrated on a single substrate. Obviously, this also results in an increased number of transistors, which isn’t now being met with an equal and opposite improvement in standard process technology. Under these conditions, there is still a danger of imbalance, so some further innovations are needed to counterbalance the demands of the transistors.
|
| |
Author
Philip Ling
|
| |
| |
|
This material is protected by Findlay Publications copyright 2008.
See Terms and Conditions.
One-off usage is permitted but bulk copying is not.
For multiple copies contact the sales team.
|
| |
|
| Email this article |
| |
|
|
|
|