Computer Systems- Computer Architecture

Página 1

Dynamic energy

1.Transistor switch from 0 -> 1

or 1 -> 0

2.½ x Capacitive load x Voltage2

3.Dynamic power

½ x Capacitive load x Voltage2 x Frequency switched

4.Reducing clock rate reduces power, not energy

To be studied

--Cost of an integrated circuit

Being able to expand memory and the number of processors and disks is called scalability.

Pipelining is the best-known example of instruction-level parallelism.

Temporal locality states that recently accessed items are likely to be accessed in the near future. Spatial locality says that items whose addresses are near one another tend to be referenced close together in time.

Amdahl's law states that the performance gain that can be obtained by running in the faster mode is limited by the fraction of time it is run.

There are three type of dependences

1.Data Dependence

2.Name Dependence : In name dependence there is no real flow of data between the instructions i and j

They can be classified into two types.

Anti-dependence : i reads the data in the same register

j writes to the same register

Output-dependence : Both the instructions i and j are writing to the same register.

The goal of both our software and hardware techniques is to exploit parallelism by preserving program order only where it affects the outcome of the program.

A simple scheme for increasing the number of instructions relative to the branch and overhead instructions is loop unrolling. Unrolling simply replicates the loop body multiple times, adjusting the loop termination code.

New Page